[00:00:09] *** hrist has quit IRC
[00:00:29] *** hrist has joined #opensolaris
[00:00:37] <nikolam> http://www.gnu.org/software/gnufi/
[00:01:05] <lewellyn> nikolam: that was pretty much superceded by tianocore
[00:01:29] <lewellyn> things like ovmf are under a "Freer" license
[00:02:01] <nikolam> lewellyn, you know a manufacturers that use tianocore? is it all free licensed.
[00:02:16] <ubuntujenkins3> hello i am using ubuntu and am doing hg clone http://src.opensolaris.org/source/xref/jds/time-slider/ how ever i am getting abort: 'http://src.opensolaris.org/source' does not appear to be an hg repository! what would the proper address be?
[00:02:17] <lewellyn> sun, afaik
[00:02:23] <lewellyn> iirc, virtualbox uses ovmf
[00:02:56] <lewellyn> vbox only boots linux via efi right now though :(
[00:03:04] *** dunc has joined #opensolaris
[00:03:06] * lewellyn so wants to boot 64-bit windows guests with efi!
[00:03:30] <lewellyn> ubuntujenkins3: have you loaded that url in your browser
[00:03:30] <lewellyn> ?
[00:03:32] *** RoyK^ has quit IRC
[00:04:09] <ubuntujenkins3> lewellyn: yes, i have never used hg before though the error shoudl have been abort: 'http://src.opensolaris.org/source/xref/jds' does not appear to be an hg repository!
[00:04:17] <nikolam> so you say, with vbox module that run on top of efi, you can run multiple OSes and main OS is EFI
[00:04:33] <lewellyn> ubuntujenkins3: if you loaded it in your browser, you'd see that's not an hg repo
[00:04:35] <seanmcg> ubuntujenkins3, thats the src browser i think, not a hg repo..
[00:05:46] <ubuntujenkins3> lewellyn: i wouldn't know a hg repo if i saw one. is it a case fo file save as to get them all then?
[00:05:47] <lewellyn> nikolam: vbox has an efi implementation. right now, as it's new and experimental (read: "broken by default, like anything shiny and new"), it only can boot linux.
[00:06:09] <lewellyn> ubuntujenkins3: no. there should be a readme if you load the page
[00:06:19] <seanmcg> read the jds pages about building it: http://hub.opensolaris.org/bin/view/Project+jds/building
[00:06:39] <lewellyn> maybe not though
[00:06:59] <alanc> ubuntujenkins3: the jds sources are kept in svn, not hg
[00:07:06] * lewellyn makes a note to file a bug that jds needs a readme
[00:07:12] <nikolam> So basically question of an OS on the machine is not between OSes like we know today. But between Bios replacements. Efis and Coreboot
[00:07:22] <lewellyn> alanc: yeah. i made the  naive assumption that they had a readme telling him that :(
[00:07:33] <ubuntujenkins3> alanc: how would i find them then
[00:07:37] <lewellyn> nikolam: please rephrase
[00:07:47] <lewellyn> ubuntujenkins3: read seanmcg's link
[00:08:02] *** raichoo has quit IRC
[00:08:28] * lewellyn hopes no one complains and reboots the server during *this* image-update...
[00:08:37] * lewellyn has no clue why pkg needs to be such a damned ram hog
[00:08:39] *** piwi has quit IRC
[00:08:41] <nikolam> no just thinking about, what will run on my machine 5 yrs from now. first OS, second os, and how it can be remotely controlled if Efi is closed on machine.
[00:08:53] <lewellyn> you choose via efi
[00:09:13] <lewellyn> use something like refit, if your manufacturer gives a sucky menu
[00:09:33] <nikolam> well if efi is closed source implementation. what is there to choose.
[00:09:46] <lewellyn> closed source doesn't matter to the user
[00:09:49] <nikolam> they can take over machines any time.
[00:10:32] <jbit> hrm, is there anyway to pass the "-f" flag to the initial root pool import? :)
[00:10:37] <nikolam> I was reading the other day about some NIC firmware exploits. Today
[00:10:41] <lewellyn> yeah. the post-bios/pre-os boot system is going to randomly call to the fbi and grab a replacement for itself, because they'll be making a new efi firmware for every machine. right.
[00:11:13] <lewellyn> does it significantly impact you that your current bios is closed source?
[00:11:14] <nikolam> lewellyn, no, other side will have and it already had ¨plug in¨ to system
[00:12:02] <nikolam> lewellyn, current bios is stupid. It only boots up. Efi and coreboot are OSes
[00:12:06] <jbit> i think the curent legacy bios "service mode" stuff is more scary
[00:12:20] <lewellyn> no. efi is no more of an os than obp/of is
[00:12:34] <jbit> it is code that runs on a non-maskable interrupt that the OS can no intercept or disable
[00:12:34] <nikolam> jbit, yes, that too.
[00:12:56] <nikolam> jbit, yup
[00:13:05] <jbit> at least EFI disappears if you kill it
[00:13:12] <jnss> my silly eth0 marvell yukon card is not supported with any drivers. what do you suggest i do now
[00:13:17] *** kim0 has left #opensolaris
[00:13:24] <jbit> jnss: buy a new card :)
[00:13:38] <nikolam> jnss, check out HCL
[00:13:45] <jnss> okay
[00:13:46] <lewellyn> think of it this way: forth scared people, so openfirmware didn't catch on with the x86 crowd, so they had to reimplement it and call it efi.
[00:14:37] <jbit> lewellyn: forth was a silly choice
[00:14:48] <nikolam> lewellyn, any closed source implementation of such low level thing is unchecked gate as I see it.
[00:14:50] <lewellyn> jbit: and imo, the only reason it's not on PCs
[00:15:05] <jbit> lewellyn: yeah, otherwise openfimrware was actually pretty rockin'
[00:15:10] <lewellyn> nikolam: so build your own bios and use a fork of ovmf or something
[00:15:37] <jbit> nikolam: and what about the firmware on your wifi/network/disk controller
[00:15:43] <jbit> or hte firmware on your HDDs?
[00:15:55] <lewellyn> the bios, being burnt onto a chip, is pretty much opaque to the user and by definition closed source
[00:16:09] <nikolam> lewellyn, well, there is no 2 core sparc or other low cost sparc eather, too.
[00:16:12] <jbit> i think we can consider the bios a firmware
[00:16:14] <nikolam> jbit, yup
[00:16:20] <lewellyn> nikolam: iirc, obp is closed source, too
[00:16:57] <lewellyn> istr that the "Open" refers to the specification
[00:17:02] <lewellyn> just like efi
[00:17:05] <jbit> lewellyn: you are correct
[00:17:22] <lewellyn> jbit: i know i am. i'm trying to not sound condescending ;)
[00:17:23] *** bobbyz has quit IRC
[00:17:29] <jbit> there is actually only one or two opensource ofw implementations, and they all suck
[00:17:39] <nikolam> lewellyn, aha. you are saying about sparc support.
[00:17:53] <lewellyn> and tianocore is the open-source efi, which is still immature
[00:18:09] <jbit> nikolam: the mac openfirmware implementation was closed source too :)
[00:18:31] <lewellyn> but between efi and coreboot, i'll take efi. it's already supported by hardware and you can boot with efi on windows xp and newer, osx, linux, ...
[00:18:35] <nikolam> jbit, well BSD -> closed source. nothing new
[00:18:50] <jbit> nikolam: what
[00:18:51] <lewellyn> nikolam: you'll see that most of tianocore is bsd, too.
[00:18:55] <lewellyn> so still nothing new, with efi
[00:19:43] <lewellyn> the exceptions, of course, are the things which are not able to be relicensed.
[00:19:46] <lewellyn> e.g. fat32
[00:19:47] <nikolam> bah, we will see. Future is ahead of us. One thing for sure, opensolaris needs PC Bios / Gpt disk support
[00:20:09] <lewellyn> it needs other things more urgently, as there will be dual-mode pcs for a year or two yet
[00:20:34] <lewellyn> and pc bios + gpt = impossible
[00:20:37] <nikolam> dual-mode. like bios and coreboot
[00:20:49] <lewellyn> unless you rewrite your bios to be incompatible with the rest of the world
[00:20:57] <nikolam> http://en.wikipedia.org/wiki/GUID_Partition_Table#OS_support_of_GPT
[00:20:58] <lewellyn> (like coreboot)
[00:21:10] <lewellyn> yes, for gpt boot, the bios needs to know how to boot gpt
[00:21:21] <lewellyn> the os knowing how is irrelevant
[00:21:28] <lewellyn> (at that point)
[00:21:44] <jbit> is there a list of comman line arguments i can provide to the opensolaris kernel?
[00:21:53] <lewellyn> jbit: src.os.o! ;)
[00:22:02] <jbit> lewellyn: yes, i'm looking at that :)
[00:22:04] <alanc> or man boot
[00:22:04] <lewellyn> but, in seriousness, i don't know of a list
[00:22:08] <lewellyn> oh
[00:22:10] <TommyTheKid> you can provice whatever command line arguments you want
[00:22:14] * lewellyn didn't think of man boot
[00:22:27] <ubuntujenkins3> thanks for the help seanmcg lewellyn and alanc I think i understand now
[00:22:35] <TommyTheKid> whether they actually *do* anything useful is an entirely different argument :)
[00:22:37] <lewellyn> TommyTheKid: his question was more "is there a list of useful commandline options"
[00:22:40] <nikolam> lewellyn, well, linux, bsd use gpt now.
[00:22:49] <lewellyn> nikolam: on an efi system, yes.
[00:23:00] <nikolam> lewellyn, on bios as i read table
[00:23:03] <lewellyn> as do osx and xp, vista, 7, 2003, 2008, 2008 r2.
[00:23:18] <jbit> i'm trying to figure out if i can pass "-f" to the initial zfs import
[00:23:22] *** ubuntujenkins3 has quit IRC
[00:23:25] <nikolam> table is saying pc/bios
[00:23:29] <lewellyn> only if the disk has a mbr protective partition, as i keep saying
[00:23:43] <nikolam> aha yes thats it
[00:23:55] *** pizdec has quit IRC
[00:23:55] * lewellyn starts feeling like a broken record
[00:24:17] <lewellyn> big disks are what will drive efi adoption in the next couple years
[00:24:20] * TommyTheKid bonks lewellyn on the side so he skips to the next song
[00:24:22] *** Wicked has joined #opensolaris
[00:24:26] <system5> so the OpenSparc project didn't give out any of the code for the OpenBoot firmware, that's interesting, so I guess the only two still used OpenBoot implementations are both closed source and one implementation is owned by Sun / Oracle and the other one by IBM, or does HP Itanium have OpenBoot as well?
[00:24:29] <nikolam> so basically, no osx and opensolaris on same pc/bios machine , except on separate disks ;)
[00:24:37] * lewellyn installs vista64 on a mac without bootcamp
[00:24:51] * system5 wishes he could afford a mac
[00:24:53] *** leito has quit IRC
[00:24:57] <lewellyn> system5: afaik, itanium is efi
[00:25:04] <reflect> correct
[00:25:10] <jbit> dont you want zfs on a seperate disk anyway? :)
[00:25:15] <lewellyn> nikolam: yes, osx creates a protective mbr
[00:25:19] <jbit> otherwise you dont get the awesome write cache stuff (by default)
[00:25:28] <reflect> SGI:s altix systems has been EFI for.. 4-6 years now
[00:25:30] <lewellyn> jbit: you don't get it on rpool anyhow
[00:25:37] <system5> SGI is a Linux shop now
[00:25:51] * system5 misses IRIX and wishes that Sun had bought SGI instead of Cobalt / Mysql
[00:26:09] <Wicked> hello all. new opensolaris user here. Im looking to see if there are additional repositories i should add? the selection is ok..but kinda limited...im not looking for any particular package...im just testing out solaris and id like to get an idea of whats availible and how recent/updated everything is
[00:26:11] <lewellyn> then alanc would be bitching about more than cde ;)
[00:26:12] <reflect> system5: yes, their altix line has always been linux, I think
[00:26:21] <TommyTheKid> yes, Sun did sooo well with Cobalt :p
[00:26:26] <lewellyn> smrt: explain repostiories
[00:26:27] <smrt> Um. I seem to not know anything about repostiories...
[00:26:29] <system5> anybody besides me ever heard of the 3d file browser that they used to have in SGI Irix?
[00:26:29] <aghaster> quick question... how do I configure services? I need to configure something for the mysqld service
[00:26:32] <lewellyn> smrt: explain repositories
[00:26:32] <smrt> Um. I seem to not know anything about repositories...
[00:26:35] <lewellyn> smrt: explain repos
[00:26:36] <smrt>  There are a few repos available for OpenSolaris. Most have factlets. Search me for more info about any of them! There are: release (what you get off a fresh install), dev ("IPS" in /topic), extra (see https://pkg.sun.com/register for access), contrib and pending (see jucr), webstack, ha cluster (also at https://pkg.sun.com/register for access) , /support (valid support contract required)
[00:26:37] <richlowe> I'm about 90% sure that opensparc did include at least some of the prom code
[00:26:39] <richlowe> at least for T1
[00:26:39] <nikolam> Wicked, update to /dev and see.
[00:26:40] <jbit> system5: hahah, that'd have been awesome (sun+sgi)
[00:26:41] <alanc> aghaster: svccfg
[00:26:43] <system5> it was featured in the movie "Jurassic Park", all the computers in that movie were SGI boxen I think
[00:26:52] <lewellyn> Wicked: there are some third-party repos, too. but try the official ones first :)
[00:26:52] <reflect> system5: yes, it's in Jurrasic Park :)
[00:26:59] <Wicked> thanks guys :)
[00:27:07] *** cybernd has joined #opensolaris
[00:27:09] *** synegy34 has quit IRC
[00:27:14] *** leito has joined #opensolaris
[00:27:20] <lewellyn> system5: and macs
[00:27:28] <system5> I think the 3d file browser is open source, if only I had the free time I'd port it to OpenSolaris and integrate it with Project Looking Glass or Compiz somehow
[00:27:30] <Wicked> nikolam, in the solaris world...is the dev version of solaris considered semi stable?
[00:27:31] <lewellyn> and i think there were a few suns
[00:27:31] <jbit> okay, i gave in, i booted a livecd and force imported my rpool
[00:27:46] <nikolam> system5, and T2 was modeled on Amiga ;)
[00:27:54] <lewellyn> Wicked: the current dev version is the basis for the upcoming release
[00:28:00] <reflect> system5: when did you last use this filebrowser? it's actually not very good :)
[00:28:02] <Wicked> yea.
[00:28:03] <nikolam> Wicked, very semi and its Opensolaris world
[00:28:03] <TommyTheKid> i would consider dev stable for unstable people :)
[00:28:06] *** moepmoep has quit IRC
[00:28:08] <Wicked> ok
[00:28:09] <system5> the old Amiga's were impressive for what they can do
[00:28:14] <Wicked> ill stick with something more stabled/tested
[00:28:25] <lewellyn> TommyTheKid: and release stable for people who like broken stuff ;)
[00:28:30] <TommyTheKid> would I install the companies finance system onto it.. NO
[00:28:33] <system5> if any noobs come in and ask about OpenSolaris 2010.05, the official response should be that it's only being released on the Amiga, hehehe
[00:28:39] <TommyTheKid> do I run it as my primary workstation *sunray ... yes :)
[00:28:43] <jbit> system5: yeah, having tonnes of special purpose chipsets really helped :)
[00:28:48] <TommyTheKid> do I ahve a backup incase the sunray fails, yes ;)
[00:28:54] <lewellyn> Wicked: if you want "stable", use solaris 10
[00:28:55] <nikolam> Basically Wicked its a minefield over time. but Much more stable for development releases of other systems
[00:29:12] <lewellyn> opensolaris is still very young and lots of stuff is being tested with it, so it's inherently a bit "unstable"
[00:29:24] <TommyTheKid> we tried s10u9, and fmd continuously coredumped :p
[00:29:24] <Wicked> ah
[00:29:26] <Wicked> well
[00:29:27] <nikolam> Wicked, you can always have multiple versions of system in parallel in zfs, so, no sweat ;)
[00:29:53] <system5> hey jbit, I just want to say thanks for the work you did on Linux 2.6 zones, very cool stuff :-)
[00:30:09] <lewellyn> now, start on the 2.8 support! ;)
[00:30:14] <Wicked> ah. yea. i know nothing about doing all that right now...brand new to solaris....and last time i tried zfs was on a p4 with 1 gig of ram using freebsd...needless to say performance sucked....
[00:30:16] <reflect> system5: have you seen this; http://www.a-eon.com/6.html
[00:30:19] <nikolam> Wicked, only thing you do not get is Security and other updates for stable osol releases. You need to pay $ Oracle for that.
[00:30:27] <jbit> system5: it's far from done, i only did a crutch :)
[00:30:28] <Wicked> like 3mb/s on a raidz
[00:30:50] <Wicked> wait.
[00:30:59] <Wicked> pay for security updates?
[00:31:10] <system5> A-EON, I thought that link was going to talk about the OpenSolaris NAS platform called EON, and then I noticed it was about something else
[00:31:12] <Wicked> what kinda madness is that?
[00:31:14] <lewellyn> smrt: explain support
[00:31:17] <smrt>  Sun sells support for OpenSolaris at varying subscription levels. http://www.sun.com/service/opensolaris/ and http://www.sun.com/service/serviceplans/software/ for info on them. See also: support repo
[00:31:20] <jbit> "this USB EHCI hostcontroller is unusable" "panic" "BAD TRAP".... yyyaaayy.. one of those days
[00:31:31] <lewellyn> Wicked: if you are ballsy enough to not need support, there's dev
[00:31:41] <reflect> system5: it's about what they call the next generation amiga..
[00:31:42] <lewellyn> jbit: via?
[00:31:50] <jbit> lewellyn: virtual box in fact
[00:31:52] <Wicked> i see. im pretty familar with GNU/BSD/linux
[00:32:04] <system5> I contest the fact that Sun sells support for OpenSOlaris and Solaris since I haven't been able to buy any contracts in the last two weeks
[00:32:08] <lewellyn> jbit: i don't really use usb in vbox, so i don't remember what they emulate
[00:32:10] <system5> :-)
[00:32:16] *** dajhorn has joined #opensolaris
[00:32:27] <nikolam> That smrt info is outdated, I am afraid. You cant buy Opensolaris support now directly. You get it together with Solaris support from Oracle as I know
[00:32:30] <jbit> lewellyn: i'm trying now with the EHCI feature turned off, since i don't really care about it
[00:32:30] <lewellyn> system5: dude. they're merging. there are roadbumps. it's not just you.
[00:32:35] <system5> I guess I don't know the secret Oracle handshake or password or whatever it is
[00:32:46] <lewellyn> nikolam: i'm not updating smrt until oracle brings out the revised support policies
[00:32:47] <system5> to get in to Club Solaris where all the cool people are
[00:33:01] <TommyTheKid> neither do we :)
[00:33:04] *** moepmoep has joined #opensolaris
[00:33:21] <wereHamster> ah, outdated information > no information at all...
[00:33:21] <nikolam> lewellyn, thats good reason. Let everyone know ¨they can¨ buy it..
[00:33:36] <lewellyn> nikolam: i'd rather provide outdated information than uninformed information
[00:33:53] <lewellyn> wereHamster: the url still goes somewhere sane :)
[00:33:56] <system5> what we should do as a community is reach out to Oracle's Linux developers and see if they want to help out with lx brand zones
[00:34:02] <lewellyn> and there's a "contact us to buy" link!
[00:34:12] <nikolam> lewellyn, I tend to call outdated as uninformed but wont sweat about it.
[00:34:20] <jbit> system5: i was thinking about porting freebsds linux stuff to opensolaris
[00:34:23] <lewellyn> system5: it'd be smart to see what oracle wants to do first
[00:34:28] <system5> better Oracle Enterprise Linux (a.k.a. CentOS / Redhat) support and integration to Solaris
[00:34:29] <wereHamster> at least it's not misinformation..
[00:34:31] <system5> I'm down with that jbit
[00:34:32] <lewellyn> system5: they may already have the idea :)
[00:34:37] <system5> hey jbit are you still in school?
[00:34:39] <jbit> system5: i spoke to a freebsd guy about it and he sugested it might be possible
[00:34:52] <system5> I've been trying to figure out how to turn hacking on OpenSOlaris into a project that I can get credit for
[00:34:57] <jbit> system5: no, i've been a professional developer for three years
[00:35:04] <system5> ah, explains why you are so good
[00:35:18] <lewellyn> nikolam: any replacement factlet is pure speculation aside from "it appears that it's difficult to buy support right now. we look stupid for mentioning the word in here. HI MOM!"
[00:35:21] <nikolam> Wicked, there is also Nexenta/StormOS. but they use .deb s ;)
[00:35:21] <jbit> not really, my field is completly unrelated to opensolaris :)
[00:35:31] <Wicked> smrt, explain dev
[00:35:31] <smrt>  see "dev repository"
[00:35:40] <Wicked> smrt, dev repository
[00:35:42] <system5> but yeah, for a computer science degree you need to do at least two projects in your senior year, I don't want to do the lame pre-canned ones
[00:35:43] <nikolam> hi lewellyn ;)
[00:35:43] <jbit> lewellyn: smrt needs symlinks
[00:35:46] <Wicked> smrt, explain dev repository
[00:35:47] <smrt>  To switch to the /dev repository (OpenSolaris development releases, see the IPS version in /topic): pfexec pkg set-publisher -O http://pkg.opensolaris.org/dev opensolaris.org && pfexec pkg install SUNWipkg && pfexec pkg image-update -v # Then wait. You *must* read the release notes before starting out or you *will* be unhappy. See also: ips mirrors
[00:36:17] <jbit> system5: if i had time i'd totally port ZFS to windows :)
[00:36:21] <jbit> just for the hilarity factor
[00:36:25] <nikolam> Wicked, you can also /msg smrt
[00:36:44] <system5> jbit, what do you develop on, since you're a C programmer, right?
[00:36:46] <Wicked> ah ok.
[00:36:49] <Wicked> :)
[00:36:57] <jbit> system5: i'm a ps3 platform programmer
[00:37:14] <system5> I had to go the Linux sysadmin route since I'm low on cash and need money for food and bills
[00:37:24] <TommyTheKid> holy *****, the zip'd ilom/bios for the x4450 is 133MB
[00:37:33] *** victori has quit IRC
[00:37:40] <system5> how's this for an idea jbit- the PowerPC port needs something to focus on since Pegasos is dead and took the port down with it
[00:37:51] <system5> whynot transform the POwerPC port into a Play Station port?
[00:38:01] <Meths> jbit: You develop games or do you develop the stuff that has just broken if people upgrade their firmware?
[00:38:03] <system5> we could re-use the code from the halfway finished NetBSD playstation port
[00:38:06] <jbit> system5: well, sony just killed the ps3 otherOS boot method
[00:38:13] <system5> really?
[00:38:14] <system5> why?
[00:38:16] <jbit> Meths: i'm a game developer
[00:38:20] <jbit> system5: security reasons
[00:38:22] <system5> so Linux doesn't even work on play station anymore?
[00:38:32] <jbit> system5: correct
[00:38:34] <nikolam> system5, playstation3 no more runs linux on slim version. Sony killed ¨other OS¨ on PS3
[00:38:36] <wereHamster> security reasons is just an excuse
[00:38:40] <system5> that is weak
[00:38:57] <system5> how did they kill it? what can they do if you already have Linux installed?
[00:39:05] <wereHamster> fw update
[00:39:09] <jbit> system5: upgrading to a new  firmware removes the option
[00:39:09] <nikolam> sony: we wanted to push usage of our own OS. eol.
[00:39:11] <Meths> system5: It's killed in the latest firmware
[00:39:15] <Stric> system5: you have 3 options.. 1) upgrade and lose "other os"..  2) don't upgrade and lose online gaming.. 3) install hacked firmware to keep online + otheros
[00:39:29] <wereHamster> guess what I'll be doing :)
[00:39:34] <system5> so basically Sony is becoming the next Apple
[00:39:47] <alanc> so now PS/3's will have jailbreak instructions posted for each firmware rev too?
[00:39:51] <wereHamster> except that sony isn't removing features once a product has shipped
[00:39:57] <wereHamster> s/sony/apple/
[00:39:57] <snuff-home> not often u have ppl removing features with each fw update ;)
[00:40:03] <nikolam> wereHamster, I gess you will only delay innevitable bricking later.
[00:40:05] <reflect> nikolam: well, that's been opened up again, from a third party already
[00:40:24] <system5> this is lame, they killed my idea for finishing the NetBSD Playstation port along with an OpenSolaris port that moves forward in concert with NetBSD
[00:40:29] <jbit> wereHamster: i'd recommend against using hacked firmwares since it can be detected and you will get banz0red
[00:40:40] <jbit> system5: there's always the playstation2!
[00:40:40] <nikolam> reflect, nevermind. Why would anyone loose time, when can use only 6 SPE even with that
[00:41:08] <nikolam> and with 256MB.. I would rather have 8GB on PC
[00:41:17] * system5 wonders if he should e-mail Oracle linux devs to see what they think about LX brand zones
[00:41:19] <reflect> nikolam: it's mostly for those who already used linux on their systems.. or who had installed clusters, I guess
[00:41:29] <aghaster> I get the following error message:
[00:41:29] <aghaster> Use 'mysqld -O thread_stack=#' to specify a bigger stack.
[00:41:43] <jbit> reflect: i doubt people with clusters care about upgrading their firmware so they can play MAG online...
[00:41:43] <aghaster> I tried changing thread_stack in my.cnf
[00:41:45] <nikolam> jbit, I wander is PS2 could be used as good X/Terminal on Sunray replacement or something ;)
[00:41:45] <aghaster> didn't work
[00:41:52] <jbit> nikolam: it's not :)
[00:42:08] <reflect> jbit: ofcourse not, but firmware upgrade often brings other things.. ?
[00:42:33] *** blizzow1 has quit IRC
[00:42:37] <jbit> system5: eitherway, i think powerPC is pretty dead these days
[00:42:45] <Stric> nikolam: you know that the ps2 has a 33MHz mips or so, right?
[00:42:45] <jbit> system5: i think the ARM port of opensolaris is much more interesting
[00:42:55] <jbit> Stric: 300mhz mips, 32 mb ram.
[00:43:12] <nikolam> Stric, it is less then Sunray? ;)
[00:43:26] <system5> I thought PowerPC and System Z were dead 10 years ago, but how come IBM doesn't go belly up? That's what I can't understand? and HP didn't go belly up over Itanium?
[00:43:39] <system5> how do they stay in business? it's a mystery
[00:43:46] <jbit> system5: service contracts
[00:43:51] <system5> makes sense
[00:43:58] <reflect> system5: PPC is huge..
[00:43:59] *** ewdafa has quit IRC
[00:44:05] *** Yu\2 has quit IRC
[00:44:07] <system5> the one advantage IBM has over competitors is they maintain 100% backwards compatibility
[00:44:14] <nikolam> system5, novell have suse ent. for that
[00:44:21] <Stric> sorry, I was thinking of ps1 which has a 33MHz mips..
[00:44:34] <jbit> reflect: POWER is huge, PowerPC is mainly used in embedded products these days
[00:44:37] <reflect> your washing machine probably sports ppc.. or your car..
[00:44:39] <jbit> Stric: indeed, with 2megs of ram :)
[00:44:44] <system5> like OpenSolaris just broke all my scripts for "pkg install SUNWpackagename" in the nweest builds, the temptation for me to work on IBM stuff is that I can write something and ignore it for 30 years and it will still work 30 years later
[00:44:48] <system5> that is a huge advantage
[00:45:08] <system5> I wish Sun / Oracle would pickup on that so they could compete successfully against IBM, break their monopoly
[00:45:20] <reflect> jbit: POWER is miniscule compared to the embedded market, I think
[00:45:37] *** Animal-X has quit IRC
[00:45:37] <jbit> reflect: well, i assumed we were talking about hte server/workstation space :)
[00:45:56] * nikolam heard the other day MacroShoft is doing Ford 7 OS ;)
[00:46:02] <system5> I don't like having to relearn the way I do something every 4 years, if I wanted to do that I would develop on Micro$oft
[00:46:04] <reflect> jbit: ah.. ok.. well I just answered "how can they still be around" question :)
[00:46:24] <jbit> reflect: i don't think IBM make much money from embedded PPC CPUs, that's mainly freescales area
[00:46:47] <system5> or Linux, Linux is pretty good at saying- "oh look at feature X, it's working too well, let me break that for you in the next update."
[00:47:03] <reflect> jbit: it would be interesting to see some statistics.. I do know that PPC in that area is.. huge.. and I mean.. humongous..
[00:47:21] <nikolam> reflect, military? government? etc.
[00:47:35] <system5> I think it's the psychology of developers, if you're a new developer just coming in to a project, you don't want to take the time to learn and understand someone else's code, it's easier to just delete large portions of the old code and redo everything from scratch sometimes
[00:47:38] <reflect> nikolam: more like your washing machine
[00:47:44] <reflect> your car stereo
[00:47:46] <nikolam>  reflect lol
[00:47:55] <system5> military / governments are big SPARC customers, I thought
[00:47:57] <reflect> that's how you get volume
[00:48:15] <nikolam> power. as I heard.
[00:48:15] <jbit> reflect: alot of pople are moving from ppc to arm though
[00:48:34] <alanc> system5: pkg install SUNWpackagename works fine, pkg tracks renames and maps the old names to the new ones
[00:48:43] <alanc> we're not all flaming idiots here after all
[00:48:46] <reflect> jbit: yes, seems like it. still, will take some time yet
[00:48:49] <system5> there's no government customer's using AIX, how could there be, Trusted AIX just came out a couple of years ago
[00:48:54] <nikolam> One thing it is lacking is :...  low-cost Sparc.
[00:49:00] <system5> AIX is a noob in this sector that has been SOlaris's backyard for years
[00:49:07] <nikolam> Sparc T1,2 are GPL`ed
[00:49:15] <jbit> reflect: indeed, you don't really want to rewrite your engine managment code base just to move processor arch :)
[00:49:16] <alanc> Sun is the company the Linux people always hold up as being overly obsessed with preserving backwards compatibility
[00:49:34] <TommyTheKid> "was" :)
[00:49:36] <system5> governments have IBM mainframes though which they use for printing out people's paychecks as big batch processing jobs, kind of like what ADP does
[00:50:23] <jbit> system5: so anyway, go work on the opensolaris ARM port :)
[00:50:30] <nikolam> low-cost one/two core Sparc, produced in China? Anyone? :)
[00:50:38] <jbit> i totally want an ARM based opensolaris server appliance :P
[00:50:49] <system5> no, IBM is the company that is overly obsessed with backwards compatibility and Sun / Oracle has to do the same to compete with them
[00:50:58] <jbit> nikolam: there's nothing lowcost about custom asic :)
[00:51:16] <system5> the same COBOL code written in the 1960's is still running today, no one else can say that, but I'm hoping S'norcle will be able to compete
[00:51:29] <system5> competition is good for everyone
[00:51:36] <jbit> low-cost dualcore ARM, running opensolaris.... now that's a good idea :)
[00:51:38] <alanc> OpenSolaris on SPARC still runs SunOS4 binaries from the 1980's
[00:51:57] <nikolam> jbit, you know, one you can put in Opensolaris-powered washing machines.. and RAID cards with Zfs ;)
[00:52:02] <jbit> whack two gigs of ram on it... and you ahve a low cost / low power fileserver
[00:52:11] <system5> jbit: I head the arm port was broken, something to do with czfs (the czech file system) ?
[00:52:13] <wereHamster> alanc: speaking of backwards compatibility.. that's why you acked removing DBE? ;)
[00:52:25] <system5> what exactly was broken about czfs on arm again?
[00:52:29] <jbit> system5: which is why you should fix it!
[00:52:37] <system5> haha
[00:52:44] <system5> ARM is only 32-bit still though, right?
[00:53:13] <system5> I wish we had a Real Time Solaris based OS for robotics, kind of like how they have a real time linux for robotics
[00:53:21] <system5> that is something I would be interested in
[00:53:39] <system5> QNX or whatever it is they use for industrial robotics is ueber-proprietary and closed
[00:53:39] <jbit> system5: czfs is compat zfs btw
[00:53:42] <system5> isn't it?
[00:53:52] <jbit> and yes arm is 32bit
[00:53:53] <alanc> wereHamster: I don't think Sun ever shipped DBE, but it's not breaking binary compatibility
[00:53:54] * system5 isn't an expert in embedded realtime systems
[00:54:06] * system5 hopes someone else here is an expert
[00:54:11] <jbit> system5: i am :)
[00:54:17] <jbit> i worked in embedded field for two years
[00:54:26] <alanc> any X client that uses an X extension is required to test for it's presence first
[00:54:50] <relling> RIM just bought QNX
[00:54:59] <jbit> alanc: where's my NeWS backwards compat damn it! :P
[00:55:30] <richlowe> relling: whoa
[00:55:36] <wereHamster> if you take backwards compatibility just as 'binary compatibility', then I think linux is just as good as linux.
[00:55:42] <CIA-21> Jesse Butler <Jesse.Butler at Sun dot COM>: 6937159 firmware update should update both images
[00:55:53] <nikolam> Symbian is open now, too
[00:55:58] <system5> jbit: have you ever used real time LInux for robotics or other embedded real / time apps ?
[00:56:07] *** blizzow1 has joined #opensolaris
[00:56:07] <system5> I wonder if there's a *BSD equivalent for it
[00:56:30] <jbit> system5: i never used real time linux because back then it was a huge joke
[00:56:36] <system5> I think AuroraUX might shape up to be an OpenSolaris equivalent
[00:56:40] <nikolam> system5, I suppose if there is, its closed sourced by now ;)
[00:56:42] <system5> jbit: so you used QNX or what?
[00:56:51] <jbit> system5: we wrote our own kernel
[00:56:55] <brandon> system5, I don' think QNX is closed source now ... but it isn't exactly under the BSD either
[00:57:31] <jbit> system5: we had two CPUs, one for real time processing (running 100% custom code) and the other for hte user interface and such which ran linux
[00:57:37] * system5 goes to wikipedia to check, and yes he knows that anybody can edit wikipedia making the accuracy of dubious value
[00:57:40] <jnss> automount does not seem to work on a sdhc stick on which i have the marwell drivers
[00:57:51] <TommyTheKid> fast reboot is a bit disconcerting when you are not expecting it :)
[00:57:52] <jnss> i've never really mounted anything on solaris
[00:58:00] <jnss> is it quite different from linux or bsd?
[00:58:07] <jbit> TommyTheKid: hah, i've had that a couple of times
[00:58:18] *** psychicist has quit IRC
[00:58:29] <jbit> TommyTheKid: it's great at like 4am when you start thinking "did i just fall asleep for the BIOS screens?"
[00:58:41] <TommyTheKid> exactly
[00:58:51] <TommyTheKid> only its 5pm and I thought the same thing
[00:59:00] <TommyTheKid> and on an x4450 thats a LOOONG nape ;)
[00:59:24] <jbit> hrm
[00:59:28] <TommyTheKid> these type of machiens are THE reason for fast reboot, they drone on in bios for like 5 mins
[00:59:45] <jbit> my short project to try to get opensolaris running with my pools under virtual box in windows turned out to be a bit longer than i was hoping
[01:00:03] <nikolam> so, played any good games lately on opensolaris ;) Or just doing some OpenCL ;)
[01:00:28] <jbit> nikolam: I'm actualyl writing a game engine which i'm developing on opensolaris right now :)
[01:00:45] <nikolam> jbit, that is just great :)
[01:00:49] <jbit> mainly because of dtrace...
[01:01:06] <nikolam> jbit, will there actually be opensolaris version/release at the end? :)
[01:01:13] <jbit> nikolam: doubtful :)
[01:01:26] <nikolam> shame on you :))
[01:01:50] <jbit> development engnie is very different to the end "player" engine
[01:02:03] <nikolam> I believe number of bugs debugged for Gnome on Solaris/Opensolaris is countless til. now.
[01:02:10] <jbit> unles you like waiting 20 minutes for a level to load while it calculates lighting :)
[01:02:42] <TommyTheKid> wow, there are some interseting errors on the screen booting 134,, something about a malformed adapter name (sorry it scrolled away before I got a good glimpse)
[01:03:15] <system5> ah, you are correct jbit, source for QNX was released: http://en.wikipedia.org/wiki/Qnx
[01:03:23] <nikolam> jbit, but since you know more about that, obviously.. what is difference between it being released for linux and for Opensolaris. You develop it on osol, then why not being on osol, too?
[01:03:42] <system5> had to go start an fsck on a downed Linux server so I was gone for a minute
[01:03:46] <jbit> nikolam: it won't get released on linux either, probably
[01:04:09] <jbit> nikolam: i'll probably target the player for windows and (undisclosed network based platform)
[01:04:26] <nikolam> jbit, Why (Oh why)
[01:04:35] <jbit> nikolam: i could release for solaris, in theory
[01:04:45] <jbit> but since this will be semi-commercial the support would be annoying :)
[01:05:00] <nikolam> release yes yes yes :))
[01:05:01] <jbit> just because an engien runs on something doens't mean it's suitable for public consumption
[01:05:07] <system5> I thknk there is some German company that dominates the industrial robotics industry right now, can't remember their name though
[01:05:19] *** dajhorn has quit IRC
[01:05:53] *** Spencer_tt has joined #opensolaris
[01:05:54] <jbit> nikolam: although that would be an awesome box shot "Requirements: Windows Vista, Windows 7, Solaris 10+"
[01:06:05] *** cybernd has quit IRC
[01:06:20] <jbit> and to be honest, supporting solaris would be far far far far far easier than supporting linux
[01:06:33] <nikolam> :)))
[01:06:33] <jnss> is it possible that my marvell yukon driver is available on the install cd?
[01:06:37] <jbit> since solaris at least has a consistent platform and has a clue what "ABI compatability" means
[01:06:53] <brandon> jbit, do it - i need more distractions on my workstation .. ;)
[01:06:56] <nikolam> well, lets push Sparc version too :)
[01:07:03] <system5> there is a Marvell Yukon driver that Garret D'amore wrote in some newer OpenSolaris /dev/build
[01:07:12] <nikolam> Imagine it on pair of T2+ ;))
[01:07:32] <system5> I'm still using the old YukonX driver for my OpenSolaris and Solaris systems that are cursed with that awful card, and the YukonX driver works fine
[01:07:37] *** falconn1 has joined #opensolaris
[01:07:42] <jbit> nikolam: i don't think many T2 based systems have openGL 3.x graphics cards :)
[01:07:52] <system5> stay away from the skge driver though, that one will panic your system, or it panicked mine anyway
[01:08:03] <nikolam> And the look on the face when gamers read requrements for sparc version ;)
[01:08:07] <jnss> system5, i dont have any inet access on this machine without a working eth0 card. is the driver available by other means, like on the cd
[01:08:13] <jnss> or on the system already, somehwere
[01:08:47] <system5> hey jnss, you can view web pages, and chat on IRC right? I'll give you a link that shows how to install the Yukonx0 driver
[01:08:55] <nikolam> jbit, But wait, imagine Larry`s stage appearence with it, also ;)
[01:08:58] <jnss> yeah
[01:09:04] <jnss> thanks system5
[01:09:11] * jnss gets more water and coffeee
[01:09:12] <jbit> nikolam: unfortuantly, T2 wouldn't make a good gaming CPU
[01:09:23] <system5> jnss, run this command with root level permissions and give me the output:
[01:09:24] <system5> scanpci | grep -i eth
[01:09:30] <jbit> nikolam: while it's better than the T1 since it has more float units (T1 having one float unit.. wtfh)
[01:09:32] <system5> pfexec scanpci | grep -i eth
[01:09:34] <system5> or sudo scanpci
[01:09:39] <nikolam> ok, joking. But windows/Solaris could look interesting :)
[01:09:39] <strawf> jnss: I've a mobo with double yukon gigabit integrated NICs, both working in snv_133, 2009.06 did not recognize them :)
[01:09:40] <jbit> nikolam: it still lacks SIMD and other quite vital things :(
[01:09:45] <system5> | grep -i eth, however you have it set up
[01:10:04] <jbit> nikolam: yeah, maye as a "experimental" release or something
[01:10:12] <system5> does anybody else besides me dislike the use of pfexec as a passwordless sudo?
[01:10:25] <jbit> nikolam: openAL seems to hate boomer though
[01:10:26] <nikolam> jbit, would be glad to put details on Wikipedia ;)
[01:10:27] *** RoyK^ has joined #opensolaris
[01:10:37] <system5> I never log in as that "primary administrator" user that was set up during the install, I created a second user with useradd and set him up for sudo
[01:10:45] <jbit> system5: it's not a passwordless sudo, it's running something in a profile
[01:10:59] <system5> I know, but the profile for the first user is basically full root permissions
[01:11:00] <jnss> system5, it's marvell techg group 88e8039 pci fast ethernet
[01:11:18] <system5> jnss, mine is Marvell Technology Group Ltd. 88E8056 PCI-E Gigabit Ethernet Controller
[01:11:26] <system5> hmm, we might be different jnss
[01:11:32] <jnss> that may be why yours work. yeah
[01:11:33] <jbit> system5: you an remove the admin profile from your account and use sudo if you wish
[01:11:40] *** nitrile has quit IRC
[01:11:46] <jnss> the card is not listed at all in the hcl
[01:11:47] *** harukomoto has quit IRC
[01:12:09] <system5> I just created a second user, I'm worried if the system is bricked I might not be able to fix it in single user mode without that user created during install having all of his permissions
[01:12:35] <strawf> mine is 88e8056, too :(
[01:12:47] <system5> but I"m always worried that I'll forget to lock my monitor screen if I"m logged in as that initial user, and some two year old kid walks up to me computer and types in: pfexec su - root
[01:12:52] <system5> and pwns my box
[01:12:57] <system5> I've never been pwned before and intend to keep it that way
[01:13:07] *** nitrile has joined #opensolaris
[01:13:11] <jnss> i should try changing the id for the driver and see if that will have solaris recognizing it
[01:13:19] <system5> watch out jnss
[01:13:27] <system5> let me research your problem before you do something dangerous
[01:13:41] <system5> I think Murayama's driver might cover your chipset
[01:13:44] <strawf> jnss: are you working with the 2009.06?
[01:13:50] <jnss> strawf, sure
[01:14:00] <jnss> hold on system5. i think i may be on it
[01:14:03] <lewellyn> Currently seeding
[01:14:09] <lewellyn> erp. wrong paste buffer
[01:14:11] <nikolam> slow and quiiet pwned is much worse
[01:14:27] <lewellyn> pkg: There is not enough memory to complete the requested operation.  At least 581MB of virtual memory was in use by this command before it ran out of memory.
[01:14:30] <nikolam> like all those lucky guys that still run 2009.06 without updates.
[01:14:41] <jbit> lewellyn: leak?
[01:14:44] * lewellyn wonders WHY pkg needs to be such a resource hog :(
[01:14:46] <system5> don't you love IPS lewellyn
[01:15:00] <lewellyn> jbit: image-update 133->134
[01:15:11] <system5> imagine poor Olga, trying to build OpenSolaris from source on a computer with 1GB of RAM running OpenSolaris snv_134
[01:15:35] <lewellyn> as soon as it finishes getting around to dying, i'll reboot since i increased the swap zvol
[01:15:38] <jamesd2> lewellyn, because you didn't rewrite to use less and they gave the guy who wrote it 16GB of ram and a small set of packages to manage so he has no need to do a better job.
[01:15:44] <lewellyn> system5: the machine has 600 megs of ram :)
[01:15:52] <nikolam> my nbook has 4Gig But only 3.3Gig is available, due to Unmentioned Intel chipset/Dell bios restrictions etc..
[01:15:58] <lewellyn> jamesd2: don't get me started... :P
[01:16:03] <system5> whowever it is who writes IPS, they should, from here on only be allowed to develop on a Sun Netra T1
[01:16:08] <system5> with an UltraSparc 2 CPU
[01:16:09] <snuff-home> lewellyn: my machine had 512.. i made swap 2gb.. all good for pkg ips ;)
[01:16:21] <system5> who writes this stuff anyway, Shawn Walker writes a lot of it doesn't he?
[01:16:24] <snuff-home> just sloow
[01:16:41] <lewellyn> oh nice... it seems to have killed the machine. i get to rdp into the host now :(
[01:16:46] *** infomomo has joined #opensolaris
[01:16:48] <RoyK^> does anyone here use vbox?
[01:16:53] <system5> Sun Netra T1 should be the official development platform for IPS, I can donate one I have sitting in my bedroom if you want
[01:16:56] * lewellyn spawns some ssh tunnels and sighs
[01:16:59] <RoyK^> seems vbox fucks up quite badly on osol
[01:17:01] <nikolam> jamesd, so thats idea. torture that guy with 1Gb machine for a week? :)
[01:17:16] <lewellyn> nikolam: 512mb
[01:17:25] <lewellyn> and no access to grow the swap zvol
[01:17:38] *** infomomo has left #opensolaris
[01:17:41] <system5> yeah, I want to lock all the IPS developers in a room full of Sun Netra T1's, Sun Netra X1's and Sun Blade 100's all with only 768 megabytes of RAM
[01:17:43] <nikolam> lewellyn, dont be so cruel ;)
[01:17:48] <system5> lock them in their for 8 months
[01:17:54] <system5> and see what IPS looks like when they come out
[01:18:00] <lewellyn> nikolam: i'd LOVE to get by with less ram in my VMs
[01:18:11] <jamesd2> nikolam, i think all testers/qa people should have to do there final testing on a 5 year old machine with minimal config including small slow hardrives in a raidz2 config with dedup enabled
[01:18:20] <system5> exactly!
[01:18:22] <strawf> jnss: just out of interest, have you tried the developer versions 133 or 134? :)
[01:18:33] <system5> the only guy who still develops on this stuff is Martin Bochnig
[01:18:36] * nikolam is crying
[01:18:40] <system5> no wonder he is so pissed off and cranky all the time
[01:19:51] <jnss> strawf, are they more recent=
[01:19:52] <jnss> ?
[01:19:54] <jbit> try developing for systems with 256mb of ram and no upgrade path :)
[01:19:56] <jbit> then we'll talk
[01:20:09] <nikolam> jamesd, and shall we not forget encryption and compression
[01:20:27] <jnss> i have snv 111b
[01:20:36] <jnss> strawf, you may be onto something
[01:20:54] <lewellyn> jbit: if you remember, i was complaining about no way to burn a bin/cue the other day
[01:21:04] <jbit> lewellyn: bchunk
[01:21:07] <jbit> lewellyn: it's your friend :)
[01:21:10] <lewellyn> jbit: no, that turns it into an iso
[01:21:12] <nikolam> someone would say ¨well there will be arm port for your mobile phone with 512 MB, soon¨
[01:21:38] <jbit> lewellyn: and? :)
[01:21:51] <jbit> bin/cue -> iso -> disc -> profit
[01:21:53] <system5> hey jnss, I found your bug: http://defect.opensolaris.org/bz/show_bug.cgi?id=1548
[01:22:04] <lewellyn> turns out the embedded system has a firmware bug where it doesn't see anything till after the first audio track on the disc, and it has a compounding bug where it expects audio gaps to be 1.75 seconds
[01:22:23] <system5> ah, a better version of the bug is here: http://bugs.opensolaris.org/view_bug.do?bug_id=6819638
[01:22:31] <jbit> lewellyn: oh
[01:22:33] <lewellyn> oh, and the data was mode2
[01:22:37] <system5> bookmark the second page jnss
[01:22:40] <system5> the second link
[01:22:55] <jbit> lewellyn: that sounds pretty amazing, but ironicly it sounds like it adheres to hte specification
[01:23:00] <lewellyn> jbit: at least i now know why converting to iso causes the disc to fail when i mail it off :D
[01:23:10] <strawf> jnss: the 133 installation livecd got working driver for my yukon so it might work for you, too
[01:23:12] <jbit> lewellyn: what is the system
[01:23:32] <lewellyn> some embedded crap that i need to know little about, so i ask little more than i need to know :)
[01:23:42] <Spencer_tt> lewellyn: Hiya :) How do I get the jeos vdi image working in VirtualBox
[01:23:45] <jbit> i think cdrecord can burn from bin/cue though?
[01:23:49] <system5> here is your solution jnss: http://opensolaris.org/jive/message.jspa?messageID=318975
[01:23:59] <jnss> thanks
[01:24:04] <jbit> argh
[01:24:04] <lewellyn> Spencer_tt: you follow its readme, then you give up because the performance is terrible :)
[01:24:05] <jbit> windows
[01:24:10] <system5> try the murayama drivers from here:
[01:24:11] <system5> http://homepage2.nifty.com/mrym3/taiyodo/eng/
[01:24:11] <Spencer_tt> :)
[01:24:15] <system5> check if they support your chipset
[01:24:16] <jbit> sees my raid array and wants to format it, scary
[01:24:16] <lewellyn> jbit: it didn't like my burner
[01:24:30] <jnss> ah excelletn
[01:24:41] <Spencer_tt> lewellyn: giving is for romance ;)
[01:24:42] <lewellyn> jbit: vlc and mplayer don't like my drive either, but fluendo's dvd player has no problem
[01:24:46] <system5> what kind of computer / laptop is it jnss ?
[01:24:47] <jbit> zfs doesn't put any protection MBR on raw disks :(
[01:24:53] <system5> so I know not to accidentally buy one
[01:24:54] <system5> heh
[01:25:01] <system5> I avoid Broadcom wireless like the plague too
[01:25:02] <jnss> i'll try that system5.. it's an lg e200, i think
[01:25:22] <lewellyn> Spencer_tt: i got sstallion in on my rdp session and the vbox guys and jmcp claimed ignorance. so i put it toward the bottom of my stack :)
[01:25:32] <system5> jnss do you have Microsoft Windows currently installed on it?
[01:25:35] <system5> or Linux?
[01:25:45] <jnss> system5,  neither
[01:25:49] <system5> if so, how much RAM do you have
[01:26:01] <system5> oh, what's running on that thing, I can't believe that OpenBSD would have a driver for that chipset
[01:26:13] <lewellyn> jbit: but yeah. it apparently does follow all specs, which is why no one's taken the time to fix the firmware bug. since you can work around it so easily with a bin/cue
[01:26:17] *** easyE has quit IRC
[01:26:25] <nikolam> I got working 3G card I put in my Nbook, reading some instructions online, btw
[01:26:27] <Spencer_tt> lewellyn: a good reason to adopt another system.. for OpenSolaris.
[01:26:38] <jnss> system5, do you happen to know how to mount an sdhc card?
[01:26:45] * Spencer_tt waves @ reflect
[01:26:54] <lewellyn> Spencer_tt: have you noticed that i've been advocating using hyper-v as a hypervisor for opensolaris lately? it works pretty well ;)
[01:26:56] <reflect> Spencer_tt: ?
[01:27:11] <system5> sdhc card, are those the small "sim cards" that are used in digital cameras? I just plug them in to the laptop and OpenSolaris automatically mounts it in the /media directory I think
[01:27:20] <system5> same thing with USB thumb drives
[01:27:30] <jnss> my solaris doesn't mount a thing
[01:27:36] <system5> do you have GNOME running?
[01:27:43] <system5> the graphical desktop environment?
[01:27:46] <Spencer_tt> lewellyn: yeah, still that means more hardware..
[01:27:47] <jnss> yes
[01:27:51] <nikolam> Sony Erricson / Toshiba F3507g 3G data card is working (GPS part still not I suppose)
[01:28:13] <system5> yeah, sdhc cards should work
[01:28:15] <system5> this is weird
[01:28:26] <jnss> well, i tried a 500gb usb drive too!
[01:28:27] <Spencer_tt> lewellyn: I'm looking for more with less without the server room ambience
[01:28:30] <lewellyn> Spencer_tt: either that or install via ai into vbox
[01:28:34] <lewellyn> they give a recipe
[01:28:34] <Spencer_tt> as in sound :)
[01:28:43] <system5> jnss, run the device driver utility and tell me how many red lines there are
[01:28:47] *** jayk- has quit IRC
[01:28:58] <lewellyn> bah. you can use a core2duo to run windows server 2008 just fine :)
[01:29:00] <system5> Applications --> System Tools --> Device Driver Utility
[01:29:00] <Spencer_tt> reflect: pm
[01:29:14] <system5> click on it with your mouse (hopefully the mouse works, heh)
[01:29:21] <lewellyn> and then virtualize a bunch of jeos osol instances :D
[01:29:35] <system5> yeah, so is the text installer kind of like an installable JeOS?
[01:29:38] <system5> anybody tried it yet?
[01:29:38] <lewellyn> not that i have a plan *anything* like that up my sleeve...
[01:29:41] <system5> the one from genunix?
[01:29:43] *** Macor has quit IRC
[01:29:45] *** stewart has joined #opensolaris
[01:29:50] <lewellyn> jeos isn't an installer thing
[01:29:56] <lewellyn> jeos is a ai recipe
[01:29:57] <balrog> hi lewellyn
[01:30:01] <lewellyn> balrog: moo
[01:30:07] <Spencer_tt> like a ai text install
[01:30:24] <jnss> system5, no red lines, really. just says driver problems: 1 << and that is the yukon
[01:30:27] <system5> well, there is a text installer for Osol Indiana available now at genunix.org
[01:30:30] <balrog> lewellyn: no working framebuffer driver for osol for blade 150 == stuck on solaris 10 :(
[01:30:40] <system5> I have an angry customer on the phone, be back later
[01:30:49] <jnss> heh
[01:30:52] <Spencer_tt> that's means a weekend for getting jeos to work
[01:30:52] <balrog> but the text installer works
[01:30:56] <balrog> on SPARC, too
[01:30:57] <richlowe> alanc: wasn't blade150 xvr-100?
[01:31:09] <balrog> no, it's mach64
[01:31:17] <balrog> alanc knows, so don't bother him
[01:31:37] <richlowe> huh, learn something new every day.
[01:31:57] <alanc> right, Blade 150 was PGX64
[01:32:08] <balrog> yea, that's more accurate :)
[01:32:15] <system5> ah, the pains of Linux
[01:32:18] <balrog> I think they sold it with an xvr-100 card
[01:32:26] <balrog> but that's no good because I don't have one
[01:32:34] *** stewart has quit IRC
[01:32:36] *** hajma has quit IRC
[01:32:46] *** hajma has joined #opensolaris
[01:32:48] <lewellyn> i wonder whatever happened to the xorg sparc support effort
[01:33:04] <alanc> PGX64 is ATI Rage XL - XVR-100 is Radeon 7000-based
[01:33:27] <lewellyn> ragexl is supported on x86, right?
[01:33:29] <system5> jnss, what operating system do you use when you're not booted up into OpenSolaris? I don't think 2009.06 comes with a c compiler by default so I'm trying to brainstorm a way to get GCC on there if you need it
[01:33:31] <alanc> lewellyn: Martin was the only one interested in doing anything on it
[01:33:48] <system5> since you said you had no network connectivity, neither wireless or wired are working, right jnss?
[01:33:55] <alanc> lewellyn: all the ATI based chipsets Sun used are supported on x86
[01:33:56] <lewellyn> alanc: did he give up on it? has he tried recruiting people since sxce died?
[01:34:21] <alanc> he tries to get interest on the mailing list from time to time
[01:34:30] <jnss> system5, well, wireless is working
[01:34:35] <lewellyn> ok. good. i am half debating a server board which still ships with ragexl, and i've not had time to hcl it yet
[01:34:43] <system5> so you can get out to the internet with wireless jnss?
[01:34:51] <system5> that's good, gives us some room to manuever
[01:34:54] <richlowe> I would assume there's a minimal set of people wanting to use desktop sparcs, and an even more minimal range of framebuffers.
[01:34:58] <lewellyn> alanc: k. i don't recall seeing anything about it this year yet. but it's not like i get very little email :(
[01:35:00] <jnss> system5, eh, i tried creating a adhoc network earlier
[01:35:10] <jnss> let me try that again. just a moment
[01:35:14] <richlowe> pgx32, 64, xvr-100, and c3d would be my guesses.
[01:35:15] <system5> jnss- you don't have a wireless router you can connect to?
[01:35:21] <jnss> system5, nah
[01:35:36] <system5> no wireless cafee or free wireless nodes in your neighborhood?
[01:35:38] <lewellyn> richlowe: however, since there's x86 support, and sparc support on linux, it shouldn't be insurmountable
[01:35:46] <alanc> lewellyn: he sent out some mail in the last week offering to do work on it if someone paid him
[01:35:53] *** bklang has quit IRC
[01:36:00] <lewellyn> ah. i must not have caught up to that point yet
[01:36:02] <alanc> driver-discuss or xwin-discuss I think
[01:36:03] <system5> I know OpenSolaris doesn't ship with aircrack by default, so if there isn't a wide open one, you're going to have problems
[01:36:08] <richlowe> lewellyn: Wasn't saying it would be impossible, just I can't imagine there being many people with the right combination of hardware, interest and ability
[01:36:11] <jnss> system5, course not! everything is wpa at least ;)
[01:36:15] * lewellyn was sick most of last week, and ended up about 15000 emails behind
[01:36:23] *** aghaster_ has joined #opensolaris
[01:36:26] <system5> ok, so jnss, what operating system are you using to chat with me on IRC right now then?
[01:36:37] <jnss> fedora
[01:36:40] <system5> ok
[01:36:48] <aghaster_> does anybody have mysql running on opensolaris? I need to know if you can set the thread stack size
[01:36:48] <lewellyn> richlowe: there are still people interested in sparc32, so who knows. at least the machines with unsupported framebuffers are semi-useful
[01:36:54] <system5> how much RAM do you have on your fedora box?
[01:37:10] <jnss> just 1 gig. it's a netboot
[01:37:13] <jnss> netbook
[01:37:21] <system5> ouch, so virtualbox is a no go there
[01:37:22] <richlowe> lewellyn: I knew there were few folks who cared about US-I running 64bit, there are really people who care about sun4m?
[01:37:23] <lewellyn> richlowe: Not here.
[01:37:32] <richlowe> (tell me, at least, that it's sun4m, not 4c...)
[01:37:56] <lewellyn> there are people who pipe up about running bsd on sun4c and sun4m in various channels, yes
[01:38:01] <system5> I don't know why it doesn't automatically mount the sdhc card, try plugging in the sim card and wait 45 seconds and see if OpenSolaris picks it up
[01:38:06] <system5> it always picked it up on my laptop
[01:38:18] *** ivo_ has quit IRC
[01:38:20] <system5> maybe the sdhc card isn't formatted as fat 32 ?
[01:38:22] <lewellyn> and since i'm /away, i probably should go /away :)
[01:38:24] <alanc> the Linux drivers are out there if people want to port them
[01:38:37] <jnss> system5, sec
[01:38:38] <system5> which Linux drivers, SPARC linux drivers?
[01:38:45] <alanc> yes
[01:39:02] <alanc> well, the SPARC graphics drivers that are used on Linux & BSD
[01:39:07] <system5> what about OpenBSD and NetBSD SPARC drivers, seems the license would be more compatible with CDDL
[01:39:08] <alanc> from the Xorg & DRI proejcts
[01:39:28] <alanc> as far as I know they're exactly the same
[01:39:28] <system5> OpenBSD is the only other operating system besides Solaris that seems to be fixated on SPARC
[01:39:37] <system5> NetBSD SPARC drivers have been flakey for me at times
[01:39:47] *** ivo_ has joined #opensolaris
[01:39:56] <alanc> DRM is dual-licensed BSD & GPL so it can be shared between the Linux & *BSD & Solaris kernels
[01:40:01] <ivo_> any ideas
[01:40:02] <alanc> Xorg is just MIT licensed
[01:40:38] * lewellyn thinks it's silly to dual-license bsd/gpl, still
[01:40:40] <lewellyn> just make it bsd :P
[01:40:55] <alanc> ivo_: no, we have no idea why you just showed up in channel and asked for idea
[01:40:56] <alanc> s
[01:41:15] <system5> hey jnss, plug the sdhc card in to your laptop, wait 20 seconds or so, then try commands like this: ls -lFha /media    or this command: pfexec format < /dev/null
[01:41:29] *** jt_ has quit IRC
[01:41:31] <jnss> okay system5
[01:41:33] <system5> or this one: pfexec rmformat < /dev/null
[01:41:35] * lewellyn has an idea and wanders off to take care of it
[01:41:45] <system5> you might need to use rmformat for a removable device but I'm no Solaris expert
[01:41:47] <lewellyn> echo | pfexec rmformat
[01:41:55] <ivo_> .. hm I just got disconnected
[01:42:08] <aghaster_> in which package is mysqldump?
[01:42:10] <ivo_> what is the best way to autostart a VM with vortualbox
[01:42:12] <alanc> and hadn't said anything in the hours before that
[01:42:21] <ivo_> I have just added a line to a rc script
[01:42:26] <system5> I think pfexec rmformat < /dev/null was the one that always worked for me, but that pip thing lewellyn posted probably works too
[01:42:36] <ivo_> but I don't feel this is the optimal way
[01:42:49] <system5> ivo, try the vbox_manage command line
[01:42:54] <ivo_> I do this
[01:42:55] <jnss> system5, holy crap that worked ;) it's showing the logcial node name
[01:42:56] <system5> you could make an SMF manifest for it
[01:43:06] <system5> rmformat < /dev/null worked ?
[01:43:06] <jnss> i can mount it from here ;)
[01:43:09] <jnss> yes
[01:43:15] <system5> weird that GNOME didn't automount it
[01:43:20] <system5> always automounted it for me
[01:43:22] <ivo_> this with the SMF is good Idea
[01:43:29] <ivo_> but I have no Idea how it works
[01:43:32] <jnss> thanks for your time system5 ;)
[01:43:32] <ivo_>  :(
[01:43:45] * system5 loves the way Solaris names hard drives, hopefully that won't ever get changed
[01:44:08] <ivo_> can I pay somebody a couple of bucks to write the manifest for me
[01:44:09] <ivo_> ?
[01:44:22] <system5> ivo, I bet someone at blastwave.org would do it for you
[01:44:30] <richlowe> system5: 'cos c0t600C0FF00000000007FC940DE0717A00d0s11 is just oh so memorable?
[01:44:49] <system5> dude, it's controller, target, disk, slice
[01:44:53] <jbit> system5: better than windows where it's just "PhysicalDrive<random number>"
[01:44:55] <system5> so much better than the LInux / windoze way
[01:45:20] <system5> I've had problems on Linux where you unplug a drive and now sdb becomes sda and everything gets fubared
[01:45:30] <richlowe> right, and target can be the WWN, which is no fun at all.
[01:45:47] <system5> drives should be named after the sata port on the motherboard or the controller / target / disk / slice
[01:45:48] *** jamesd2 has quit IRC
[01:45:52] <system5> something logical that makes sense
[01:46:09] <system5> the fact that AIX moved away from c0t0d0s0 is a big strike against them in my eyes
[01:46:24] <RoyK^> drives should be named after their god - roy-d1 is a start
[01:46:25] <balrog> system5: linux has a by-uuid and by-port method too
[01:46:31] <alanc> but you just said IBM is perfect and never changes anything!
[01:46:40] <system5> yeah, well they screwed the pooch with AIX
[01:46:43] <system5> hehe
[01:46:58] <balrog> //dev/disk/by-path, /dev/disk/by-id, and /dev/disk/by-uuid
[01:47:03] <balrog> one slash *
[01:47:11] *** hajma_ has joined #opensolaris
[01:47:14] *** jamesd2 has joined #opensolaris
[01:47:53] *** victori has joined #opensolaris
[01:48:03] *** hajma has quit IRC
[01:48:10] <balrog> problem is, most apps reference drives by /dev/disk/
[01:48:11] <system5> but some of their other stuff doesn't change, and I'm telling you keeping a consistent interface makes things better over the long term (think 20 or 30 year period- an eternity in IT years)
[01:48:24] <balrog> err
[01:48:28] <system5> the consistent interfaces drew me to Solaris way before I knew anything about ZFS or dtrace
[01:48:31] <balrog>  /dev/hd* or /dev/sd*
[01:48:41] <balrog> linux has a kernel ABI that always changes, though
[01:48:52] <system5> tell me about it
[01:48:59] <balrog> new kernel? you're probably going to have to recompile, perhaps even modify, your driver
[01:49:09] <system5> the Linux DDI pisses me off even more than the ABI does
[01:49:12] <balrog> allows for rapid improvements though
[01:49:14] <balrog> yea?
[01:49:17] <system5> but Linux is *hype* so everyone wants to use it
[01:49:26] <system5> which means I have to work on it to make money
[01:49:33] <balrog> system5: I can't use opensolaris on ARM
[01:49:45] <system5> what about NetBSD on ARM, is it any good?
[01:49:51] <system5> I haven't tried it
[01:50:53] <balrog> I could try it, but I have only one ARM machine right now and I don't feel messing with it as it's running a ton of services
[01:50:59] <balrog> at least Debian wasn't that messy ...
[01:51:09] <balrog> but I'm not doing system level programming
[01:51:15] <system5> I think Sun could have made money off of the Linux hype if they had talked the Linux devs into getting the mainstream distros (Red Hat, Suse, etc.) to work inside LDOMs
[01:51:39] <system5> well Debian stable is still one of the better Linux distros out there
[01:51:55] <balrog> I'm running testing
[01:52:03] <balrog> stable has too many old packages for my use :(
[01:52:17] <system5> yeah, well how much stuff would break on your ARM system Ballrog, if you upgraded from a 2.4.x kernel to a 2.6.x kernel?
[01:53:02] <system5> and how much will break in the future, since Linus and friends don't plan on changing their ways anytime soon
[01:53:57] <balrog> luckily my ARM system is pretty well supported
[01:54:04] <samc> Any of you guys played with sndr?  Specifically, is it safe to export both sides of an sndr mirror with iscsi and have a client use multipath to access either or both?
[01:54:10] *** winstonw0 has quit IRC
[01:54:35] <balrog> system5: ubuntu is worse though
[01:54:40] *** TommyTheKid has quit IRC
[01:54:44] <balrog> they dropped the ARMv5 platform completely
[01:54:52] <balrog> 10.04 drops ARMv6 as well
[01:54:53] <system5> yeah, I've got some funny stuff for you to read about Ubuntu ballrog
[01:54:59] <system5> since most of my friends use it as their main desktop
[01:55:16] <system5> did Ubuntu drop their port to SPARC Sun4v or is that still alive?
[01:55:31] <balrog> system5: you should use tab-completion for irc nicks ;)
[01:55:39] <CIA-21> David Hollister <David.Hollister at Sun dot COM>: 6938956 firmware upgrade occasionally fails the first time, 6941402 receptacle-pm should be a string array, 6941407 better synchronization between the driver debug log and firmware event log, 6941862 Increase default log buffer size
[01:55:41] <CIA-21> Reed <Reed.Liu at Sun dot COM>: 6933787 scsav3: for mpxio, devid should be registered/transferred to client node
[01:55:43] <CIA-21> Guoli Shu<Kerry.Shu at Sun dot COM>: 6932221 boot hangs in acpi_drv hotkey driver on Dell Adamo XPS laptop
[01:55:53] <system5> now only *Solaris and OpenBSD support Sun4v ? I'm not sure
[01:56:12] *** Aria has joined #opensolaris
[01:56:28] <system5> ballrog, if you want a good laugh, read this: http://linuxhaters.blogspot.com/2009/11/bad-karma.html
[01:56:37] *** Mdx4 has quit IRC
[01:56:39] <balrog> system5: I'm not a fan of that blog
[01:56:44] <balrog> and please use tab completion
[01:57:01] <balrog> ubuntu dropped sparc back in 2008 apparently
[01:57:10] <system5> balrog: my best friend had Problem #2 with his graphics card
[01:57:14] <system5> the strobe light prompt
[01:58:08] <balrog> system5: I don't use xconfig, because I have a custom xorg.conf for two monitors and the nvidia driver on the ubuntu system I maintain
[01:58:24] <balrog> I maintain a bunch of systems, including Mac, Ubuntu, Debian, and Solaris
[01:58:24] <system5> so if S'norcle is gonna sell lots of Sun4v's they need to get other operating systems on board besides just Solaris / OpenSolaris and OpenBSD
[01:58:30] <balrog> debian?
[01:58:53] <system5> balrog: I think Debian might still support Sun4v
[01:59:12] <balrog> there's an official port that supports sun4m and sun4u
[01:59:24] <system5> ldoms seem kind of pointless if you can only run Solaris inside the ldoms
[01:59:28] <system5> why not just use zones then?
[01:59:33] *** |synegy34| is now known as synegy34
[01:59:34] <system5> maybe for people who are security fanatics?
[01:59:42] <balrog> and newer ones
[01:59:54] <balrog> yea, sun4u and sun4v for debian
[02:00:00] <sickness> openbsd supports sun4m and sun4u :P
[02:00:01] <system5> still not enough impetus to sell the millions of Sun4v's you need to sell to make $$$$ to pour into more R&D to make better SPARCs
[02:00:02] <balrog> I wonder what's their driver support
[02:00:05] <sickness> (and really well :)
[02:00:12] <system5> OpenBSD supports Sun4v, I thought
[02:00:13] <balrog> sickness: that was an old wiki post :P
[02:00:18] <sickness> oh
[02:00:19] <system5> I heard you can install OpenBSD in an LDOM
[02:00:30] <balrog> my question is who supports the framebuffer
[02:00:33] <system5> the OpenBSD developers are SPARC die hard fanatics
[02:00:36] <balrog> so far I'm stuck with Sol10
[02:00:38] <balrog> :(
[02:00:46] <system5> kind of what Martin Bochnig would be like if he worked on BSD instead of on Solaris
[02:00:54] *** richb has quit IRC
[02:01:00] *** Shoggoth has joined #opensolaris
[02:01:09] <system5> hey, you got a support contract for patches on your Sol10 balrog?
[02:01:45] <system5> balrog: Solaris 10 isn't that bad if you know how to use www.blastwave.org to install and configure software, i.e. get /opt/csw in your $PATH and $MANPATH, etc.
[02:02:01] <Shoggoth> hi all!
[02:02:16] <system5> Shoggoth is always so excited to be in OpenSolaris IRC
[02:02:29] <system5> makes me more excited about OpenSolaris
[02:02:31] <Shoggoth> indeed!
[02:02:55] <Shoggoth> system5: how's your kmf chops?
[02:02:59] <sickness> system5: yeah, you can :)
[02:03:26] <system5> kmf?
[02:03:30] <balrog> system5: how do you manage with updates and all?
[02:03:38] <balrog> I hate the official update tool
[02:03:45] <balrog> also we're not on a support contract :(
[02:03:56] <system5> apropos kmf says saomething about "key management policy"
[02:04:12] <system5> BTW, the apropos command doesn't work on SOlaris or OpenSolaris if you don't do a catman -w first
[02:04:15] <aghaster_> is there a way to use the mysql 5.0.x package instead of 5.1.x in the amp package?
[02:04:16] <system5> see man catman for more info
[02:04:26] <aghaster_> I installed 5.0.x, but I need to switch to it instead of 5.1.x
[02:04:38] <system5> kmf = key management framework ?
[02:04:51] <Shoggoth> yes
[02:05:11] <Shoggoth> I'm trying to figure out how to use it
[02:05:32] <balrog> system5: ?
[02:05:36] <system5> all I know is that the all knowing and all wise "apropos kmf" command says to look at the kmfcfg command
[02:05:40] <system5> so man -s 1 kmfcfg
[02:05:53] <system5> balrog- I don't think you can patch Solaris 10 anymore without support unfortunately
[02:05:59] <bubbva> pktool is your friend wrt kmf as well
[02:06:04] <Shoggoth> yes I've done that... the problem is I'm not sure exactly what it is I want to do
[02:06:10] <Shoggoth> :)
[02:06:12] <balrog> even for security updates? :(
[02:06:17] <system5> Shoggoth: that is a good problem
[02:06:21] <balrog> I used to use pca for that
[02:06:30] <system5> balrog: yeah, Solaris 10 is now a 90 day free trial
[02:06:37] <balrog> even if you have RTUs?
[02:06:39] <Shoggoth> I've currently got a box running linux with encrypted disks
[02:06:50] <Shoggoth> and I'd like to convert the whole setup into a solaris one
[02:06:51] <system5> balrog: there are places where you can illegally download Solaris 10 patches though, but I won'd mention them here, I don't want to get banned
[02:07:07] <balrog> well that's not an option here
[02:07:07] <system5> balrog: explain RTUs?
[02:07:12] <balrog> right-to-use
[02:07:27] <system5> balrog: if you can still download patches and patch your system, I guess your good then?
[02:07:34] <Shoggoth> so what I really need to know is what's the idiomatic way of doing this in solaris
[02:07:35] <balrog> they'd send you this thing that you could use Sol10 with a certain number of SPARC machines
[02:07:38] <Shoggoth> I'm sure I
[02:07:40] <system5> balrog: you could try #solaris if I'm not helpful enough
[02:07:46] <balrog> system5: pca could download allowed patches
[02:07:50] *** bubbva has quit IRC
[02:08:24] <system5> balrog: if you can download the patches still, then S'norcle must feel that you still have the rights to use Solaris 10 then
[02:08:24] <Shoggoth> need to use pktool but I'm not sure exactly what the correct "solaris way" of doing it is
[02:08:38] <balrog> hm
[02:09:11] <system5> balrog: I think the 90 day trial period only applies to people that downloaded SOlaris 10 for free off the internet within the last week
[02:09:18] <balrog> ahh.
[02:09:21] <Shoggoth> system5, bubbva: any thoughts ?
[02:09:22] <system5> balrog: if you bought SPARC with Sol10 pre-installed you should be fine
[02:09:31] <balrog> no, we bought SPARC with Sol9 pre-installed
[02:09:59] <system5> balrog: interesting, you might be interested in the Solaris 8 and Solaris 9 brand zones for Solaris 10
[02:10:21] <system5> balrog: it's a pretty slick technology for lightweight virtualization of a bunch of Sol8 or Sol9 environments inside Sol10, nothing can beat it
[02:11:11] *** Spencer_tt has quit IRC
[02:11:30] <system5> Shoggoth: I think you're advancing past my level of expertise, it's like when Darth Vader came back in the first star wars movie and now he's better than Obi Wan Kenobi and kicked his but in a light saber fight
[02:11:47] <Shoggoth> lmao
[02:11:50] * system5 ducks as Shoggoth swings his light saber at him
[02:13:17] <system5> Shoggoth: worst case scenario is that nobody here knows and you an try #Solaris if you know at what times there are sysadmins hanging out there
[02:13:21] <melbogia> hey guys, for syslog.conf I can say "daemon.* /var/adm/messages" and it will log daemon specific messages there, but can I filter by a specific daemon, say puppetd.daemon /var/log/puppet ?
[02:13:26] <Shoggoth> what about you bubbva? are you some kind of light saber Wunderkind?
[02:13:39] <Shoggoth> ;)
[02:14:08] <Shoggoth> hmmm... interesting thought
[02:14:28] <jbit> i am now concerned that zfs doesn't put a protective mbr on raw disks :(
[02:14:51] <balrog> system5: all I get are 401 Unauthorized for all patches :(
[02:14:59] <system5> balrog: uh-oh
[02:15:00] <tsoome> protective mbr?
[02:15:02] <balrog> yea, looks lie they're locking us out
[02:15:36] <system5> balrog: is there anyone in #solaris ? there are a couple of guys in there who are *REALLY* good, but they're not always around
[02:15:39] <jbit> tsoome: an mbr that contains a random partition covering the entire disk
[02:16:02] <jbit> tsoome: if you put a raw zfs drive on a windows machine the first thing it asks you is if you want to "initialize" it
[02:16:29] <balrog> system5: well we do license Oracle 11g (educational environment) so Solaris may not be that much more
[02:17:14] <system5> balrog: do you have an oracle rep, maybe you can borrow mine if you don't have one
[02:17:32] <balrog> we probably do
[02:17:37] <balrog> I'll have to talk to administration
[02:17:54] <balrog> we license Oracle DB so I'm sure we do
[02:17:55] <system5> me s firefox just crashed
[02:18:17] <system5> balrog: if you can't find your rep, you can borrow the rep I'm currently negotiating with
[02:18:36] <system5> since I do educational stuff, funneling other universities to the same rep might create a better atmosphere for .edu types
[02:19:21] <balrog> yea, but if we already have one we don't need two
[02:19:26] <system5> ok
[02:19:31] <system5> well talk to your rep then
[02:19:51] <system5> I think they should give Sol 10 security patches away for free to increase adoption of the O.S.
[02:20:04] <system5> for one and two socket Intel x86 / x64 compatible systems
[02:20:12] <system5> more than two sockets, they're free to charge IMO
[02:20:14] <system5> make some $$$$$
[02:20:23] <balrog> we have 0 x86/x64
[02:20:30] <system5> well for SPARC too
[02:20:37] <system5> they're all two CPU systems right?
[02:20:40] <system5> two socket
[02:20:48] <balrog> we have about 30 Blade 150's and like five 280r Sun Fire servers
[02:20:48] <system5> you're not running a Sun Fire e25k
[02:20:55] <balrog> the former are slow workstations
[02:21:02] <balrog> the latter are nice dual-proc servers
[02:21:04] <system5> yeah, I have a blade 100 at home
[02:21:12] <system5> they're good for Solaris 9 and OpenBSD
[02:21:26] <system5> only Martin Bochnig is crazy enough to run OpenSolaris on a Sun Blade 100
[02:22:16] * system5 needs to go do some work
[02:23:41] <balrog> according to the list, they charge $11,568.00 for up to 100 systems for a year
[02:23:47] <system5> balrog: have you tried an OpenSOlaris distro called "Milax" before? I bet it run's well on a Sun Blade 100, same thing for whatever distro Martin Bochnig is writing these days (used to be called MartUX but I don't know what it's called now)
[02:24:05] <system5> $11,568 for Solaris 10 support for up to 100 systems?
[02:24:38] <balrog> http://www.sun.com/products-n-solutions/edu/promotions/download/solaris_support.html
[02:25:02] <system5> that's $115 per system
[02:25:09] <system5> not bad if I had 100 systems but I only have 3
[02:25:17] <system5> heh
[02:25:17] <balrog> we have at most 40
[02:25:35] <system5> I work in a Linux data center with over 50,000 Linux servers though, guess we won't be adopting Solaris anytime soon then
[02:26:12] <balrog> that's for education
[02:26:19] <balrog> so you'd be paying more
[02:26:56] <balrog> do you think they'd cut a better deal as we license oracle? (we have it so that students get oracle for free to take home; I'm sure that's not cheap)
[02:27:47] *** MattBenjamin has quit IRC
[02:29:05] <melbogia> how do i show cpu core information?
[02:29:39] *** aghaster_ has quit IRC
[02:29:43] <jamesd2> melbogia, psrinfo
[02:29:50] <melbogia> thanks
[02:30:02] <balrog> system5: ?
[02:31:28] * system5 went to go work on a Linux server
[02:31:58] <system5> balrog: what university are you at? price might vary depending on where in the world you are located
[02:32:23] <balrog> east coast, Philadelphia to be specific
[02:32:28] <balrog> what Linux stuff are you doing?
[02:32:40] <system5> Red Hat Enterprise Linux / CentOS
[02:32:53] <balrog> hah
[02:32:59] <balrog> that's what we use for servers :\
[02:33:05] <system5> I work in a large data center (well actually 4 large data centers within a few blocks of eachother)
[02:33:31] <system5> I do some network engineering (lightweight CISCO stuff) and RHEL sysadmin work (i.e. managed services)
[02:33:50] <system5> basically people outsource their servers and system administration to us
[02:33:56] <balrog> I'd like to see the SPARC-based Sun Fire 280r's being utilizied more
[02:34:03] <balrog> ahh I see
[02:34:09] <system5> we wanted to do OpenSolaris servers but we evaluated the OS and found it wasn't suitable for servers like Solaris 10 was
[02:34:19] *** swift_ has joined #opensolaris
[02:34:19] <balrog> well yea, I can see that :(
[02:34:28] *** forquare-macbook has joined #opensolaris
[02:34:28] <system5> since it didn't have Flasharchive / text installer for the longest time
[02:34:40] <balrog> well now it finally does
[02:34:43] <system5> Flash Archive is the thing that really makes Solaris 10 the BEST data center OS
[02:34:50] <balrog> yea?
[02:34:54] <system5> balrog: you ever heard of Flash Archive on Solaris 10?
[02:35:01] <balrog> yea but I couldn't figure it out, really
[02:35:15] <balrog> from what I've read it would be the easiest way to do net installs
[02:35:31] *** swift has quit IRC
[02:35:31] <balrog> then again I didn't spend much time trying
[02:35:38] <system5> see: http://www.cuddletech.com/blog/pivot/entry.php?id=245
[02:35:41] *** swift_ is now known as swift
[02:35:51] <system5> basically flash archive or Flash(TM) or whatever you call it works like this
[02:36:01] <system5> say you want to deploy a million web servers
[02:36:07] <system5> or a few thousand mail or DNS servers or whatever
[02:36:27] <system5> you create a "golden image" of what you think the ideal Solaris 10 web server looks like, how every bit of it should be configured
[02:36:40] <system5> then you clone that image out thousands of times via automated Jump Start installs
[02:37:04] *** lattera has joined #opensolaris
[02:37:05] <system5> and the process automatically strips away the sense of system identification (i.e. each box can have a different hostname, IP address, etc.)
[02:37:25] <balrog> nice
[02:37:35] <balrog> that's what I should have done, instead of using postinstall scripts
[02:37:38] <system5> but they are all configured identically (i.e. if they are desktops, they all have the same icons in the same place, if they are servers they all have the same version of Apache and Mysql identically configured)
[02:37:40] *** pjama has joined #opensolaris
[02:37:57] <system5> yeah, Flash Archive is a truly scary Terminator 3 Rise of the Machines type of technology
[02:38:22] <system5> allows one sysadmin to rapidly deploy thousands of servers in a short amount of time provided that they are already racked up by someone else (the shipping guy racks them up maybe?)
[02:38:54] *** swift_ has joined #opensolaris
[02:38:56] <jnss> are dev releases somewhat stable?
[02:39:24] <tsoome> try it out
[02:40:51] *** swift has quit IRC
[02:41:01] *** swift_ is now known as swift
[02:41:10] <jnss> internet is too slow to just 'try things out' need to know what im getting
[02:46:04] <richlowe> Well, recently kinda hardware dependent.
[02:46:05] <richlowe> but "mostly"
[02:46:14] *** forquare-macbook has quit IRC
[02:46:17] <richlowe> some annoying bug folks with intel graphics seem to be suffering from?
[02:46:48] *** forquare-macbook has joined #opensolaris
[02:47:37] <forquare-macbook> Does anyone know if there are snapshots of the dev repo for download?
[02:48:21] <alanc> no snapshots, but I thought you could rsync it
[02:49:24] *** |Inthewings| has quit IRC
[02:55:21] <richlowe> alanc: rsync is just file data, no catalog, etc. "mirror" not "origin", but yes.
[02:55:39] *** sHellUx has joined #opensolaris
[02:55:57] <system5> jnss- the last /dev release that was stable for me was snv_129, but that's just my personal experience
[02:56:33] <forquare-macbook> alanc: cheers, just rsync pkg.opensolaris.org/dev   ?
[02:57:21] <alanc> dunno - I think there's instructions somewhere on the website, but as richlowe said, it's not the same
[02:58:01] <richlowe> I also dunno where the instructions are, because I've always found what it *does* do useless.
[02:58:15] *** jiraia has joined #opensolaris
[02:58:16] <forquare-macbook> Hmmm, ok.  I'l wait around for the next release and download the repo when it becomes available :)
[03:01:10] *** sstallion_away has joined #opensolaris
[03:02:18] *** sstallion_work has quit IRC
[03:02:27] *** sstallion has quit IRC
[03:02:30] *** sstallion_away is now known as sstallion
[03:02:35] *** sstallion has quit IRC
[03:02:35] *** sstallion has joined #opensolaris
[03:02:35] *** ChanServ sets mode: +o sstallion
[03:02:39] *** sstallion_ has joined #opensolaris
[03:02:44] *** sstallion_work has joined #opensolaris
[03:02:50] * sstallion sighs
[03:03:05] <Andys^> ?
[03:03:26] <sstallion> had to ghost a few persistent sessions
[03:03:56] <lewellyn> a few? they'er back! ;)
[03:04:18] <sstallion> unfortunately
[03:08:25] <system5> so does anyone know that if we buy Solaris 10 media whether or not we're still subject to that 90 day trial thing?
[03:08:27] *** ziyourenxiang has joined #opensolaris
[03:08:35] <system5> like if I buy the box set of DVD's / CD's whatever
[03:09:23] <lewellyn> my understanding is that if you have a certificate of entitlement, you're fine.
[03:09:33] <system5> sweet
[03:09:41] <system5> so the media comes with some kind of a certificate then
[03:09:44] <system5> I'm hoping
[03:11:04] <system5> hmmm, I'm looking at this: http://www.oracle.com/us/products/middleware/tuxedo/tuxedo-11g-feature-066057.html and I'm wondering what people usually run Oracle Tuxedo on
[03:11:14] <system5> if more people run it on Red Hat Linux or on Solaris
[03:11:39] *** mroconnor has joined #opensolaris
[03:13:59] <Shoggoth> system5: I think I figured out _what_ it is I was trying to do but I'm having a problem still....
[03:14:17] <Shoggoth> if you take a look at http://arc.opensolaris.org/caselog/PSARC/2007/001/spec.txt
[03:14:18] <system5> lol
[03:14:34] <Shoggoth> there's an example near the bottom:
[03:14:45] <Shoggoth> lofiadm -c aes-256-cbc -a /home/jru/private -T 'Sun Software PKCS#11 softtoken':::mykey
[03:14:54] <Shoggoth> what I want to know is what is the correct syntax for pktool to generate "mykey" in the first instance
[03:15:06] <Shoggoth> any thoughts?
[03:15:27] <system5> my thoughts are that you should join this forum:
[03:15:29] <system5> http://www.opensolaris.org/jive/forum.jspa?forumID=179
[03:15:39] <system5> the OpenSolaris cryptographic framework discuss
[03:15:47] <system5> do both the forum and subscribe to the mailing list
[03:15:52] <Shoggoth> yeah! that might help... thanks
[03:16:12] <system5> my crypto knowledge is limited to SSH and VPN related stuff since I am a network ops monkey type
[03:16:46] <Shoggoth> hey that's ok
[03:22:54] <jamesd2> okay has anyone else had this issue, but ssh from opensolaris to  solaris10u8 isn't working...
[03:24:17] <system5> jamesd2 try this command:
[03:24:19] <system5> ssh -v
[03:24:22] <system5> and see what happens
[03:24:43] <system5> then google for whatever the error is that happens right before it stops doing it's thing or craps out or whatever
[03:24:59] <relling> it works... keeping the riff-raff out ;-)
[03:25:02] <jamesd2> jamesd@frankie's password:
[03:25:02] <jamesd2> debug1: Authentications that can continue: publickey,password,keyboard-interactive
[03:25:02] <jamesd2> Permission denied, please try again.
[03:25:02] <jamesd2> jamesd@frankie's password:
[03:25:23] <system5> lol @ relling
[03:25:36] *** tosh has quit IRC
[03:25:37] <jamesd2> system5, yeah i tried that,  putty works,  sol10u8 to sol10u8 works... just opensolaris to sol10u8 breaks
[03:25:59] <system5> jamesd2, what is your output for cat /etc/release ?
[03:26:21] <jamesd2> amd:~$ cat /etc/release
[03:26:21] <jamesd2>                        OpenSolaris Development snv_134 X86
[03:26:36] <system5> well, I haven't gone past 129, so I don't know what's happening there
[03:26:52] <jamesd2> cat /etc/release
[03:26:52] <jamesd2>                       Solaris 10 10/09 s10s_u8wos_08a SPARC
[03:26:52] *** forquare-macbo-1 has joined #opensolaris
[03:27:00] *** forquare-macbook has quit IRC
[03:28:40] *** rmorse has quit IRC
[03:29:06] <system5> jamesd2, you could try bringing your problem up under opensolaris.org networking discuss:
[03:29:10] <system5> http://www.opensolaris.org/jive/forum.jspa?forumID=3
[03:29:18] *** forquare-macbo-1 has quit IRC
[03:29:18] *** synegy34 has quit IRC
[03:29:30] <system5> or opensolaris.org security discuss:
[03:29:30] <system5> http://www.opensolaris.org/jive/forum.jspa?forumID=37
[03:30:28] *** synegy34 has joined #opensolaris
[03:30:37] <jamesd2> system5, yes but i was hoping this was a known issue.. but i guess people that would know are at the bar on a friday night
[03:34:11] <system5> yup
[03:34:29] <system5> you have a working beadm of snv_129 that you can roll back to?
[03:34:37] <system5> see if the problem persists there
[03:35:12] <system5> I personally thought they should have just done bug fixes from 129 to 2010.04, but they crammed in all these new features in 130 and 131 that scared the crap out of me
[03:35:12] <sstallion> system5: your attachment to snv_129 amuses me.
[03:35:18] <system5> because new features have new bugs
[03:35:23] <system5> hey, I'm holding on to what works
[03:35:36] <system5> like a lifeboat or lifesaver or linus on peanuts with his blanket
[03:35:45] <sstallion> system5: for the most part they all work...
[03:35:48] <sstallion> every build has bugs.
[03:35:58] <system5> well, my 129 bugs were trivial ones
[03:36:17] *** blizzow1 has quit IRC
[03:36:27] <system5> but I would get fired in less than five minutes if I was in jamesd2's shoes and couldn't ssh in to a box I needed to ssh in to
[03:36:34] <system5> I'm using OpenSolaris as my main desktop at work
[03:36:38] <system5> and I don't want to get fired
[03:36:40] <sstallion> system5: I haven't encountered any problems with snv_134 on my test machines... doesn't mean a damned thing though.
[03:36:49] <sstallion> system5: stop using the tip
[03:36:54] <system5> well, then help jamesd2 with his ssh bug then
[03:37:29] <system5> since it's beyond what I"m capable of helping with
[03:37:56] <jamesd2> system5, i can use putty to ssh in... just can't get opensolaris ssh to tallk to solais10.. but 10 will talk to opensolaris its kinda wierd
[03:38:08] <sstallion> jamesd2: ptmx problems?
[03:38:18] <lewellyn> sstallion: i can't upgrade past 129 on most machines. i keep running into various issues ;)
[03:38:18] <jamesd2> what is ptmx?
[03:38:31] <lewellyn> sstallion: s10u8 shouldn't have ptmx woes
[03:38:35] <sstallion> ahh... snv_132 was my magic build, but 134 works fine.
[03:38:51] <sstallion> ahh sorry, I had that backward
[03:39:33] <sstallion> jamesd2: have you tried using -v on client/server?
[03:40:14] <sstallion> system5: Your Mother was a hamster and your father smells of elderberries
[03:40:18] <jamesd2> just a minute i think i found the issue... possibly a bone head error just a second
[03:40:32] <lewellyn> sstallion: ni
[03:40:40] <system5> ah good, so it's not an issue with snv_134
[03:40:58] <jamesd2> yeap... I'm  a bone head...  didnt conigure a domain name.... when i disabled nwam..
[03:41:02] <system5> don't scare me like that
[03:41:15] *** nanase has joined #opensolaris
[03:42:29] <system5> lewellyn, are you telling me that I'm not the only person who has problems upgrading past snv_129 ?
[03:43:18] <lewellyn> system5: you may be the only one not filing bugs
[03:44:24] <system5> see my job is to fix other people's broken Linux server's, not fix my own OpenSolaris desktop if I have problems with it, I had a co-worker who was running Gentoo on his desktop and he got fired 'cause his stuff was breaking all the time
[03:44:36] <system5> when he did unstable Gentoo updates
[03:46:04] <system5> lewellyn: do you have any ATI cards on your machines? I have problems post build 129 on Dell PowerEdge 1950
[03:46:46] <system5> I think the ATI card might have something to do with it, I want to update to 134 on that machine, but the developers who are working in the zones are strongly against changing anything
[03:47:09] <system5> I guess rule #1 of IT is: if it works and the customers are happy, don't touch it!
[03:47:42] *** icarus901 has joined #opensolaris
[03:47:54] <Aria> Yes. /if/ that's not gonna get you one massive vulnerability fail.
[03:47:55] <lewellyn> i only have some rage128 cards pulled from macs
[03:48:04] <sstallion> system5: rule #1 of being employed - don't predicate productivity on unstable software.
[03:48:11] <sstallion> (unless its your job to do so)
[03:48:13] <system5> too late
[03:48:16] <alanc> the only ATI card I have is the XVR-100 in my Sun Blde
[03:48:35] <system5> my PowerEdge 1950 with snv_129 has been up since build 129 was first released
[03:48:38] <system5> massive uptime
[03:48:57] <sstallion> system5: eh? snv_129 isn't massively old.
[03:49:03] <alanc> oh, and I guess there's a Radeon M700 chipset in the Ferrari 4000 laptop I've got buried somewhere - haven't turned it on in many months though
[03:49:12] <system5> yeah, but it's been up since I installed it, months ago
[03:49:16] <system5> with no problems
[03:49:26] <system5> pretty good for an "unstable" developer build
[03:49:29] <system5> solid as a rock
[03:49:50] <lewellyn>   6:54pm  up 184 day(s),  4:40,  3 users,  load average: 0.42, 0.45, 0.45
[03:49:59] <sstallion> alanc: you know, one of the guys at the office found a pair of fully loaded V480's for 80$ a pop
[03:49:59] <lewellyn> THAT's not even massive uptime
[03:50:07] <system5> one of the developers who is a c programmer bought me a beer and gave me props for using OpenSolaris instead of Linux on it
[03:50:08] <sstallion> ... they had rage 128 card installed in them.
[03:50:49] <sstallion> s/(card)/\1s/
[03:51:15] <system5> lewellyn: if you have years and years of uptime than that on Solaris 10 it might mean that you are not patching the system up to date, maybe give me the IP address to see if I can telnet into it (j/k)
[03:51:49] <sstallion> alanc: fwiw, this is lewellyn's fault
[03:52:09] <sstallion> system5: you do realize that ssh was quite available in 2005, right?
[03:52:25] <system5> but anyway, one of the c programmers who is working inside one of the zones on the Power Edge bought me a beer and gave me massive props for using OpenSolaris 129 instead of Linux on that server
[03:52:26] <Andys^> i'm trying to build software that needs C++ "tr1", any idea where i can get that into my opensolaris system?
[03:52:56] <sstallion> Andys^: assuming you are talking about boost, I would start there
[03:53:06] <sstallion> (unless there are other tr1 impls around)
[03:53:06] *** kukhuvud has quit IRC
[03:53:09] <Andys^> no idea what boost is :)
[03:53:11] <system5> he likes the delegated zfs snapshots and dtrace and the zfs compression
[03:53:30] <sstallion> Andys^: http://www.boost.org/
[03:53:33] <sstallion> Andys^: bring a snorkel
[03:53:45] <system5> and he said "I have to give you props because with Solaris, there was no question of uptime, the uptime was assumed"
[03:54:15] * Andys^ is having a string of terrible luck trying to get a bunch of stuff compiled on this osol system :(
[03:54:22] <system5> sstallion, I was referring to the fact that if you had a Solaris 10 box with 5 years of uptime, it might be vulnerable to the telnetd bug
[03:54:26] <lewellyn> Andys^: using sfe? ;_
[03:54:27] <lewellyn> ;)
[03:54:28] <Andys^> i've wasted *days* on this
[03:54:33] *** sstallion has quit IRC
[03:54:51] *** sstallion has joined #opensolaris
[03:54:51] *** ChanServ sets mode: +o sstallion
[03:54:55] <sstallion> bah, vpn tossed.
[03:54:57] <lewellyn> sstallion: stay in place
[03:55:15] <system5> Ands^ what is your $PATH when you build it
[03:55:17] <system5> echo $PATH
[03:55:21] <sstallion> I should probably use screen given my shoddy connection
[03:55:27] <Andys^> the default
[03:55:50] <lewellyn> Andys^: sfe is good for various reasons. ;)
[03:56:03] <lewellyn> Andys^: what are you building, out of curiosity?
[03:56:15] <Andys^> developer libraries of various sorts
[03:56:22] <lewellyn> that's vague :)
[03:56:22] <Andys^> nosql database-related
[03:56:33] <Andys^> well, i've tried several now and run into problems i can't solve
[03:56:58] <Andys^> usually because the authors expect mmap and semaphores and whatnot to work like how they do on linux
[03:57:01] <Andys^> (probably broken...)
[03:57:27] <lewellyn> there's nothing for nosql in sfe
[03:57:45] <lewellyn> at least nothing for ls SFE*nosql*spec
[03:58:37] *** falconn1 has quit IRC
[03:59:02] <Andys^> heh. messing with anything C++ related always makes me glad to return the comfortable, understandable world of C ;)
[03:59:58] <system5> C++ was always more difficult for me than C was, especially looking at other people's C++ code when it's not well commented, gah!
[04:00:05] <lewellyn> C++: The non-portable portable, standardized non-standard language! Look in your grocer's dairy case for it, or ask for it by name!
[04:00:10] <Andys^> hehehe
[04:00:29] <Andys^> it adds so much bloat, and yet they screwed up by not giving people what they actually wanted..
[04:00:46] <system5> there is a hilarious interview with Bjarne Stroustrup
[04:00:50] <Andys^> ahh, i bet this is all broken because i'm not using GCC v4
[04:01:06] <lewellyn> someone, in all seriousness, tried to tell me that linux was the best os out there. because... it's written in "pure c++" :P
[04:01:15] <Andys^> lolwtf?
[04:01:23] <lewellyn> ya
[04:01:42] <lewellyn> i asked him if he's read the sources. he said he didn't have a need to, since he doesn't know c++
[04:01:59] <system5> maybe he confused Linux with BeOS?
[04:02:10] <system5> I think BeOS and Haiku is written in C++
[04:03:15] <lewellyn> bits of beos were C++, sure
[04:03:42] <system5> then it must have been marketing hype then
[04:04:00] <system5> they were trying to sell it as being better because it was more based on C++ (remember this was the 90's)
[04:04:02] <jnss> is there reason to beleive that more recent builds than the one offered by opensolaris.com will have biuiltin support for marvell yukon nic's?
[04:04:06] <jamesd2> well C++  is compatible with kind of ;-)
[04:04:25] <Andys^> everything compiles and works perfectly on ubuntu... when i do finally get this to compile cleanly on osol, it just fails by segfaulting
[04:04:30] <Andys^> insert sympathy here:  [    ]
[04:04:43] * system5 has never looked at BeOS source code so he doesn't really know, but the marketing hype was that it was object oriented and written in C++
[04:05:03] * system5 inserts sympathy
[04:05:19] *** devians[w] has quit IRC
[04:05:32] *** devians[w] has joined #opensolaris
[04:05:39] <system5> Andys^ do you have Nexenta installed in VirtualBox? Try compiling on a Nexenta version that has the same kernel snv build as your opensolaris and see if it builds
[04:05:56] <system5> then you know for sure if it's a problem with OpenSolaris or a problem with how your build environment is configured
[04:06:15] <Andys^> i've actually never tried nexenta
[04:06:19] <system5> Nexenta has the same build environment as Ubuntu
[04:06:21] <system5> exactly the same
[04:06:32] <system5> try either StormOs or Nexenta
[04:06:42] <system5> both should have the same build environment set up as Ubuntu
[04:07:15] <system5> I'm guessing it's something with the build environment, and I bet it will compile smoothly on Nexenta or StormOS
[04:07:22] *** devians[w] has quit IRC
[04:07:25] *** pjama has quit IRC
[04:07:43] <system5> for people who like to talk about C++, this is the funniest interview ever:
[04:07:44] <system5> http://www.nsbasic.com/desktop/info/interview.shtml
[04:07:48] <system5> with Bjarne Stroustrup
[04:07:51] <lewellyn> jnss: if by "downloaded from opensolaris.com", you mean 2009.06, yes.
[04:08:00] <duckinator> system5: blech, C++ :P
[04:08:03] *** devians[w] has joined #opensolaris
[04:08:12] <duckinator> i'll stick to C and ooc for compiled languages, thank you ;)
[04:08:15] <system5> duckinator, read the interview
[04:08:33] <system5> I got kicked out of class once at university for criticizing C++
[04:09:30] <Andys^> wtf?
[04:09:33] <duckinator> are you serious? C++ is horrible
[04:09:45] <system5> exactly duckinator, you didn't read the interview did you?
[04:09:46] <duckinator> i've used it a few times...never again ><
[04:09:50] <Andys^> that makes sense though. alot of C++ is of academic nature
[04:09:50] <duckinator> system5: i am reading it
[04:09:54] * system5 hates C++
[04:10:01] <Andys^> it belongs in universities, not in the business world, IMO
[04:10:09] <system5> especially the part where Stroustrup says:  I thought 'I wonder what would happen, if there were a language so complicated, so difficult to learn, that nobody would ever be able to swamp the market with programmers?
[04:10:25] <duckinator> system5: you should look at ooc...it's pretty neat although not really good for "production" stuff yet i don't think (they're still working on a self-hosting compiler)
[04:10:28] <Andys^> Java is what happened :P
[04:10:34] <duckinator> Java.... *shudders*
[04:10:40] * duckinator pets C
[04:10:47] <Andys^> easy duckinator, this is a Sun-friendly channel ;)
[04:10:52] * system5 loves C
[04:10:54] <duckinator> Andys^: fine, fine. :P
[04:11:17] <duckinator> Andys^: Java's better than C++... i mostly dislike theCommonNamingPracticesPeopleTendToUseWithSomeApps *cough*
[04:11:20] *** relling has quit IRC
[04:11:28] <system5> Stroustrup said: I got some of the ideas from X10, you know, X windows. That was such a bitch of a graphics system, that it only just ran on those Sun 3/60 things.. They had all the ingredients for what I wanted. A really ridiculously complex syntax, obscure functions, and pseudo-OO structure. Even now, nobody writes raw X-windows code. Motif is the only way to go if you want to retain your sanity..
[04:11:51] <system5> the interview is fake but it's still hilarious
[04:11:58] <duckinator> Can we set fire to Xorg? Please? :D
[04:12:06] * duckinator grabs the torches and pitchforks
[04:12:17] * system5 grabs a can of gasoline
[04:12:33] * Andys^ waits for boost to install
[04:12:43] <Andys^> "make" wasn't good enough for these people, there's this "Jam" stuff instead
[04:12:57] <duckinator> make and scons <3
[04:13:05] * Andys^ wonders what this thing is doing, its a pretty powerful system, reduced to a crawl...
[04:13:13] <Andys^> 16 threads, 24gb ram.....
[04:13:59] <duckinator> i use make for smaller projects, scons for larger projects (work on two oses that use scons for their entire source, works very well)
[04:14:21] <system5> Stroustrup said: "You know, when we had our first C++ compiler, at AT&T, I compiled 'Hello World', and couldn't believe the size of the executable. 2.1MB !!"
[04:15:11] <jamesd2> system5, i guess you only used  printf  ... cout << "Hello World!" would of been more ;-)
[04:15:16] <Andys^> heh
[04:15:26] <Andys^> try hello world in Smalltalk :P
[04:15:30] *** lewellyn sets mode: +q system5!*@*
[04:15:53] <lewellyn> this isn't ##c++
[04:15:59] <Andys^> a strangely tranquil peace descends...
[04:19:23] *** sHellUx has quit IRC
[04:20:55] <ziyourenxiang> hello world in gnu smalltalk is small
[04:21:18] <ziyourenxiang> not that i can arsed to compile and see :-)
[04:27:09] <duckinator> is it true that Oracle is acquiring/has acquired Sun? i thought Oracle was just funding some stuff
[04:27:40] <duckinator> hm, apparently it's been acquired by Oracle. hadn't heard about that ^^
[04:27:47] <jamesd2> duckinator, oracle bought sun its complete.. price was about 8billion IIRC
[04:27:52] <duckinator> wow
[04:28:48] <jamesd2> duckinator, you really ought to take a look at a business page in the paper or read tech news once a month or so... that is old news just about everyone has heard about it
[04:29:05] <lewellyn> or visit sun.com occasionally
[04:29:10] *** lewellyn sets mode: -q system5!*@*
[04:30:41] <duckinator> jamesd2: january would be the time when i was using dialup due to the cable being re-ran, though... :P i normally do read some tech news every few weeks
[04:33:03] *** stewart has joined #opensolaris
[04:33:12] *** CodeWar has joined #opensolaris
[04:34:20] <lewellyn> i wonder if there's a market for opensolaris-based web hosting
[04:35:27] <duckinator> is there an (easy) way to install osol without a gui? or at least without it starting automagically?
[04:35:40] <alanc> smrt: explain text installer
[04:35:41] <smrt>  OpenSolaris does not yet officially have a text installer. BUT, there's one coming. Read more at http://sparcv9.blogspot.com/2009/08/opensolaris-text-based-installer.html and http://opensolaris.org/os/project/caiman/TextInstallerProject/ or download a build 134 prototype at http://hub.opensolaris.org/bin/view/Project+caiman/Text+Install+b134+Notes
[04:36:06] <duckinator> alanc: does that start in text mode *once installed*, though?
[04:36:15] <alanc> lewellyn: you mean other than joyent?
[04:36:20] <alanc> duckinator: I believe so
[04:36:29] <duckinator> hmm, ok, i'll look at it
[04:36:39] <alanc> of course you could always do the normal install and just svcadm disable gdm
[04:36:47] <duckinator> ahh, ok :) thank you
[04:36:48] <lewellyn> alanc: yes
[04:37:27] <lewellyn> alanc: just a basic bare-bones AMP web hosting thing :)
[04:38:04] <Andys^> are any of you guys using GCC v4 on osol?
[04:38:26] <sstallion> duckinator: it works pretty well. I used it recently to install on a V120
[04:38:28] <Andys^> lewyllen: isn't that how joyent started? i think the answer is 'no', personally
[04:38:38] <jamesd2> lewellyn, why does anyone care if its opensolaris or solaris or linux, as long as php, mysql and apache are there and confgured for them...
[04:39:22] <lewellyn> jamesd2: people care, for some reason
[04:39:23] <jnss> im planning on 133
[04:39:37] <sstallion> wow
[04:39:39] <sstallion> hrmm
[04:39:50] <jnss> what, i shouldnt go with 133?
[04:40:23] <jamesd2> 134 seems nice for me..
[04:40:29] <lewellyn> Andys^: and, yes, some of the stuff in sfe requires gcc4
[04:40:33] <jnss> or is it 134
[04:40:45] <lewellyn> jnss: 134 is the current dev version
[04:40:48] <jnss> jamesd2, 134 it is
[04:41:55] <Andys^> if they ONLY need php/mysql/apache, then yeah... but usually end users need to request all sorts of other packages over time
[04:44:48] <lewellyn> Andys^: yeah. i've already thought about that part, too ;)
[04:45:45] *** aghaster_ has joined #opensolaris
[04:47:58] <aghaster_> Hi
[04:50:15] <jamesd2> hello
[04:51:26] <lewellyn> moo
[04:51:33] <system5> moo
[04:52:02] <jamesd2> 0/~ who let the cows in ~\p
[04:53:00] <system5> lewellyn: I don't think web hosting on OpenSolaris will work unless you can get all those control panels that that particular industry uses to run on osol
[04:53:40] <system5> customer's are used to having things like Cpanel, Directadmin, Plesk, hsphere, etc. that only run on Linux
[04:53:41] <jamesd2> system5, seems to work okay for joyent
[04:53:53] <system5> Joyent is more than just "web hosting"
[04:54:41] <system5> Joyent markets to a particular group within the rails community, and they have serious cred since some of the company founders at Joyent are the same developers who wrote and popularized rails, mongrel, etc.
[04:55:40] <system5> joyent also has partnerships with facebook and yahoo, etc. where they give free Solaris zones to facebook developers
[04:56:16] <Andys^> and they have to do alot of work to make that happen
[04:56:23] <system5> yup
[04:56:35] <Andys^> whereas someone like myself, instead, can just install linux on a VM...
[04:57:14] <system5> I deal with a lot of PHP developers on a daily basis (since I work for a company that charges them $70 an hour to fix their broken Linux servers) and my experience with them is that most of them are helpless given only SSH and a command line
[04:57:14] <Andys^> since i don't really have the resources to effectively make my own distribution of opensolaris
[04:57:57] <system5> the typical PHP / Lamp developer would have a hard time in a Joyent Accelerator / zone because he doesn't have the cpanel GUI to make it easy for him to admin things by clicking with a mouse
[04:58:11] <system5> but I think Joyent might have a GUI product too, not sure
[04:58:27] <system5> the closest thing we have on OpenSolaris is a webmin package
[04:58:53] <system5> if you can sell PHP developers on the OpenSolaris webmin package and take regular snapshots of their zones (because they probably will get hacked) then you could make it work
[04:58:56] *** x58 has joined #opensolaris
[04:58:59] <Andys^> system5: i think they offer virtualmin/webmin
[04:59:16] <lewellyn> system5: some work on solaris just fine :)
[04:59:20] *** mariusz has joined #opensolaris
[04:59:30] <Andys^> right.. cept you can take snapshots of any sort of virtualised disk if you are storing it on a ZFS-based NAS/SAN
[04:59:36] <system5> oh I know some do lewellyn, but web hosting is a very un-profitable low-margin business
[04:59:43] <system5> you need to attract a large number of developers
[04:59:47] <x58> How do I remove a zone from zoneadm which was never installed but has the same path as a zone that I did initialise and install correctly (I apparently mistyped the zone name when i created it ...)
[04:59:56] <lewellyn> i'm just looking for enough to cover the costs of sticking a machine in a colo
[05:00:14] <Andys^> lewellyn: you're probably better off just using an opensolaris-friendly VPS provider
[05:00:30] *** nanase has quit IRC
[05:00:39] <system5> and a large number of them are very dependent on web gui interfaces like cpanel (see image at link): http://www.hostingzoomreview.org/images/hosting-zoom-cpanel.gif
[05:01:04] <system5> lewllyn, have you tried doing Voip on Solaris? set up an Asterix PBX and it might cover the cost of the colo and use very little bandwidth and CPU
[05:01:21] <lewellyn> Andys^: that's a different project
[05:01:34] <system5> check this out lewellyn: http://www.thrallingpenguin.com/articles/asterisk-solaris.htm
[05:01:40] <lewellyn> system5: i can't get * working on an "easily"-supported environment
[05:01:45] <lewellyn> so OCS ftw
[05:01:47] *** gm152 has quit IRC
[05:01:57] <lewellyn> yay transcription to email!
[05:02:23] <system5> every small business needs a PBX right, so you sell a few of them and you've covered the cost of the server using very little bandwidth and CPU and RAM in the process
[05:02:58] <system5> lewellyn: have you tried these asterix packages-   http://www.solarisvoip.com/
[05:02:59] <lewellyn> i'd rather sell OCS. i make more off it ;)
[05:03:11] <lewellyn> 20:06 @lewellyn: system5: i can't get * working on an "easily"-supported environment
[05:03:39] <system5> what's OCS, is that some Microsoft thing?
[05:03:42] <lewellyn> yes
[05:04:01] <system5> ic
[05:04:12] <lewellyn> pretty much voip with fewer cert advisories than * gets
[05:04:15] * lewellyn wanders off
[05:04:43] <system5> I'm just saying I know people who run small Callweaver or Asterix setups and pay for the server that way, which is what you seemed to be trying to do
[05:04:55] <system5> they run these setups on Linux, but should work even better on Solaris
[05:05:10] <sstallion> system5: oh?
[05:05:13] *** anibal has joined #opensolaris
[05:05:30] <system5> well Solaris has better multithreading than Linux, so it should run VOIP applications better than Linux does
[05:06:10] <sstallion> system5: it does?
[05:06:24] <Andys^> great, but the last thing i'm worried about when running a voip app is "oh no, 100% cpu time is being used"
[05:06:35] <Andys^> unless you're doing transcoding i guess
[05:06:54] <system5> sstallion, check this link: http://www.thrallingpenguin.com/articles/asterisk-solaris.htm
[05:07:33] <system5> if you guys at Sun want to sell lots of Sun4v boxes, you should make a case that the superior multithreading makes it handle Voice over IP better than x86 while using less resources (electricity, air conditioning, etc.)
[05:07:39] <sstallion> thats some pretty outdated information there
[05:07:45] <system5> yeah it is outdated
[05:07:49] <system5> you got more updated info?
[05:08:45] <system5> hey sstallion, aren't you the guy who integrated FreeRadius into OpenSolaris IPS?
[05:08:49] * lewellyn peers in on his way out and realizes no one active works for sun and wonders who system5 is talking to
[05:08:56] <sstallion> system5: yes.
[05:09:09] <system5> thought that was you
[05:09:18] <lewellyn> unless sstallion does and i didn't realize it, from the cloak
[05:09:23] <system5> thanks for that btw
[05:09:34] *** stewart has quit IRC
[05:09:39] <sstallion> lewellyn: nope. simple bystander here.
[05:09:43] <lewellyn> thought so
[05:09:53] <lewellyn> did you catch my jucr versioning question earlier, btw?
[05:09:54] <jamesd2> system5, most of sun/oracle's customers are using cisco's call manager for voip
[05:09:57] <sstallion> system5: sure. glad you found it useful.
[05:09:59] <lewellyn> about zeroes?
[05:10:06] <sstallion> I did
[05:10:16] <sstallion> ISTR there being problems. I solved it a little differently
[05:10:20] <sstallion> let me find an example specfile
[05:10:27] <lewellyn> istr there are, too. but i'm failing to find docs
[05:10:32] <system5> yeah, I had problems getting your version of Radius to work
[05:10:38] <system5> so I was using the blastwave one
[05:10:54] <lewellyn> i expect a 1.0 "soon", so i'm reluctant to chop off the leading digit
[05:11:08] <system5> the CSWfreeradius
[05:11:38] *** jiraia has quit IRC
[05:11:43] <system5> sstallion: have you tested your radius package in a zone?
[05:12:02] *** kwinz3_ has quit IRC
[05:12:26] <sstallion> system5: yes. that is how I use it.
[05:12:35] <sstallion> lewellyn: http://arf.ubound.org/repos/opensolaris/specfiles/file/7a0e006040a4/media/firefly.spec#l1
[05:12:44] <sstallion> thats a good example of a strange version number
[05:13:19] <sstallion> system5: odd. I've used that package for quite a while.
[05:13:41] <lewellyn> sstallion: but that doesn't really address my case, which is a not-terribly-unusual version number
[05:13:53] <lewellyn> 0.8.12
[05:14:05] *** kwinz3 has joined #opensolaris
[05:14:05] <sstallion> right
[05:14:21] <lewellyn> (the .13.1 is to get around the fact that i know that it has to be all-numeric, and it's 0.8.13-RC1)
[05:14:26] <sstallion> you need to make up a substitute version number that will work with pkgbuild
[05:14:33] <sstallion> yeah
[05:14:40] <sstallion> I think I have a zero version in here somewhere let me dig
[05:14:53] <lewellyn> it'd be great to find a page that documented this :P
[05:15:07] * lewellyn debates just trying and filing a bug if it pukes
[05:15:13] <sstallion> http://arf.ubound.org/repos/opensolaris/specfiles/file/7a0e006040a4/media/ffmpeg.spec#l1
[05:15:26] <sstallion> 0.5 seems to work ;)
[05:15:37] <lewellyn> ok. good
[05:15:50] <lewellyn> i've got to get some deps into jucr and promoted first, anyhow, i fear
[05:15:54] <sstallion> course I never fed that one through jucr - only built locally.
[05:16:01] <sstallion> thats always a PITA
[05:16:02] * lewellyn hopes jucr voting/promotion starts happening again soon
[05:16:16] <sstallion> I need to update my packages for the release
[05:16:17] *** wrapster has joined #opensolaris
[05:16:21] <wrapster> http://hg.genunix.org/onnv-gate.hg/rev/308d70e4dce8
[05:16:32] <wrapster> can any one pls tell me a bit more about this..
[05:16:35] <wrapster> what is elpaso?
[05:16:42] <lewellyn> 4 and a half months since the last big push to contrib.
[05:16:45] <lewellyn> so long :(
[05:16:51] <sstallion> wrapster: a city in texas.
[05:17:09] <wrapster> sstallion: :) yes i know in the context of opensolaris...
[05:17:11] <lewellyn> sstallion: btw, can you vote on jucr? :)
[05:17:12] <wrapster> i meant.. :)
[05:17:22] <sstallion> lewellyn: unfortunately no.
[05:17:24] <lewellyn> wrapster: i'd search src.os.o
[05:17:27] <lewellyn> sstallion: bah :(
[05:17:38] <wrapster> and yes I have been to that place.. too damn hot so couldnt see much of it.
[05:17:54] <sstallion> wrapster: looks like some lock cleanups
[05:18:25] <sstallion> lewellyn: I'm a lowly device contrib - my packages are just products of me trying to do something useful ;)
[05:19:00] *** Calibellus has quit IRC
[05:19:42] <lewellyn> sstallion: my packages are mostly to scratch my own itches, too ;)
[05:20:01] <sstallion> lewellyn: mine are more like rashes :D
[05:20:59] *** chendy has joined #opensolaris
[05:21:17] <lewellyn> i'd *love* to start seeing jucr promotions more often though. i have lots of stuff to contribute. i may end up sticking to sfe at this rate, if i can't get deps promoted :P
[05:21:32] <lewellyn> anyhow. i have a gf who is now home therefore it's time for me to go :)
[05:22:11] <sstallion> haha
[05:22:17] <sstallion> I need to hit the sack. Long day tomorrow.
[05:22:35] <sstallion> I still have a few packages that are pending as well. I've had a finalized amanda package for almost a year now.
[05:23:34] <sstallion> night all
[05:23:37] <system5> when you create IPS packages is there a standard for how the SMF FMRI is supposed to be named?
[05:23:42] <system5> ok, good night sstallion
[05:23:50] <sstallion> there is
[05:23:59] <sstallion> I want to say 'packagename'.xml is expected
[05:24:09] <system5> blastwave calls their FMRI: svc:/system/radius:default
[05:24:22] <sstallion> blastwave is very wrong
[05:24:23] <system5> while yours is svc:/network/radius/server:default
[05:24:26] <system5> ok, ic
[05:24:33] <sstallion> last I checked, radius is not an essential system service
[05:24:37] <system5> ic
[05:25:09] <sstallion> look at conventions used by dns/ldap etc.
[05:25:11] <system5> why not just call it svc:/network/radius:default or does that violate a rule?
[05:25:14] <system5> ok, I will
[05:25:28] <sstallion> because there could be client fmri's added later as well
[05:25:33] <sstallion> think dns/client v.s. dns/server
[05:25:42] <system5> just seemed easier to type in "svcadm enable radius" than to type in "svcadm enable system/radius:default"
[05:25:44] <sstallion> (there is a freeradius-client package)
[05:25:49] <system5> I don't know the standards though
[05:26:09] <sstallion> anyhow, I'm out
[05:26:12] * sstallion &
[05:26:19] <system5> errr I mean easier to type in "svcadm enable radius" than to type in "svcadm enable radius/server:default" my bad
[05:26:23] <system5> ok good night
[05:26:29] *** sstallion has quit IRC
[05:33:48] *** fOB_2 has joined #opensolaris
[05:38:17] *** fOB has quit IRC
[05:47:41] *** wrapster has left #opensolaris
[05:52:30] *** ziyourenxiang has quit IRC
[05:55:27] *** andersenep has joined #opensolaris
[05:58:18] *** leito has quit IRC
[05:58:52] *** brandon is now known as Jondice
[05:59:42] *** chendy has quit IRC
[06:03:30] *** anibal has quit IRC
[06:05:21] <CodeWar> noticed a statement earlier about Solaris/Linux threading one better than the other .. hard to quantify subjective statements like that and I m not aware of anything superior in Solaris scheduling other than the FSS class over Linux (please educate me if there is) *but* ..my observation being
[06:05:43] *** leito has joined #opensolaris
[06:05:59] <CodeWar> most concurrency issues.. you tend to get better over tiem .. finding bottlenecks having good tools to narrow them down and fix them .. Solaris running on SPARC machines with crazy lotta threads does have that experience
[06:11:03] *** blizzow1 has joined #opensolaris
[06:12:55] *** CodeWar has quit IRC
[06:13:54] <jamesd2> CodeWar, solaris has a more complex and flexible scheduling,  it has 3 diferent schedulers that you can use and assign to processes as necessary including a soft realtime scheduling option that is included out of the box, linux soft real time is an optiion and isn't enabled by default.
[06:19:06] *** Nemesis has joined #opensolaris
[06:19:28] <Nemesis> whats the command to get an i/o address readout of all the hardware in your box
[06:19:46] <Nemesis> something like the resource tab in windows
[06:20:15] <Nemesis> prtconf does not tell me i/o addresses
[06:22:55] <jamesd2> prtconf should have an option check its manpage
[06:24:20] <jamesd2> what exactly are you trying to do... most of the time unless you are writing kernel drivers you don't need to worry about it
[06:26:39] <Nemesis> well..i have an old nic that has a static i/o address of 100...I need to see if any other device in the box also likes that address
[06:27:09] <Nemesis> i may have found a utility to help...its called ioview.zip
[06:27:18] <Nemesis> or ioview
[06:27:36] <Nemesis> runs from msdos
[06:28:42] *** corretico has joined #opensolaris
[06:28:47] *** andersenep has quit IRC
[06:30:51] *** RoyK^ has quit IRC
[06:31:49] *** winstonw has quit IRC
[06:32:27] <jamesd2> Nemesis, are you sure that opensolaris supports a nic that old?
[06:34:18] *** vivid has quit IRC
[06:36:29] *** vivid has joined #opensolaris
[06:37:27] *** winstonw has joined #opensolaris
[06:37:39] <Nemesis> jamesd2 how would I be sure that opensolaris supported the nic when the hardware compatibility list is totally incomplete..
[06:38:00] <Nemesis> all of the nics provided on the list "as supported" are either pcmicia or pci
[06:38:12] <Nemesis> in my manufacturer category
[06:38:22] <jamesd2> and what is your nic?
[06:38:34] <Nemesis> 3com 3c507tp
[06:38:59] <Aria> Yow. Dspam is utterly hammering our database.
[06:39:22] <richlowe> that's not the HCL being incomplete, that's your NIC being ancient.
[06:39:59] <JeremyK> isn't that just a tulip driver?
[06:40:15] <JeremyK> or rather.. is there an opensolaris version of the tulip driver :)
[06:40:40] <Nemesis> [00:44] <richlowe> that's not the HCL being incomplete, that's your NIC being ancient. <-- ancient to who?
[06:40:49] <jamesd2> does opensolaris even support  isa slots... its a 10mbit card.. better off getting a used 100mbit or even better intel gigabit card... of course this may be a 366mhz p2 ... with 512MB of ram max.
[06:40:57] <Aria> 3c507 /is/ pretty damn old...
[06:41:01] <Nemesis> no equipment is "ancient" if it does the job..
[06:41:13] <jamesd2> Nemesis, even the geico cavemen have moved on to pci 100mbit nics.
[06:41:47] <Nemesis> why do I need 100mbit when I only get 2.3mb/s over samba
[06:42:00] <jamesd2> Nemesis, latency...
[06:42:31] <Nemesis> look...i only need a 10mip nic for the job
[06:42:39] <Nemesis> and ISA will suffice
[06:42:56] <Nemesis> but if opensolaris dont support it..then I guess I cant use it..
[06:43:08] <Nemesis> but to answer your question from earlier jamesd2...
[06:43:11] <jamesd2> okay plug it in and get an opensolaris live cd, and see if it boots, and see's your nic.
[06:43:38] <Nemesis> there is no possible "way" for me to know what opensolaris totally supports with an incomplete and partial hardware compatibility list...
[06:43:45] <jamesd2> Nemesis, you do know that opensolaris needs at least 512MB of ram to boot.. probably more is better.
[06:44:08] <Nemesis> the only way to know if something works with it..is to simply try it..but if it is tried..someone needs to document that it did work or did not work when they tried it
[06:44:35] <Nemesis> i have plenty of ram for it
[06:44:39] <Nemesis> it runs fine
[06:45:09] <Nemesis> is there a place for me to document where I tried these nics and they did not work..
[06:47:30] <JeremyK> jamesd2: does it really need 512MB to boot at all?  or just to boot up into X?
[06:47:51] *** plompi has joined #opensolaris
[06:47:51] <JeremyK> first thing I do on my opensolaris systems is enable root ssh logins.  second is I disable gdm
[06:47:54] <jamesd2> JeremyK, it is just to boot,  zfs is a pig.
[06:47:55] <JeremyK> :P
[06:48:08] *** andersenep has joined #opensolaris
[06:48:17] *** plompi has left #opensolaris
[06:48:45] *** deet has quit IRC
[06:48:56] *** blizzow3 has joined #opensolaris
[06:49:07] *** deet has joined #opensolaris
[06:49:17] *** blizzow1 has quit IRC
[06:50:23] <jamesd2> JeremyK, and good luck trying to run pkg  update in /dev ... it takes over 1GB of complete, so  swap will be necessary and patence is a must.
[06:51:08] <lewellyn> Nemesis: afaik, there's no isa support anymore
[06:52:13] * lewellyn just had pkg run out of ram on a system with 1.2gb :P
[06:52:22] <Nemesis> [00:55] <lewellyn> Nemesis: afaik, there's no isa support anymore <--- whose bright idea was it to not support ISA hardware especially in nix?
[06:52:51] <lewellyn> Nemesis: solaris only recently, by hardware standards, started supporting x86
[06:52:53] <Nemesis> look..im not asking to put nix on a mainframe here
[06:53:05] <Nemesis> isa nics work just fine..
[06:53:16] <lewellyn> you're asking to use hardware which was already dead by the time solaris ran well on x86
[06:53:16] <Nemesis> ic
[06:53:28] <lewellyn> you're welcome to write the driver yourself ;)
[06:53:30] <Nemesis> i understand what you are saying..
[06:54:22] <lewellyn> yes. i'm saying that it's unix. it wasn't originally written with commodity hardware in mind ;)
[06:54:59] <Nemesis> i see what you are saying...its a solaris specific issue probably...just like freebsd also has one with my "up to date" raid card
[06:55:16] <Nemesis> in fact freebsd wont even boot
[06:55:24] <lewellyn> of course, when we move to efi in a year or so, nothing will support your isa card ;)
[06:55:34] <lewellyn> Nemesis: does windows like your 3c507?
[06:55:41] <lewellyn> does 3com have a vista driver for it?
[06:55:42] <Nemesis> because of the raid card..anyway..I will just stay with my pci nic for now
[06:56:06] <Nemesis> lewellyn...Windows supports "ALMOST EVERYTHING"
[06:56:14] <lewellyn> hahaha. as if.
[06:56:18] <Nemesis> i have no intention of running windows though
[06:56:26] <Nemesis> i am quite happy with nix
[06:56:52] <lewellyn> does any other unix besides netbsd support it?
[06:57:07] <Aria> Linux.
[06:57:12] <lewellyn> i said unix
[06:57:13] *** winstonw has quit IRC
[06:57:29] <Aria> Heh.
[06:57:43] <jamesd2> lewellyn, what's the difference, the box will probably die before even linux can crash on it...
[06:58:15] <lewellyn> oh. Nemesis is the one with the low-ram?
[06:58:33] <Nemesis> low ram?
[06:58:35] <lewellyn> i sat down with the3c507 at the top of my screen and didn't scroll up
[06:58:45] <Nemesis> osol has 512mb free right now...LOW??
[06:58:46] <lewellyn> also, i've tested booting as low as about 300 megs. ;)
[06:59:00] <jamesd2> lewellyn, its a box with isa... how can it handle much ram at all.. and jump buffers will kill performance anyway..
[06:59:01] <Nemesis> out of a total 768 osol has 512 free
[06:59:14] <lewellyn> jamesd2: do i need to link you things? ;)
[06:59:31] <Nemesis> its like I said before osol is not doing anything over here, but sitting here idly by catching files..
[06:59:38] <lewellyn> jamesd2: like this http://www.adek.com/ATX-motherboards.htm
[07:00:11] <Nemesis> im sure we could pound this server is some environments...
[07:00:19] <Nemesis> but thats not the issue here
[07:00:30] <Nemesis> i dont need all that to do what I simply need to do here
[07:00:32] <lewellyn> it's unfortunate they don't make the boards with pcie, pcix, pci, vlb, and isa anymore
[07:01:07] <jamesd2> lewellyn, i would settle for pci-x, pci-e and pci
[07:01:10] <lewellyn> it's also unfortunate i'm spacing on the company's name who did
[07:01:18] <jamesd2> and dual sockets...
[07:01:24] *** blizzow3 has quit IRC
[07:01:26] *** blizzow1 has joined #opensolaris
[07:01:29] <jamesd2> that less than $100 each.
[07:01:48] <lewellyn> heh. those boards aren't cheap
[07:02:12] <lewellyn> the "we have everything" boards were WATX and used every sqmm
[07:02:48] *** blizzow1 has quit IRC
[07:02:49] *** blizzow2 has joined #opensolaris
[07:04:07] <lewellyn> (and, of course, since no one *makes* wtx cases anymore, the intention was that you'd build your own or run it bare in a lab)
[07:04:33] <Nemesis> so...I am correct in saying there is no..NOOO..."ISA" support for osol right?
[07:05:12] <Nemesis> so there would be no use in trying an isa 10/100 card probably
[07:06:21] *** aghaster has quit IRC
[07:07:30] <lewellyn> there is no use, no.
[07:08:43] <Andys^> Nemesis: lol..wtf?
[07:08:52] <lewellyn> i think you'd have no luck unless you had a pcnet card
[07:08:54] *** aghaster_ has quit IRC
[07:09:04] <lewellyn> and even those are ill-supported atm, i think
[07:10:28] <lewellyn> aha. i found what i was misremembering. the isa pseudo-node was removed, not isa support altogether
[07:10:43] *** winstonw has joined #opensolaris
[07:12:40] *** winstonw has quit IRC
[07:16:51] *** winstonw has joined #opensolaris
[07:16:53] <Nemesis> osol dont like wireless nics either
[07:17:05] <Nemesis> is that up to date enough for joo all?
[07:17:19] <lewellyn> it does fine with wifi
[07:17:25] <lewellyn> as long as you aren't using broadcom
[07:17:34] <lewellyn> yell at broadcom about that
[07:17:39] <Nemesis> netgear
[07:17:53] <Nemesis> wpn111
[07:17:54] <lewellyn> netgear doesn't make the chips
[07:18:10] *** leito has quit IRC
[07:19:55] <Nemesis> http://forums.opensolaris.com/thread.jspa?threadID=1603
[07:20:26] <lewellyn> ah. you didn't say it was usb
[07:20:33] * lewellyn has no clue what usb support is like
[07:21:17] <lewellyn> ural exists...
[07:21:47] <lewellyn> and 1119 added arn
[07:21:48] *** easyE has joined #opensolaris
[07:21:49] <lewellyn> 119, even
[07:22:02] <lewellyn> arn apparently supports that nic, from a cursory search
[07:22:31] <lewellyn> oh no. misread
[07:22:57] *** biant92 has joined #opensolaris
[07:23:06] *** devians[w] has quit IRC
[07:23:19] *** devians[w] has joined #opensolaris
[07:23:20] <lewellyn> yeah. no atheros usb wifi support yet, it seems
[07:23:54] <lewellyn> oh. i'm wrong. again! :D
[07:24:31] <lewellyn> the wg111 should work with uath
[07:24:31] *** winstonw has quit IRC
[07:24:54] *** swift_ has joined #opensolaris
[07:24:59] *** devians[w] has quit IRC
[07:25:40] *** devians[w] has joined #opensolaris
[07:25:45] <lewellyn> Nemesis: as mentioned before, if you have a lot of exotic hardware you wish to see support for, you should either write your own driver, or convince someone to do it for you :)
[07:26:30] * lewellyn wishes pkgbuild-friendly folk would wake up/get home from the bar :P
[07:26:49] *** swift has quit IRC
[07:27:02] *** swift_ is now known as swift
[07:27:32] <Nemesis> i may consider learning "how to write drivers"
[07:28:38] *** jaywalk has quit IRC
[07:31:19] *** leito has joined #opensolaris
[07:34:49] *** smrt has quit IRC
[07:35:10] *** smrt has joined #opensolaris
[07:39:32] *** MrHeavy_ has joined #opensolaris
[07:40:06] <MrHeavy_> Anyone else having NFS server issues on snv_134?
[07:40:14] <MrHeavy_> Anytime I try to mount from one of my ESX hosts, mountd crashes
[07:42:33] *** winstonw has joined #opensolaris
[07:48:38] <MrHeavy_> Actually
[07:48:48] <MrHeavy_> It looks like it dumps core if there's no reverse DNS for the connecting host
[07:49:10] <lewellyn> Nemesis: there's a book :D
[07:49:18] <MrHeavy_> Once I added RDNS for my ESX hosts' VMkernel ports and restarted nfs/server it looks like it's working
[07:49:40] <lewellyn> MrHeavy_: it shouldn't dump core, but by default you need rdns, yes
[07:49:47] * lewellyn needs to test smrt real quick
[07:49:51] *** smrt was kicked by lewellyn (<3)
[07:49:56] *** smrt has joined #opensolaris
[07:50:02] <lewellyn> interesting.
[07:50:14] <MrHeavy_> I didn't specify the ACLs as hostnames, though
[07:50:23] <MrHeavy_> They were IP addresses, so it should have worked fine
[07:50:34] <palowoda> lewellyn: You should be arrested for child abuse.
[07:50:34] <MrHeavy_> Poking through the core dumps it looks like it might be a logging function to blame
[07:50:38] <lewellyn> well, it still shouldn't have dumped core :)
[07:50:56] <lewellyn> palowoda: so should alanc, then
[07:51:16] <lewellyn> MrHeavy_: if you have useful info, file a bug :D
[07:51:32] <lewellyn> "nfs dumps core" is a great thing to file before they finalize the next release
[07:51:39] <MrHeavy_> I'm checking now to see if it's reproducible
[07:51:45] <lewellyn> :)
[07:51:53] <MrHeavy_> I've got to fix my ESX servers first :(
[07:51:54] <lewellyn> my newest nfs server runs 130
[07:54:01] <palowoda> lewellyn : alanc has been arrested several times. It makes no difference now. :)
[07:55:23] *** derchris has quit IRC
[07:55:41] *** derchris has joined #opensolaris
[07:56:58] *** iFire has quit IRC
[07:58:09] *** Zauberpo1y has joined #opensolaris
[07:59:16] *** Zauberpony has quit IRC
[08:00:21] *** blizzow2 has quit IRC
[08:00:51] *** swift_ has joined #opensolaris
[08:01:11] *** Aria has quit IRC
[08:01:29] *** Bartman017 has joined #opensolaris
[08:01:39] *** Bartman017 has quit IRC
[08:02:32] *** swift has quit IRC
[08:02:44] *** swift_ is now known as swift
[08:07:47] *** Dagobert has joined #opensolaris
[08:08:24] *** Dagobert_ has joined #opensolaris
[08:08:24] *** Dagobert has quit IRC
[08:08:26] *** Dagobert_ is now known as Dagobert
[08:09:19] <jnss> i don't have the root password to the livecd
[08:09:22] *** easyE has quit IRC
[08:09:55] *** easyE has joined #opensolaris
[08:09:58] <jamesd2> jnss, root is a role you can't login as root...
[08:10:20] <jnss> i don't want to login
[08:10:25] <jnss> i just need to su
[08:10:34] <jamesd2> pfexec bash
[08:11:01] <jnss> thanks
[08:11:40] <jnss> also, would you happen to know how the hell one mounts a device? i have this thing at /dev/dsk/c6t0d0p0 and nothing i try will actually mount it
[08:11:58] <jnss> knowhing this would really help me out of  a jam
[08:11:58] <jamesd2> what type of device is it?
[08:12:02] <jnss> i
[08:12:11] <jnss> it goes into the laptops sdhc slot
[08:12:18] <jnss> a memory card, in other words
[08:12:18] <Thrae> jnss: Do you know what it's formatted as?
[08:12:27] <jnss> vfat
[08:12:36] <jnss> i can format it to something more suitable if needed
[08:12:43] <jamesd2> svcadm enable rmvolmgr
[08:12:52] <jnss> oh wow
[08:12:55] * jnss types it
[08:12:56] <jamesd2> restart it if doesn't show up
[08:13:18] <jamesd2> with restart == svcadm restart rmvolmgr
[08:13:31] <richlowe> root pw on the livecd is 'opensolaris', if I recall correctly
[08:13:34] <Thrae> jnss: Or just mount -F pcfs /dev/dsk/partition /mnt/mountplace
[08:14:57] <jnss> hmm thanks. i just need to fix something real quick before i can benefit from any of those commands
[08:15:13] *** Mousk4r has joined #opensolaris
[08:15:47] *** hsp has joined #opensolaris
[08:16:03] <richlowe> not sure the sdcard bits are on the livecd
[08:16:22] <richlowe> can't recall if you can still use pkg install from the live env, either.
[08:18:34] <jnss> have to reboot ;)
[08:19:00] <jnss> but thanks guys,
[08:20:52] *** Mousk4r has quit IRC
[08:21:21] *** PrickelPit has quit IRC
[08:22:27] <lewellyn> richlowe: you can't
[08:23:15] <richlowe> shame, that was pretty neat :)
[08:27:15] *** Mousk4r has joined #opensolaris
[08:27:29] *** jaywalk has joined #opensolaris
[08:27:39] <lewellyn> perhaps i'd have less to bitch about if the livecd were actually useful ;)
[08:28:52] <lewellyn> i've been spoiled by solaris, i suppose. where you can install different profiles from either text or gui mode
[08:29:21] <lewellyn> and where a network install doesn't require invalid xml
[08:29:28] <richlowe> file bugs.
[08:29:29] <richlowe> or RFEs
[08:30:30] <lewellyn> well, the ai issue is well-known. i think there's already a bug, even. and theoretically, it'd be really easy to make the xml valid. but the tools can't handle that yet
[08:30:54] *** PrickelPit has joined #opensolaris
[08:30:55] <lewellyn> i had that conversation already
[08:31:32] <lewellyn> probaly the only pkg(5) complaint i've had which was immediately "yeah. that sucks that it's not fixed, as it's entirely broken behavior and counterintuitive."
[08:31:41] <lewellyn> immediately replied to with, even
[08:31:47] * lewellyn is overly multitasking, sorry
[08:32:13] <richlowe> AI is the caiman folks, largely different group of people.
[08:32:16] <richlowe> (as far as "pkg(5) complaint")
[08:32:34] <lewellyn> ok, (tangentially) pkg(5) ;)
[08:32:53] <lewellyn> considering that AI needs an ips repo to be useful
[08:32:55] <richlowe> Well, given how much shit gets flung at the actual pkg(5) folks, it's best to correct such things regardless :)
[08:33:37] <lewellyn> yes, and i understand the reason for caiman's existence more than pkg(5)'s, so i probably should be fairer :)
[08:34:51] *** swift_ has joined #opensolaris
[08:34:58] <richlowe> Now there's an argument that's been had a thousand time.
[08:34:59] <richlowe> s
[08:35:29] <richlowe> lewellyn: in my experience, both groups of folks are pretty eager to fix stuff, but lacking in people to do it quickly (necessarily)
[08:37:32] *** swift has quit IRC
[08:37:35] *** swift__ has joined #opensolaris
[08:37:53] *** swift__ is now known as swift
[08:40:57] *** swift__ has joined #opensolaris
[08:41:04] *** swift_ has quit IRC
[08:42:19] <lewellyn> richlowe: yes, i realize. that doesn't make me not grumble about the lack of package compatibility between solaris and opensolaris ;)
[08:43:15] <lewellyn> i mean, if someone would make the packaging tools coexist somehow, that'd rock. but as pkg is still apparently "Unstable", that won't be me.
[08:43:21] <richlowe> one-way, SVR4 packages install on osol .
[08:43:29] <lewellyn> which pkg doesn't know about
[08:43:50] <lewellyn> so if i install, e.g. GVsomething, pkg will gladly clobber anything it installed
[08:44:31] <richlowe> I'm betting any bug against that contains "But then we'd have to read the contents file...", too. :\
[08:44:55] *** swift has quit IRC
[08:45:06] *** swift__ is now known as swift
[08:45:06] <lewellyn> imo, at a minimum, the packaging system should go "hey, this thing wants to step on the toes of a package that is on the system, but isn't managed by this packaging system!"
[08:45:34] <richlowe> agree.
[08:45:44] <lewellyn> in pkgadd's case, it'd be logical to prompt to continue. in pkg(1)'s case, it seems to puke in that type of scenario.
[08:46:17] <lewellyn> when i get a newer environment up that i can actually use IPS, i plan on giving some love to both pkg and pkg*
[08:46:44] <lewellyn> i have a workspace here with svr4 packaging changes collecting bitrot
[08:47:22] <lewellyn> i do have a tendency to poke at stuff that bugs me, along with complaining :)
[08:47:53] <lewellyn> i'd also love to put up a webrev of the svr4 stuff, but i've not tested it with anything newer than 129 or so...
[08:47:57] <richlowe> and on that note, I finally found what's causing double dependencies!
[08:48:01] <lewellyn> oh?
[08:48:26] <richlowe> finds dependency that only satisfies one variant, tries again with installed bits, finds same dependency only satisfying same variant.
[08:48:27] <richlowe> uses both.
[08:48:37] *** Yu\2 has joined #opensolaris
[08:50:22] *** |AbsyntH| has joined #opensolaris
[08:50:43] <richlowe> this is what tinkering as well as complaining gets you.  Utter bloody confusion. :)
[08:55:59] <lewellyn> i'm not sure if i've seen that particular issue yet
[08:57:07] <richlowe> lewellyn: build ON > 136, and you'll find it in a couple packages.
[08:57:24] <richlowe> completely non-fatal, since foo at 0 dot 134 and foo at 0 dot 136 are both satisfied by foo at 0 dot 136 :)
[09:00:17] *** vivid has quit IRC
[09:01:00] *** vivid has joined #opensolaris
[09:03:34] <lewellyn> richlowe: ah. i can't even get most of my boxes upgraded to 134, so... :)
[09:03:40] *** CodeWar has joined #opensolaris
[09:15:11] *** Odin- has quit IRC
[09:16:12] *** merzo has joined #opensolaris
[09:16:23] *** zazenrasta has joined #opensolaris
[09:19:47] *** xyyz has joined #opensolaris
[09:20:39] <xyyz> this question is a little different form what i've asked before, but i have 2 sata raid controllers (amcc 9650se) handling 4 disks each
[09:21:02] <xyyz> but each controller or external enclosure acts as a point of failure
[09:21:24] <xyyz> what i'd like to do, is run raidz on each of the 4 disks on each controller
[09:21:30] <xyyz> and then mirror the two under 1 pool
[09:21:33] <xyyz> is this possible?
[09:22:26] <echobinary> yes, it is
[09:22:43] *** zazenrasta has quit IRC
[09:22:53] <echobinary> you pretty much just said it -   you'd create a raidz out of the two banks of drives...
[09:23:00] <xyyz> okay... this is what i did
[09:23:03] <xyyz> can you tell me if it was right
[09:23:05] <richlowe> that stripes across the two raidzs
[09:23:10] <jamesd2> possibly.. but is it a good idea, NO!!!... just mirror the two controllers...  or create two raidz zdevs... each vdev with devices from each controller..
[09:23:28] <echobinary> I cannot - I am a noobled myself  - I just know that you should be able to do what you asked - I've seen it in the forums a few times
[09:23:31] *** biant92 has quit IRC
[09:23:37] <echobinary> nooblet*
[09:23:39] <xyyz> okay wait.... way above my understanding level
[09:23:44] <xyyz> trust me i'm a super noob too
[09:23:46] <xyyz> this is what i did
[09:24:42] <xyyz> zpool disk-pool raidz (da1) (da2) (da3) (da4) raidz (db1) (db2) (db3) (db4)
[09:25:07] <xyyz> that's what i wanna do actually... mirror the controllers...
[09:25:18] <jamesd2> zpool create poolname raidz2  drive1 drive5 drive2 drive6  raidz drive3 drive7 drive4 drive8 ... then one controller or 4 drives can die and no dataloss.
[09:25:19] <xyyz> sadly, that's what i thought i was doing
[09:25:35] <xyyz> so then my syntax is sound?
[09:25:50] <xyyz> well it worked at least... but then what i am doing is mirroring both controllers?
[09:25:59] <jamesd2> syntax is right, but if you loose a controller/enclosure, you have data loss
[09:26:06] <xyyz> so if one falls to shyte,,, no harm no foul
[09:26:44] <lewellyn> raidz2 of 4 drives? :P
[09:26:57] <xyyz> ack... yeah that's exactly what i want to avoid... initially i had raidz2'ed all 8 but then i started thinking that it was a crap design
[09:27:03] <lewellyn> what's wrong with a mirror of 2 raidz vdevs?
[09:27:20] <xyyz> well i want a redundance of two things
[09:27:30] <lewellyn> allows you to lose one controller + 1 disk
[09:27:37] <xyyz> 1. controller failure  and  2. external enclosure failure
[09:27:45] <jamesd2> lewellyn, yeah that could work, but layout of his was not the best since... one each drive box was in each enclosure failure
[09:27:48] <xyyz> this will have all my backups and my iSCSI stuff on it
[09:28:05] <lewellyn> i'm not really paying much attention
[09:28:17] <lewellyn> but why make 8 drives external in the first place
[09:28:31] * lewellyn noted the sata above
[09:28:58] <xyyz> well i got a good deal on the 3ware sidecarts with the 9650se cards
[09:29:07] <xyyz> they've been sitting here all this time... so you know :)
[09:29:11] <lewellyn> you can hotswap 8 sata in 2u
[09:29:21] <xyyz> well i can hotswap these drives too
[09:29:32] <lewellyn> getting a good deal on something doesn't mean it's right for the project at hand
[09:29:45] <lewellyn> you can always let them hang out till you actually need them for something :)
[09:29:48] *** lblume has joined #opensolaris
[09:29:51] <mui> pff
[09:29:53] <mui> moin
[09:30:07] <lewellyn> like lblume. we let him hang out in case we need him to moan at alanc :)
[09:30:13] <xyyz> it's not a production environment... it's a home environment
[09:30:17] <xyyz> more of a learning experience really
[09:30:22] <mui> do mirrors
[09:30:31] <lewellyn> xyyz: so use a standard pc case. most can handle 8 drives :)
[09:30:46] <xyyz> that would require an additional purchase now
[09:30:49] <lblume> You guys had a nice break last night because my server lost connectivity. It's over now.
[09:31:03] <lewellyn> ("designed to" is irrelevant, with a trip to the hardware store, a dremel, and some welding!)
[09:31:18] <xyyz> i just gotta make due with what i have
[09:32:07] <xyyz> the configuration i did, i thought i was mirroring each enclosure/enclosure.... no?
[09:32:11] <lewellyn> xyyz: if you have 2 enclosures, i'd say to stick 4 spindles in each, on a single controller, and mirror a raidz made from each enclosure.
[09:32:19] <lewellyn> i didn't see yours
[09:32:28] <lewellyn> and my gf is bitching at me to finish a project, so i should :)
[09:32:32] <lblume> And where's my favourite font-hacker? I want to moan about X, SDL apps are corrupting Compiz again.
[09:32:48] <xyyz> well much obliged for the input you provided....  :)
[09:32:56] *** CodeWar has quit IRC
[09:33:12] <xyyz> zpool create disk-pool raidz (da1) (da2) (da3) (da4) raidz (db1) (db2) (db3) (db4)
[09:33:18] <xyyz> there's my configuration just in case
[09:33:34] <mui> how do you mirror two raidz? creating two pools and mirroring dataset from those?
[09:33:50] <xyyz> totally ADDing right now, but does anyone familiar with In'n'Out know when they close on friday nights?
[09:34:01] <lblume> Uhoh..... When gmake says "Virtual memory exhausted", it's a bad sign, right?
[09:34:22] <lewellyn> lblume: sdl apps are corrupting compiz? you're using mednafen now? ;)
[09:34:38] <lewellyn> xyyz: same time as every night
[09:34:40] <lblume> ScummVM and DosBox.
[09:35:00] *** forquare-macbook has joined #opensolaris
[09:35:13] <lewellyn> xyyz: http://www.in-n-out.com/locations.asp
[09:35:21] <mui> hey btw
[09:35:34] <mui> I must say that basic support for osol worked very well for me couple days ago
[09:35:42] <mui> so it's not oracle coming afterall!
[09:36:00] <xyyz> thnanks.... so they do close at 130
[09:36:18] <xyyz> i'm gonna run a test and powerdown an enclosure
[09:36:21] <xyyz> i wanna see what happens
[09:36:37] <xyyz> okay this is weird...
[09:36:50] <xyyz> it says that the powered down enclosure/disks are still online
[09:37:00] <lewellyn> xyyz: run a scrub :)
[09:37:09] <lblume> You should cfgadm them off first
[09:37:27] <lewellyn> lblume: he's testing what happens in case of catastrophic failure
[09:37:36] <lblume> Oh, yes, sorry., I just realized
[09:37:48] <lblume> The OS will notice that at the first write
[09:38:18] <mui> xyyz: what is your failmode btw?
[09:38:21] <xyyz> cfgadm?  more stuff over my head.... i just wanted to see what happens in the event of a sudden failure
[09:38:28] <lblume> However, drives disappearing is not the kind of failure ZFS handles best (though it has improved)
[09:38:49] <mui> afaik failmode=continue is not that much continue yet?
[09:38:53] <xyyz> umm okay... the scrub isn't returning anything so far... does it take a while?
[09:39:16] <xyyz> well if it's mirrored right.... there's not too much to handle no?
[09:39:16] <lewellyn> xyyz: zpool status
[09:39:23] *** leito has quit IRC
[09:39:25] <|woody|> scrub doesn't return anything
[09:39:27] <mui> there's bit lenghty timeout for disks that vanish
[09:39:28] <lewellyn> xyyz: zfs is learning :)
[09:39:33] <lblume> For your test, you need to do some writes to the disappeared disk before it notices. You also will need cfgadm to put the disks back online properly without reboot.
[09:39:35] <xyyz> yeah it's not doing anything....  ahhh learning... :)
[09:39:55] <lewellyn> lblume: i thought a scrub would also make it notice
[09:40:03] <|woody|> it does lewellyn
[09:40:15] <xyyz> haha sweet... okay it recognized that the disks are no longer there
[09:40:32] <xyyz> ahhh
[09:40:40] <lblume> It should, but a write is more typical. What happens then (i/o hanging) is good to learn :-)
[09:40:40] <mui> zfs is quite amazing handling disasters
[09:40:42] <xyyz> but it's saying that the pool is unavailable
[09:40:51] <mui> once I had like 80 disks in jbods
[09:40:58] <mui> one shelf disappeared, 16 disks
[09:41:02] <xyyz> 80 disks?
[09:41:03] <mui> another shelf disappeared, other 16disks
[09:41:09] <lblume> mui: I'd say zfs is amazing not losing your data. But so far, I haven't been impressed at its handling of disasters.
[09:41:16] <mui> and data that was spread over other shelves could be read
[09:41:37] <xyyz> hey... i was impressed in that video when they took a sledgehammer to the disks
[09:41:48] <xyyz> seemed to recover pretty well :)
[09:41:50] <mui> I could pinpoint exactly what data was still accessible and what was not
[09:42:13] <mui> after loosing 32 disks in a pool
[09:43:16] <mui> if you think what would happen with ordinary raid-setups
[09:43:18] <mui> :P
[09:43:34] <|woody|> which reminds me I still need to fine tune mpxio because evertime the paths switch at least one disk is lost in the pool because even recoverable scsi error count for write error. which is kind of bad
[09:44:12] <lewellyn> xyyz: if the pool is unavailable, you did it wrong :)
[09:44:17] <lewellyn> pastebin the zpool status?
[09:44:42] <xyyz> me doing it wrong.... yeah that happens a lot
[09:44:51] <xyyz> it's still sitting on the scrub
[09:44:59] <lewellyn> pastebin it and people can help :)
[09:45:13] <xyyz> you got it...
[09:45:16] <xyyz>   pool: enetwerks-zfs  state: UNAVAIL status: One or more devices are faulted in response to IO failures. action: Make sure the affected devices are connected, then run 'zpool clear'.    see: http://www.sun.com/msg/ZFS-8000-HC  scrub: scrub completed after 0h0m with 0 errors on Sun Apr 11 00:42:40 2010 config:
[09:45:22] <lewellyn> we may get cranky, but we do like to problemsolve
[09:45:25] <lewellyn> smrt: explain pastebin
[09:45:26] <smrt>  Anything larger than 2-3 lines of output should be pastebin http://opensolaris.pastebin.com
[09:45:32] <lewellyn> do it there :)
[09:45:34] <lewellyn> the whole output
[09:45:39] <xyyz> 	NAME        STATE     READ WRITE CKSUM 	enetwerks-zfs  UNAVAIL      0     0     0  insufficient replicas 	  raidz1-0  UNAVAIL      0     0     0  insufficient replicas 	    c6t0d0  UNAVAIL      0     0     0  experienced I/O failures 	    c6t1d0  UNAVAIL      0     0     0  experienced I/O failures 	    c6t2d0  UNAVAIL      0     0     0  experienced I/O failures 	    c6t3d0  UNAVAIL      0     0     0  experienced I/O failures
[09:45:43] <lewellyn> not here :P
[09:45:47] <lewellyn> smrt: explain spamming
[09:45:48] <smrt>  Please *always* use a pastebin for any paste, even 3 lines.  It causes the following problems:  Your paste gets interrupted by IRC chatter.  IRC chatter gets interrupted by your paste.  And after a minute or two, the paste scrolls off and is difficult to reference.  Please be a considerate IRC user and use a pastebin for pastes.
[09:45:50] <mui> well thats one line
[09:45:50] <xyyz> oh there
[09:46:02] <xyyz> sorry... still learning :)
[09:46:04] <mui> lewellyn: you just need bigger screen (:
[09:46:12] <lewellyn> that's why smrt tries to be polite about it :)
[09:46:18] <lblume> What would cause gmake to complain that virtual memory is exhausted? It's not even swapping....
[09:46:35] <xyyz> okay... pasted
[09:46:46] <lewellyn> mui: 232x55 atm
[09:46:54] <mui> inches?
[09:46:56] <mui> :P
[09:47:07] <lewellyn> colsxrows
[09:47:11] <mui> ah
[09:47:28] <lewellyn> i should expand the window...
[09:47:31] <xyyz> and the scrub command still hasn't finished running....
[09:48:15] <xyyz> and... i powered the enclosure back on... just in case i didn't mention that
[09:48:25] <lewellyn> $ echo "${COLUMNS}x${LINES}"
[09:48:26] <lewellyn> 278x72
[09:48:35] <lewellyn> mui: high enough for you?
[09:48:56] <lewellyn> xyyz: scrubs take a while. and zfs won't know about it till you tell the os it's back
[09:49:07] <xyyz> ummm how do i do that?
[09:49:09] <lewellyn> xyyz: also, you haven't linked the pastebin paste to us yet :)
[09:49:15] <xyyz> cfgadm?
[09:49:19] <lewellyn> yup
[09:49:20] <taemun> lewellyn: what screen res was that on lol
[09:49:23] <xyyz> oh... i thought it just magically appeared
[09:49:24] <lblume> xyyz: And you absolutely need to use cfgadm to do it properly.
[09:49:29] <taemun> 1920x1200 default putty font is 240x75
[09:49:30] <lewellyn> taemun: 1440x900
[09:49:35] <xyyz> http://opensolaris.pastebin.com/Z69Q2xE9
[09:49:36] <taemun> ah lol\
[09:49:36] <lewellyn> menlo 8pt
[09:49:40] <taemun> must have tiny font
[09:49:40] <taemun> yeah
[09:49:48] <xyyz> i guess i outta start reading up on cfgadm 'cause i never heard of it
[09:50:05] <mui> lewellyn: more than I have!
[09:50:31] <lewellyn> mui: well then!
[09:50:35] <mui> !
[09:50:44] <lblume> Sorry, my connection is rather paingfully slow atm, can't follow much :-(
[09:50:47] * lewellyn wishes Terminator.app wasn't in java so menlo8 wasn't quite as suck
[09:50:52] * mui sits middle of freezing datacenter floor
[09:50:56] <lewellyn> lblume: feed it coffee
[09:50:58] <mui> pillow would be cool
[09:51:05] <lewellyn> mui: late night? :(
[09:51:23] *** leito has joined #opensolaris
[09:51:28] <mui> just watching ppl over
[09:51:34] <xyyz> alright ladies and gents... i'm gonna go get some food and hope that the scrubbing finishes by the time i get back
[09:51:57] <lewellyn> oh hey. i forgot. i have office on here!
[09:52:15] <lewellyn> i have consolas! <3 consolas 10 is the best tiny scalable font, imo
[09:52:55] *** ma3x_ has joined #opensolaris
[09:52:56] <lewellyn> i lost some rows, but so worth it
[09:53:23] <lewellyn> menlo's too wide for my tastes, anyhow
[09:55:01] <lblume> xyyz: Since the zpool is unavailable, the scrub won't finish.
[09:55:24] <lewellyn> that makes sense
[09:55:28] <lewellyn> it's slow anyhow :)
[09:55:28] *** forquare-macbook has quit IRC
[09:55:57] <lblume> AFAICT, it's not a mirror you did, but a stripe. No redundacny.
[09:56:24] <lblume> So removing half of it just breaks it down completely.
[09:56:44] *** ma3x has quit IRC
[09:57:01] *** gebi has joined #opensolaris
[09:57:05] *** ma3x_ is now known as ma3x
[09:58:48] <lewellyn> oh yeah. i didn't hit my browser again yet
[09:58:51] * lewellyn looks at the paste
[09:59:04] <lewellyn> yeah. that looks like a stripe
[09:59:18] <lewellyn> cfgadm on and you should be fine
[09:59:33] <mui> urm
[09:59:44] * lewellyn urms at mui
[09:59:50] <taemun> uurrrmm
[10:00:01] <mui> can you mirror top vdevs instead of stripe without creating two pools and mirroring files from them?
[10:00:10] <taemun> yes?
[10:00:21] <taemun> you could mirror a pair of raidz vdev's
[10:00:25] <taemun> not sure why you would though lol
[10:00:26] <mui> oh
[10:00:45] <mui> like zpool create foo mirror raidz2 bar raidz2 foo?
[10:00:58] <taemun> yeah?
[10:01:00] <mui> oh
[10:01:07] <mui> didnt know that
[10:01:07] <lewellyn> taemun: in xyyz's case, he can do a raidz per enclosure, so he can lose an enclosure + 1 disk :)
[10:01:23] <taemun> i'd just mirror across each enclosure
[10:01:30] <mui> me too
[10:01:39] <lewellyn> then he's screwed if he loses a disk in the still-online enclosure
[10:01:46] <taemun> lol
[10:01:49] <lewellyn> i've met murphy enough times myself :P
[10:02:06] <lewellyn> if i'm giving up that much capacity already, may as well add in more safety
[10:02:25] <lblume> mirroreed RAID-Z3 is the way to go.
[10:02:31] <taemun> fuck it
[10:02:36] <taemun> just have 48 drives in a single mirror
[10:02:43] *** h3x0r has quit IRC
[10:02:48] <taemun> at least you can't have too many bytes to backup
[10:03:14] <lewellyn> lblume: can you raidz3 4 disks? i've not tried that yet ;)
[10:03:20] <taemun> yes, you can
[10:03:23] <taemun> its technically stupid
[10:03:42] <taemun> lewellyn: still, wouldn't you be better with raidz ( mirror (4 drives)) rather than mirror (raidz (4 drives))
[10:03:49] <taemun> I guess I'm arguing semantics here
[10:05:07] <echobinary> are you anti-semantic?
[10:05:34] <lewellyn> taemun: i'd expect zfs to freak less if half a mirror died than half of two raidzs
[10:06:00] <taemun> the raidz vdev wouldn't even know that it'd been compromised
[10:06:03] <lewellyn> i mean, that's how a mirror works ;)
[10:06:04] <taemun> just the mirrors
[10:06:09] <richlowe> didn't relling graph all this stuff out, at one point?
[10:06:18] <lewellyn> i think so
[10:06:29] <lewellyn> but he's not here when these convos come up
[10:06:32] <taemun> richlowe: not raidz(mirror(4)) vs mirror(raidz(4))
[10:06:38] <lewellyn> and his posts are hard to grok for me :)
[10:06:45] <lblume> Great, I'm just hitting a gmake bug.....
[10:06:46] <lewellyn> taemun: yes
[10:06:54] <lewellyn> lblume: *shock* *awe*
[10:07:36] <lblume> Indeed!
[10:07:38] <lewellyn> relling's graphs are what made me decide on a single raidz2 for six drives
[10:07:57] <taemun> http://blog.richardelling.com/
[10:07:59] <taemun> ding
[10:12:40] <mui> are we there yet are we there yet
[10:14:27] <taemun> quit screwing around in the back of the car
[10:14:28] *** seanmcg has quit IRC
[10:15:40] *** seanmcg has joined #opensolaris
[10:17:18] <lblume> Why can't we get some scandisk-like look at the disk when scrub runs? So we could know where we are yet?
[10:17:43] <mui> there is percentage?
[10:18:13] <lblume> It would be so fascinating watching those little bits being picked one by one, looked at, evaluated, weighted, then put back in place.
[10:18:48] <mui> you seriously need coffee
[10:19:21] <lblume> I only had one.
[10:19:38] <lblume> And no more than 2 glasses of wine.
[10:19:53] <lblume> Or 3.
[10:20:44] <mui> wine doesnt count
[10:23:10] *** swift has quit IRC
[10:24:22] *** sndcrb has joined #opensolaris
[10:24:31] * lewellyn installs QQ in lblume's wine
[10:25:12] *** Yu\2 has quit IRC
[10:26:26] *** dalibor has joined #opensolaris
[10:27:09] <lblume> Chinese wine! It'll support QQ!
[10:27:25] <lewellyn> i think it may actually run soon
[10:27:40] <lblume> With video support?
[10:27:44] <lewellyn> probably not
[10:27:49] <lewellyn> but logging in would be a start!
[10:28:18] <lewellyn> looks like there's a patch that may make 1.1.43 which may fix the issue
[10:28:34] <lewellyn> which will save me from having to get a spare tuit
[10:44:36] *** ubuntujenkins3 has joined #opensolaris
[10:49:17] *** bengtf has joined #opensolaris
[10:49:29] <zzpat> I've got a boomerang tuit, it keeps coming back
[10:50:05] *** gebi has quit IRC
[10:53:10] <mui>  /W 125
[10:55:06] *** bengtf has quit IRC
[11:04:01] <lewellyn> only... :P
[11:04:06] <lewellyn> mui: irc moar!
[11:05:10] <mui> :P
[11:05:41] <trochej> http://tinyurl.com/yauo5op
[11:07:07] *** merzo has quit IRC
[11:09:13] *** master_of_master has quit IRC
[11:09:43] *** PrickelPit has quit IRC
[11:10:17] *** nitrile has quit IRC
[11:10:24] *** jollyd has joined #opensolaris
[11:10:41] *** master_of_master has joined #opensolaris
[11:10:50] *** plat- has quit IRC
[11:11:00] *** nikolam has quit IRC
[11:11:29] *** pjd_ has quit IRC
[11:11:42] *** pjd- has joined #opensolaris
[11:11:48] *** dalibor has quit IRC
[11:12:15] *** jollyd has left #opensolaris
[11:14:07] *** merzo has joined #opensolaris
[11:15:42] <dizko> sigh...we really need faster/distributed ips repos
[11:15:47] <dizko> incredibly slow...
[11:16:15] <lewellyn> or an official way to set up third-party mirrors
[11:16:23] <lewellyn> (including metadata)
[11:17:22] <dizko> i think latency seems to be a big part of it, not very latency tollerant protocol
[11:17:27] <dizko> and...im in japan
[11:18:39] <lewellyn> it's http
[11:18:58] <dizko> doesnt it grab things by the file though vs downloading a package stream?
[11:19:08] <dizko> lots of little requests, etc
[11:19:23] <lewellyn> http pipelining should take care of that
[11:19:23] <lblume> IIRC, it does pipelining, so it doesn't matter.
[11:19:55] <lewellyn> istr that the svr4 stuff "supports" pipelining too
[11:20:13] <dizko> i tend to get fairly limit transfer speeds to other continents unless i can split the transfer up into parallel connections
[11:21:04] <dizko> 10MB/s if i can do ~50 connections
[11:21:44] *** SouLakos has joined #opensolaris
[11:24:46] *** plat- has joined #opensolaris
[11:26:15] *** twisti has joined #opensolaris
[11:31:30] *** RoyK^ has joined #opensolaris
[11:36:53] *** SouLakos has left #opensolaris
[11:43:54] *** PrickelPit has joined #opensolaris
[11:45:34] *** milek_ has joined #opensolaris
[11:46:18] *** eleusis has quit IRC
[11:47:10] *** eleusis has joined #opensolaris
[11:47:10] *** eleusis has joined #opensolaris
[11:49:20] *** Thrae has quit IRC
[11:49:30] *** vivid has quit IRC
[11:51:45] *** tosh has joined #opensolaris
[11:52:57] *** Thrae has joined #opensolaris
[12:00:55] *** fedreric has joined #opensolaris
[12:05:00] *** milek_ has quit IRC
[12:11:41] *** dalibor has joined #opensolaris
[12:15:58] *** WormDrink has joined #opensolaris
[12:16:39] *** RoyK^ has quit IRC
[12:17:24] *** Mdx4 has joined #opensolaris
[12:18:17] *** ewdafa has joined #opensolaris
[12:18:36] *** phil has joined #opensolaris
[12:21:58] *** PrickelPit has quit IRC
[12:23:13] *** eleusis has quit IRC
[12:23:28] *** |AbsyntH| has quit IRC
[12:24:38] *** eleusis has joined #opensolaris
[12:25:00] *** PrickelPit has joined #opensolaris
[12:26:26] *** milek_ has joined #opensolaris
[12:30:16] *** nitrile has joined #opensolaris
[12:38:53] *** hsp has quit IRC
[12:43:06] *** snuff-home has quit IRC
[12:45:38] <lewellyn> It's not just you! http://defect.opensolaris.org looks down from here.
[12:45:41] * lewellyn sighs
[12:48:04] <CosmicDJ> I bet they're migrating from *solaris to oracle enterprise linux ;)
[12:48:42] *** snuff-home has joined #opensolaris
[12:49:58] *** Beket has joined #opensolaris
[12:51:00] <lewellyn> join in the discussion at osol-discuss
[12:51:41] <wdp_> CosmicDJ, nah.
[12:51:45] <wdp_> they use real servers.
[12:51:52] <wdp_> windows 2000 mysql edition
[12:51:53] <lewellyn> aix, baby!
[12:51:56] * wdp_ hides
[12:52:31] <lblume> lewellyn: It's a good discussion now?
[12:54:25] <dizko> lewellyn: any suggestions on a strategy for accomplishing something like heirarchical storage management using zfs?
[12:54:33] <lewellyn> lblume: no clue
[12:55:07] <CosmicDJ> dizko: what's wrong with sam/qfs?
[12:55:15] <dizko> yea, considering that
[12:56:26] <dizko> i want to be able to turn off this external disk enclosure and have it be reasonably transparent
[12:56:40] <dizko> considering just doing something dirty like symlinks and a second zpool
[12:59:23] *** tosh has quit IRC
[13:02:54] *** tosh has joined #opensolaris
[13:04:25] *** mrp has joined #opensolaris
[13:04:30] <mrp> how do i mount a usb drive?
[13:04:49] <dizko> rmmount
[13:05:13] <mrp> dizko: how do i find the device?
[13:05:30] <dizko> you read the manpage really fast
[13:05:36] <lblume> Right click in the gnome menu is the easiest....
[13:05:46] <mrp> dizko: i read it :P
[13:06:01] <mrp> i dont know the vol name
[13:06:01] <mrp> :o
[13:06:02] <dizko> rmmount -l doesnt list it?
[13:06:16] <mrp> nope
[13:06:31] <dizko> does it have stuff on it?  are you going to use it outside of solaris?
[13:06:54] <mrp> dizko: its fat32
[13:07:16] <mrp> dizko: yes it is going to be used in ubuntu
[13:07:17] *** jollyd has joined #opensolaris
[13:07:34] *** jollyd has left #opensolaris
[13:08:14] <dizko> does it show up if you type rmformat -l ?
[13:09:00] <mrp> yes
[13:09:45] <dizko> can you run fdisk on the device ?
[13:11:12] <mrp> nathan@zoot:~$ fdisk /dev/rdsk/c2t0d0
[13:11:12] <mrp> fdisk: Cannot stat device /dev/rdsk/c2t0d0.
[13:11:15] <mrp> :s
[13:11:41] <dizko> that's really the device rmformat -l showed?
[13:11:55] <dizko> not like cXtXdXpX ?
[13:12:00] <mrp> dizko: woops forgot the p0
[13:12:03] <mrp> yes fdisk works
[13:12:06] <dizko> and the reason it cant open it is because youre not root
[13:12:19] <mrp> its actually ntfs :o
[13:12:23] <dizko> there you go
[13:12:37] <mrp> not posible in osol?
[13:12:43] <mrp> nathan@zoot:~$ pfexec fdisk /dev/rdsk/c2t0d0p0
[13:12:45] <mrp> worked ;)
[13:12:47] <dizko> i think theres some read only shit but i wouldnt bother
[13:12:53] <mrp> ok bugger
[13:12:56] <dizko> yea
[13:12:57] <mrp> thanks dizko :)
[13:13:00] <dizko> np
[13:13:13] <dizko> i run virtualbox with windows in it when i need to read an ntfs device
[13:13:21] <dizko> and attach the usb to vbox guest
[13:13:52] <lewellyn> smrt: explain ntfs
[13:13:54] <smrt>  Read/write is supported via FUSE and NTFS-3G. FUSE is still beta, but http://forums.opensolaris.com/thread.jspa?threadID=513 has an example for installing NTFS-3G. SFEntfs-3g is also available in SFE. You will need to install SUNWfusefs/SUNWlibfuse first, regardless. (See also: spec-files-extra)
[13:14:09] <dizko> does that work well?
[13:14:20] <lewellyn> people use it *shrug*
[13:14:54] <dizko> every time i touch SFE i regret it
[13:15:04] <dizko> probably because it was always related to encumbered shit though
[13:15:27] *** mavhc has quit IRC
[13:15:33] *** mavhc has joined #opensolaris
[13:16:51] <dizko> after a couple hours this pkg image-update downloaded everything and said completed, but never returned
[13:17:41] <dizko> if i were to kill it and run it again, does it cache the packages it downloaded ?
[13:17:57] <lewellyn> i think that's what it's doing: cleaning up the cache :P
[13:18:54] <dizko> truss isnt outputting anything when i attach to it
[13:19:22] <dizko> i was playing with a zpool on some new drives, i pissed off zpool and it wouldnt return or let me destroy this zpool
[13:19:53] <dizko> so i wondered if maybe it went to run something related to zpool or beadm and just hung up on whatever is going on there
[13:20:01] <dizko> was gonna reboot and try again
[13:21:43] <dizko> yea beadm list doesnt return....this box is...displeased
[13:22:29] <lewellyn> too much porn
[13:23:26] *** idonotknow_ has joined #opensolaris
[13:24:36] *** milek_ has quit IRC
[13:25:54] <phil> it's been while since a build release - anyone know when b135 will arrive?
[13:26:12] <mui> after they've fixed 134 so they can make a release
[13:26:16] <mui> :P
[13:26:28] <mui> though I'd hope release would get some of 135 fixes
[13:27:15] <phil> so it is frozen until 2010.3 release is done?
[13:27:53] <mui> think so yes
[13:27:59] <phil> thanks
[13:28:05] *** rmorse has joined #opensolaris
[13:29:11] <lewellyn> mui: stuff gets backported
[13:29:32] <lewellyn> phil: you can still build from source. development marches on, but there aren't binary blobs for the moment
[13:29:54] <lewellyn> and i seriously hope they're doing an RC cycle before release
[13:30:38] <mui> which is great
[13:31:00] <mui> it's just I hope that failmode=continue acts like failmode=wait fix gets backported
[13:31:35] <mui> it can solve great deal of annoyances I have atm with zfs
[13:33:09] <lewellyn> that's why i hope there's an rc.
[13:33:15] <lewellyn> so people can file bugs like that :)
[13:33:28] <mui> but 2010.03 could be great os for storage at least if everything goes as expected
[13:33:37] <mui> dedup gets fixed etc
[13:33:50] <lewellyn> my concern is that pkg is still too broken in 134
[13:34:11] <mui> some pkgs are broken or pkg itself is?
[13:34:42] <lewellyn> pkg itself
[13:34:47] <mui> :/
[13:35:03] <lewellyn> d.o.o is still down else i'd link a couple of the ones that i know of
[13:35:04] <lblume> Can it be unbroken?
[13:35:20] <lewellyn> lblume: well, until recently image-update tended to work :P
[13:35:46] <lewellyn> i have machines stuck on 111b-133 with various inabilities to upgrade to 134
[13:36:02] <dizko> disco@void:~# pfexec pkg image-update
[13:36:03] <dizko> Creating Plan
[13:36:05] * dizko crosses fingers
[13:36:12] <lewellyn> no --be-name?
[13:36:29] <dizko> does it matter?
[13:36:30] *** RoyK^ has joined #opensolaris
[13:36:30] <mui> I've been running 134 fine
[13:36:54] <lewellyn> dizko: opensolaris-7 is far less useful than 129, in my book :)
[13:36:59] <dizko> gotcha
[13:37:14] <lblume> lewellyn: Really? It's been broken so often since the first release, but you didn't try it :-)
[13:37:32] <lewellyn> well, i stayed on sxce as long as they'd let me, so of course i hadn't ;)
[13:37:57] <dizko> probably ill end up having to re-install this box from a newer dev image anyway
[13:38:09] <tsoome> well, i cant really complain, but all i do with pkg is like pkg-image-update
[13:38:12] <dizko> this file-server is running 111a and has never b een updated
[13:38:26] <lewellyn> tsoome: that's precisely what my complaint is about :)
[13:38:32] <lblume> Can anybody kick me for forgetting that -R$ORIGIN should be single-quoted?
[13:38:40] <lewellyn> why?
[13:38:41] <tsoome> and i havent met any too bad issues atm
[13:38:58] <mui> hm
[13:39:01] <lewellyn> tsoome: i've had a hell of a time getting to 134 :(
[13:39:03] <mui> I've had to fix pkg only once
[13:39:21] <lewellyn> i actually was trying to hit defect earlier to try the workaround they gave
[13:39:30] <lewellyn> (for one system)
[13:39:49] <lblume> lewellyn: To avoid it being interpreted by the shell, of course :-)
[13:40:04] <lewellyn> why would it be?
[13:40:28] <tsoome> but i dont do too fancy stuff with it, i let it manage be, maybe only give bename... maybe im just lucky, tho;)
[13:40:43] <lblume> With double-quotes, like I did, it is :-P
[13:40:54] <lewellyn> tsoome: i think i do it on more systems than you do :)
[13:41:05] <tsoome> probably that too, yes
[13:41:10] <lewellyn> lblume: that doesn't anwer my question :P
[13:41:21] *** mavhc has quit IRC
[13:41:54] <lblume> Well, it's certainly not desirable to have it interp[reted by the shell, isn't it? :-)
[13:42:06] <lewellyn> i'm still unclear why it'd be interpreted
[13:43:01] *** mavhc has joined #opensolaris
[13:43:18] <dizko> $ STUFF=meh; echo $STUFF; echo "$STUFF"; echo '$STUFF'
[13:43:18] <dizko> meh
[13:43:18] <dizko> meh
[13:43:19] <dizko> $STUFF
[13:43:22] <lblume> The shell replaces $ORIGIN by an empty string, since it's not defined.
[13:43:23] <dizko> i think that's what he's saying...
[13:44:30] <lewellyn> lblume: oh. you meant the literal string
[13:44:33] <lewellyn> \$
[13:44:50] *** taemun has quit IRC
[13:45:02] <lewellyn> or export ORIGIN=\$ORIGIN in .profile :D
[13:45:08] <lblume> I know, but I've just wasted a lot of time because I forgot that in a build script :-)
[13:45:51] * lewellyn wonders why the svr4 package database is still a text file :P
[13:46:17] <lblume> Because they tried to replace it by a DB already, and it sucked more.
[13:47:24] <lblume> I doubt they will try to change anything again in sysv packages.....
[13:47:29] <lewellyn> so they're only allowed one failure? :P
[13:47:31] *** nitrile has quit IRC
[13:47:44] <lewellyn> sysv is still needed for solaris's legendary compatibility
[13:48:25] <lblume> Yup, and it seems the conclusion was, "hey it didn't work, so let's start over, forget everything we know about packages, ignore everything others knooooooooooow, and call it IPS"
[13:49:14] <tsoome> its text file because thats simplest way to process it. they did attempt to convert it to binary database, but that was recalled fast...
[13:49:19] <lblume> Weird, I'm getting repeated keys.....
[13:49:30] <dizko> ips must be smart though....its still creating my plan!
[13:49:47] <dizko> to be fair this box is an atom330
[13:50:03] *** pjama has joined #opensolaris
[13:50:20] <lewellyn> lblume: that was discussed a few hours ago. see my thoughts above ;)
[13:50:21] <RoyK^> anyone here using vbox on osol?
[13:50:37] <dizko> i am
[13:50:55] <RoyK^> dizko: which version? vbox and osol?
[13:50:59] *** merzo has quit IRC
[13:51:09] <lewellyn> tsoome: yes. but the text file is horribly inefficient. and that doesn't mean that the db approach is fatally flawed. just THAT approach :)
[13:51:15] <lblume> lewellyn: I'd rather nott scroll back. I'm sure I'll get more depressed.
[13:51:29] <tsoome> agree. :)
[13:51:39] <lewellyn> lblume: basically, ips and svr4 need to be aware of each other.
[13:51:40] <dizko> running osol 134, vbox 3.1.2 until recently, now 3.1.6
[13:51:48] <RoyK^> k
[13:51:50] <dizko> 3.1.4 had some pretty broken usb stuff
[13:51:55] <Dagobert> Anyone knows why http://defect.opensolaris.org/ is down?
[13:51:56] <lblume> lewellyn: I probably still have a beta of S10 with the DB approach somewhere, want to try it?
[13:51:57] <lewellyn> 3.1.4 was pretty broken
[13:52:07] <lewellyn> Dagobert: i don't. but i mentioned it a while ago
[13:52:20] <lewellyn> i've not opened my email to see if it's been brought up on website-discuss yet
[13:52:21] <RoyK^> I've been running vbox on this 111 box for some time and it crashes repeatedly, hangs several times a day
[13:52:24] <RoyK^> same with 316
[13:52:35] <Dagobert> lewellyn: missed that. Strange. Just wanted to see if there is some progress on a specific bug.
[13:52:36] <lewellyn> lblume: i didn't say i wanted *that* implementation
[13:52:38] <RoyK^> haven't tried to upgrade osol yet, though
[13:52:47] <lewellyn> Dagobert: and i want a workaround from a specific bug
[13:52:49] <lblume> lewellyn: Yes, that would be a great idea, but I so not see it coming. I think that when all the IPS issues and lacking features are fixed, maybe they could consider that.
[13:52:59] <lewellyn> though i'm realizing that it's probably in my email if i bother opening it
[13:53:31] <lewellyn> lblume: there's a community. some have skills with a compiler. surely someone is capable of tackling it :)
[13:53:54] <dizko> royk: i think running dev is probably the right way to go with osol at this stage
[13:54:35] <dizko> you can download iso's of dev from genunix.org if youre not able to get 111 to update
[13:54:44] <lblume> lewellyn: Tackling sysv packages? I doubt it.
[13:55:05] <lewellyn> lblume: you'd be surprised... :)
[13:55:10] *** cybernd has joined #opensolaris
[13:55:27] <RoyK^> lblume: out of interest - how do you use vbox? server stuff or just test VMs?
[13:56:22] <lblume> Me? Why me? Just test VM and for some windows apps.
[13:56:22] *** merzo has joined #opensolaris
[13:56:31] <RoyK^> k
[13:57:14] <dizko> yay my plan is created and it appears to be loading from cash
[13:57:26] <dizko> if this completes and works i will be very pleased
[13:57:33] <dizko> cache
[13:58:15] <dizko> had to work this morning, getting very tired =(
[13:58:26] <dizko> first nice day we've had all year too
[14:01:01] <RoyK^> short question - I think I read somewhere about a bug or defect that made osol unbootable after upgrading from 111 to 134 or so and required some manual fix. I can't seem to find this... The reason I ask, is I have a server some 50km away from here and I'd like to upgrade it without taking yet another day off work to get there (during daytime). There is no remote console access to this location. Does anyone remember this bug?
[14:01:46] <dizko> im updating a 111a system to 134 right at this moment, so ill have the same question in about 10 minutes if so =(
[14:02:32] <lblume> RoyK^: The fancy boot bug?
[14:03:26] <dizko> haha pkg crashed!
[14:03:44] <RoyK^> lblume: huh?
[14:03:51] * dizko writes 134 usb image to a flashdrive and quits wasting time
[14:04:34] <lblume> RoyK^: In b134, the system can crash on boot if grub's graphical boot is enabled.
[14:04:50] <dizko> yea my macbook does that
[14:04:57] *** twisti has quit IRC
[14:05:24] <RoyK^> lblume: how can I turn off that?
[14:05:31] <dizko> edit grub
[14:05:32] * RoyK^ doesn't need much graphics on his server
[14:06:05] * lewellyn sends condolences to poland
[14:06:17] <lblume> Remove the foreground/background lines in menu.lst, and the console-graphics from the kernel line
[14:06:28] <lewellyn> dizko: did you update pkg first like smrt says?
[14:06:32] <RoyK^> erm - where's menu.lst?
[14:06:41] <lblume> bootadm list-menu will tell you
[14:06:51] <dizko> no i didnt
[14:07:01] <RoyK^> lblume: thanks
[14:07:48] <lewellyn> smrt: explain dev repository
[14:07:49] *** twisti has joined #opensolaris
[14:07:49] <smrt>  To switch to the /dev repository (OpenSolaris development releases, see the IPS version in /topic): pfexec pkg set-publisher -O http://pkg.opensolaris.org/dev opensolaris.org && pfexec pkg install SUNWipkg && pfexec pkg image-update -v # Then wait. You *must* read the release notes before starting out or you *will* be unhappy. See also: ips mirrors
[14:07:55] <lewellyn> it SHOULD tell you to do so anyhow
[14:08:08] <lewellyn> file a bug when d.o.o comes back if it didn't :D
[14:08:19] <dizko> thanks
[14:09:02] <lewellyn> it'd be not unheard of to see an upgrade available just before a new release :)
[14:09:25] *** taemun has joined #opensolaris
[14:09:28] <RoyK^> lblume: so something like this? http://pastebin.com/BYF5G3ek
[14:09:59] <lewellyn> imo, pkg should be smart enough to say "A newer version of the package manager is available and required for the requested action. Download it now? ("No" will abort the action.) [Y/n]"
[14:10:13] <lewellyn> i'd file a bug saying so, but... yeah. :P
[14:10:23] <dizko> # pfexec pkg install SUNWipkg
[14:10:25] <dizko> No updates available for this image.
[14:10:35] <lewellyn> you switched publishers?
[14:10:45] <lblume> RoyK^: Yes, and I forgot splashimage, that, too.
[14:10:50] <dizko> yeap
[14:11:01] <RoyK^> lblume: thanks - just did :)
[14:11:11] <dizko> ill do another pkg refresh but ... had already done t
[14:11:12] <lewellyn> dizko: fun.
[14:11:33] *** merzo has quit IRC
[14:12:42] * lewellyn updates his fastcgi-php manifest... :P
[14:12:54] * lewellyn hates running some sites with specific php versions
[14:14:14] * lblume hopes it's still time to go buy some ice cream in walmart before it closes...
[14:14:24] <dizko> noooooo
[14:14:30] <dizko> walmart is evil =(
[14:15:59] <lewellyn> dizko: it's his "buy local" market
[14:16:19] <lewellyn> for us, it's evil. for him, it spurs the local economy : )
[14:16:20] *** zazenrasta has joined #opensolaris
[14:16:23] <dizko> is he in china?
[14:17:16] <lblume> yup! Walmart is good! :-D
[14:17:50] <lblume> At least for food. Seeya all!
[14:18:07] <dizko> ive been considering making a trip to costco, but i have no idea where id put a 30gallon drum of penut butter in my tokyo apartment
[14:19:59] *** merzo has joined #opensolaris
[14:20:51] *** Dagobert has quit IRC
[14:26:08] <dizko> lewellyn: if i tell it to install 'pkg'  it gets further but it wants cherrypy
[14:26:24] <lewellyn> kinky
[14:26:28] <dizko> i installed SUNWpython-cherrypy, bu tit still bitching about pkg:/library/python-2/cherrypy at 3 dot 1.1,5.11
[14:26:32] <lewellyn> file a bug :P
[14:27:15] <lewellyn> my opinion is "if you can't upgrade to 134, don't assume you'll be able to upgrade to release"
[14:27:57] *** merzo has quit IRC
[14:29:21] <dizko> i wish the installer would let you install into a different be
[14:30:14] <dizko> probably there is very little actually configured on this server that i cant just redo and id be better off with a clean install
[14:30:48] <lewellyn> dizko: that doesn't get bugs fixed! :D
[14:31:04] <lewellyn> think of it as doing your part with helping development, if you don't pay for support :)
[14:31:13] *** psychicist has joined #opensolaris
[14:31:14] <lewellyn> (if you pay for support, bitch about that issue!)
[14:31:16] <dizko> hehe, well, ive never felt that upgrading was as critical a requirement as the newest version working well
[14:31:33] <lewellyn> you can't get to the newer version from the older version without an upgrade :)
[14:31:43] <dizko> i dont upgrade windows
[14:31:45] <dizko> always reinstall
[14:32:24] <lewellyn> i know someone who runs vista on a box that started as 98se
[14:32:32] <dizko> haha that's impressive
[14:32:45] <lewellyn> he's running with the box's max ram: 2 x 256mb :D
[14:32:48] <lewellyn> 3x even
[14:33:10] <dizko> my phone has that much ram
[14:33:19] *** twisti has quit IRC
[14:33:42] <lewellyn> you have 768mb of ram? or nvram?
[14:33:49] <dizko> oh, well 512
[14:34:23] *** merzo has joined #opensolaris
[14:34:31] *** Dagobert has joined #opensolaris
[14:34:36] <tsoome> my 2100z did start with solaris 9, upgraded to s10EA builds, then plain s10, then scxe, when sxce was declared dead, only then it got next initial install and now again only upgrades.
[14:43:29] *** syamajala has joined #opensolaris
[14:45:00] <RoyK^> erm - should I worry about this error message? http://pastebin.com/HC4CZmsC
[14:49:11] *** merzo has quit IRC
[14:49:23] *** iFire has joined #opensolaris
[14:49:27] *** PrickelPit has quit IRC
[14:50:19] *** syamajala has quit IRC
[14:51:14] *** kimc has quit IRC
[14:52:02] *** dalibor has quit IRC
[14:53:05] *** sHellUx has joined #opensolaris
[14:56:47] *** merzo has joined #opensolaris
[14:57:43] *** Mdx4 has quit IRC
[14:58:52] *** e1kg has quit IRC
[14:59:04] *** leito has quit IRC
[15:00:27] *** leito has joined #opensolaris
[15:03:18] *** merzo has quit IRC
[15:04:22] <Beket> Hey people.  I'm looking at https://shop.oracle.com/pls/ostore/f?p=ostore:2:0::NO:RP,2:PROD_HIER_ID:3760753060221631798889
[15:04:29] <Beket> What is 'perpetual license' ?
[15:04:56] <th> one that does not expire?
[15:05:06] *** fedreric has quit IRC
[15:06:10] *** leito has quit IRC
[15:06:57] <Beket> And what exactly do you buy ? i.e, how is different than the OOo that comes for 'free' ?
[15:07:08] *** Dagobert has quit IRC
[15:07:43] *** merzo has joined #opensolaris
[15:08:46] <lewellyn> Beket: you get support
[15:10:41] *** GhostInTheShell has quit IRC
[15:10:49] <Beket> Where does it say that? I can't find it in then features/benefits/tech specs
[15:11:34] <lblume> Link returns nothing for me, strange.
[15:12:00] *** merzo has quit IRC
[15:12:30] *** sHellUx has quit IRC
[15:12:43] *** vivid has joined #opensolaris
[15:14:30] <RoyK^> hi all - trying again - should I worry about this error message? http://pastebin.com/HC4CZmsC
[15:14:53] <Shoggoth> hi all!
[15:15:33] *** leito has joined #opensolaris
[15:16:02] *** nikolam has joined #opensolaris
[15:16:19] <lewellyn> Beket: also, some companies require all software installed to be licensed
[15:16:42] *** dajune has quit IRC
[15:16:53] <lewellyn> "Oracle Open Office (previously called StarOffice or StarSuite) Standard Edition "
[15:17:14] <lewellyn> that would mean it comes with the extra plugins
[15:17:16] <lewellyn> and fonts
[15:17:20] <lewellyn> the all important fonts ;)
[15:17:52] *** michelem has joined #opensolaris
[15:18:03] <michelem> hello folks
[15:18:08] <lewellyn> let them finish getting the software into the system before we bitch too much about the software not being all in the system :)
[15:18:47] <michelem> I have a question related to ZFS. I tried to produce an incremental backup for the first time today.
[15:19:31] <tsoome> with zfs send?
[15:19:58] *** merzo has joined #opensolaris
[15:20:18] <Shoggoth> so I'm still stuck with figuring out how to use pktool to import a symmetric key into the pkcs11 softtoken
[15:20:27] <Shoggoth> anyone here any good with pktool ?
[15:20:53] <lblume> lewellyn: Did I tell you how much fonts suck in the latest OOo? First time I don't have wysiwyg since Winword 2.0
[15:20:57] <michelem> yes
[15:21:12] <michelem> what I do is, I create recursive snapshots of the root dataset
[15:21:23] *** blizzow has joined #opensolaris
[15:21:32] <michelem> I then remove individual snaps of the datasets I don't want to backup
[15:22:16] *** panos has joined #opensolaris
[15:22:19] <lewellyn> lblume: hahahaha
[15:22:20] <michelem> to create the incremental snap today, I used the command:     zfs -i snap-0-zroot-3apr2010 zroot@snap-1-zroot-10apr2010
[15:22:43] <michelem> notice I didn't request recursive serialization (-R)
[15:23:23] <michelem> now, the incremental backups made this way for all filesystems are ok. The one for zroot, however, is insanely big
[15:23:24] *** dajune has joined #opensolaris
[15:23:49] <michelem> the base (full) was 102 MB, the incremental (in one week) is 10 GB
[15:24:30] <michelem> is this reasonable?
[15:25:58] <michelem> ouch!
[15:26:30] <michelem> I just noticed that in the second snapshot there were previous backup files stored
[15:26:35] <tsoome> if your full backup was from empty fs, and you have got 10GB data during the week, ofc its reasonable
[15:27:42] <michelem> this lack of support for chflags +nodump is really annoying. ..
[15:27:54] <tsoome> ?
[15:29:09] <tsoome> if you need backup, use backup tools. zfs send is more about file system replication (which coincidentally can be used backup as well).
[15:30:28] <tsoome> also, normally you have "backup" target on separate pool. it would make little sense to backup into same pool....
[15:30:37] *** Fullmoon has quit IRC
[15:30:59] <michelem> definitely, that was garbage
[15:31:04] <michelem> I have a /nobackup dataset
[15:31:21] <tsoome> at least with zfs send - as it wont protect you from pool failure and for data backup the snapshots are sufficient
[15:32:04] <tsoome> there are plenty of tools better suited for selective backups:)
[15:32:26] <michelem> i use snapshots for backup, and (other) snapshots + zfs send for disk failures
[15:32:37] <tsoome> makes sense:)
[15:33:37] <michelem> tsoome: I used to use plain dump(8) with UFS2. That did the job fine: a bit slow and complex to restore, but supported excluding and recovering of individual files to/from backup
[15:33:38] <tsoome> but yes, zfs send itself is not meant to be used to select/unselect files
[15:34:02] <tsoome> yes, but ufsdump/ufsrestore is backup tool, not replication;)
[15:34:18] <michelem> what would you suggest as a backup tool for disk failure on ZFS?
[15:34:35] <tsoome> altho you *can* use it for replication as well in some scenarios
[15:34:55] <michelem> you definitely can
[15:35:05] <tsoome> zfs send is normally fine for full dataset replication.
[15:35:33] <michelem> and for backup? :)
[15:35:42] <tsoome> but thats the whole point of unix to have different tools and now its user task to pick right too for right job;)
[15:36:32] <tsoome> you cn use zfs send for backup as well if you keep in mind it wont deselect files;)
[15:36:34] <michelem> you say there are plenty of backup tools around. True, but they're one more irritating than the other :) will you make a commitment recommending a good one usable with ZFS? :)
[15:37:10] <tsoome> tar, cpio, zfs send, cp, zip, and so on....
[15:37:13] *** gmarler has joined #opensolaris
[15:37:15] <tsoome> all are usable
[15:37:57] <michelem> to give an example, I host mail for several domains. One is a photo studeio, and exchanges lots of huge attachments a day (photos). I opted them out for backups, as that would make 50+ GB alone, backup images would not be reasonably transferable to remote locations.
[15:38:39] <michelem> so, since zfs send doesn't support +nodump, I had to create a special /nobackup pool, move this domain's directory there, and them symlink it from the appropriate location
[15:39:27] <tsoome> well, depends - with snapshots and zfs send you can create fairly frequent "backup" cycles to avoid sending data in huge amounts. ofc that would require quite amount of storage in backup site.
[15:40:20] <tsoome> if you need to conserve the backup space as well, then you need to start to apply different selections etc and in that case, zfs send might not be best tool;)
[15:41:11] <tsoome> or, maybe try to reorganize data into different datasets and apply different policies on different datasets...
[15:41:35] <michelem> yes, this is what I do now
[15:41:52] <michelem> I find ZFS a fantastic FS so far, and zfs send a nice tool to complement that
[15:41:58] *** blizzow has quit IRC
[15:42:17] <michelem> but for working around this thing I needed to create that special dataset
[15:42:46] <michelem> where the (empty) directory structure of the entire filesystem is cloned
[15:43:29] <michelem> and where I "move" data that needs not to be backed up, and then symlink it back to the "real" fs location
[15:44:02] <michelem> I understand that and why zfs send doesn't support file exclusion
[15:44:18] <michelem> I'm just saying that it's an annoying characteristic
[15:45:11] <tsoome> might be worth to file RFE;)
[15:50:36] *** macros73 has joined #opensolaris
[15:51:25] *** merzo has quit IRC
[15:54:52] <duckinator> hi
[15:57:02] *** merzo has joined #opensolaris
[16:01:09] *** merzo has quit IRC
[16:04:27] *** Rune has joined #opensolaris
[16:04:42] *** GhostInTheShell has joined #opensolaris
[16:07:09] *** merzo has joined #opensolaris
[16:07:20] *** mroconnor has quit IRC
[16:10:17] *** TomJ has quit IRC
[16:10:25] *** michelem has quit IRC
[16:11:43] *** dalibor has joined #opensolaris
[16:15:15] *** merzo has quit IRC
[16:17:47] *** hajma_ has quit IRC
[16:18:53] *** wdp__ has joined #opensolaris
[16:19:33] *** het has quit IRC
[16:21:02] *** merzo has joined #opensolaris
[16:21:45] *** myosound has joined #opensolaris
[16:22:55] *** wdp_ has quit IRC
[16:22:58] *** Shoggoth has quit IRC
[16:31:25] *** sergiusens has joined #opensolaris
[16:34:06] *** synegy34 has quit IRC
[16:36:17] *** nevdull has joined #opensolaris
[16:37:01] *** merzo has quit IRC
[16:37:04] <nevdull> dns nevdull
[16:37:16] *** Odin- has joined #opensolaris
[16:42:07] *** dalibor has quit IRC
[16:42:40] *** merzo has joined #opensolaris
[16:43:59] *** lblume has quit IRC
[16:44:43] *** duckinator has quit IRC
[16:45:34] *** lblume has joined #opensolaris
[16:47:48] *** bball has quit IRC
[16:47:51] *** d1b has quit IRC
[16:48:57] *** merzo has quit IRC
[16:49:29] *** bball has joined #opensolaris
[16:53:55] *** projectmoon has joined #opensolaris
[16:54:08] *** leito has quit IRC
[16:54:29] *** d1b has joined #opensolaris
[16:55:11] *** mmcgee has joined #opensolaris
[16:55:26] *** merzo has joined #opensolaris
[16:55:38] <CIA-21> Tim Haley <Tim.Haley at Sun dot COM>: 6572591 meta dnode lookup causes bucket lock contention in dbuf hash
[16:55:41] *** bball has quit IRC
[16:56:03] *** duckinat- has joined #opensolaris
[16:58:34] *** projectmoon has quit IRC
[16:58:56] *** projectmoon has joined #opensolaris
[16:59:14] *** mmcgee has left #opensolaris
[16:59:47] *** nachox has joined #opensolaris
[16:59:51] *** projectmoon has joined #opensolaris
[17:00:03] *** p3rror has quit IRC
[17:00:59] *** eleusis has quit IRC
[17:01:01] *** leito has joined #opensolaris
[17:01:58] *** duckinat- has quit IRC
[17:03:05] *** eleusis has joined #opensolaris
[17:03:05] *** eleusis has joined #opensolaris
[17:04:10] *** nachox has quit IRC
[17:04:31] *** ivo_ has quit IRC
[17:04:33] *** duckinator has joined #opensolaris
[17:04:55] *** Kedare has joined #opensolaris
[17:05:50] *** ivo_ has joined #opensolaris
[17:06:00] *** myosound has quit IRC
[17:08:22] *** Aria has joined #opensolaris
[17:08:24] *** bball has joined #opensolaris
[17:12:51] *** projectmoon has quit IRC
[17:16:56] *** Mousk4r1 has joined #opensolaris
[17:19:27] *** p3rror has joined #opensolaris
[17:19:35] *** Mousk4r has quit IRC
[17:22:23] *** projectmoon has joined #opensolaris
[17:22:48] *** projectmoon has quit IRC
[17:24:13] *** Yu\2 has joined #opensolaris
[17:24:40] *** het has joined #opensolaris
[17:30:00] *** PrickelPit has joined #opensolaris
[17:31:17] *** RoyK^ has quit IRC
[17:32:27] *** TomJ has joined #opensolaris
[17:37:09] *** winstonw has quit IRC
[17:40:06] *** leito has quit IRC
[17:40:07] *** Shoggoth has joined #opensolaris
[17:44:02] *** bahumbug has joined #opensolaris
[17:46:46] *** Mdx4 has joined #opensolaris
[17:49:28] *** projectmoon has joined #opensolaris
[17:55:20] *** syamajala has joined #opensolaris
[17:55:55] *** Shoggoth has quit IRC
[17:56:14] *** Jago is now known as dnaumov
[17:57:56] *** leito has joined #opensolaris
[17:58:19] *** macros73 has quit IRC
[18:00:32] *** bahumbug is now known as niq
[18:00:59] *** macros73 has joined #opensolaris
[18:01:48] *** gebi has joined #opensolaris
[18:05:59] *** leito has quit IRC
[18:06:07] <TomJ> Can I run pkg install multiple times in parallel?
[18:06:24] *** leito has joined #opensolaris
[18:07:29] <tsoome> why dont you add many pkgs at once with one command?:P
[18:08:27] *** macros73 has quit IRC
[18:08:28] <TomJ> well yes I can do that too, but it may be quicker to run two in parallel, if it's supported
[18:10:46] <lblume> Just try it.
[18:20:22] <asyd> I'll surprise if it's possible
[18:28:23] *** sergiusens has left #opensolaris
[18:29:07] <lewellyn> if it is, it's a bug
[18:30:39] <lewellyn> hrm. i have an interesting need... :P
[18:32:28] <galt> kg locks files in /var
[18:32:33] <galt> pkg even
[18:33:42] <lewellyn> so... let's say i needed to batch convert images to be a certain size AND a particular aspect ratio...
[18:33:58] <lewellyn> what, besides scripting gimp, can i use?
[18:34:31] <lblume> convert
[18:34:36] <lewellyn> (e.g. an image may be 874x500, but i need it to fill 640x480, with no cropping)
[18:35:14] <lewellyn> lblume: imagemagick doesn't seem to have a way to say "scale and make the target this image size, regardless of the image data size when scaled"
[18:35:39] <lewellyn> as long as the filler color is white or black i'm happy, even
[18:36:08] <lewellyn> no
[18:36:19] <lewellyn> scale maintains aspect ratio, fitting one dimension or the other
[18:37:57] *** lupin__ has joined #opensolaris
[18:38:36] <lupin__> why it is that there's no vx-rootdsk shown on devalias output, but its listed on nvramrc? http://pastie.org/912935
[18:39:12] <lewellyn> i'm afraid to ask ##linux, but i suspect that's the sort of thing they do more often in there :(
[18:39:22] <tsoome> make sure your nvram is enabled.
[18:39:38] <lblume> lewellyn: -resample
[18:40:29] <lblume> Honestly, I doubted your assertion that ImageMagick couldn't do something so basic ;-)
[18:40:35] <lupin__> tsoome: you mean this one use-nvramrc?  ?
[18:42:01] <TomJ> Anyone know a way to get a new zone to use a different standard list of packages to install?  Or to specify a list of packages to always be additionally installed on zone creation?   Yeah I can automate it with a script, but wondering if there's a "new zone master package list" that could be edited
[18:43:39] <tsoome> indeed.
[18:44:23] *** gmarler has quit IRC
[18:45:16] <tsoome> also, if its internal disk, you dont really need that alias, as there are aliases for all internal disks. it still can help to identify alternate boot disk.
[18:45:29] <lewellyn> lblume: this is slow... we'll see if it works :)
[18:45:57] <lblume> IT does, I tried ;-)
[18:46:21] <lblume> What's the size of your pics?
[18:49:27] <lupin__> tsoome: hmmn..I don't get, sorry I'm newbie. You mean for all internal disk its not necessary to create an alias? Then how shall I boot from alternate boot disk for example?
[18:50:03] <lewellyn> image size or file size?
[18:50:58] <lblume> resolution, yup
[18:51:41] <lewellyn> 137528 bytes 1600px × 1200px
[18:52:04] <lewellyn> trying to make them fit 500x333
[18:52:59] <lblume> Shouldnt be so slow.....
[18:53:02] <lewellyn> still running...
[18:53:07] <jbit> https://bugs.launchpad.net/ubuntu/+source/upstart/+bug/557177 is hilarious :)
[18:53:13] <lewellyn>  29778 username  783M  467M sleep   60    0   0:02:02 5.3% convert/1
[18:53:24] <lewellyn> that's a LOT of ram for such a small image
[18:53:52] *** e1kg has joined #opensolaris
[18:53:57] <lblume> wow, yes
[18:54:06] <lblume> Let me see here....
[18:54:08] <galt> hmm, is capsunlock a seperate key function?
[18:54:15] *** mes has joined #opensolaris
[18:54:15] *** nitrile has joined #opensolaris
[18:54:55] *** projectmoon has quit IRC
[18:55:19] *** RoyK^ has joined #opensolaris
[18:55:30] <tsoome> internal disks have aliases like disk0 (disk), disk1 and so on. alias is only used to make it easier to boot from, so you can do boot disk instead of boot /pci..............
[18:55:43] <lewellyn> lblume: looks like i'll have to figure out how to script gimp :P
[18:55:56] <lewellyn> ooh, or...
[18:56:06] <lewellyn> i can scale with graphicsmagick, then use netpbm!
[18:56:13] * lewellyn tries to remember how to do that with netpbm
[18:56:24] <lupin__> tsoome: I see, thanks
[18:56:50] *** leito_ has joined #opensolaris
[18:57:06] *** leito has quit IRC
[18:57:44] <galt> and yes, you can set aliases to anything that prom can see
[19:00:53] <mes> anyone know if it is possible to find a b130 copy of SXCE?
[19:01:30] <TomJ> mes: http://www.sunfreepacks.com/
[19:02:08] <lewellyn> lblume: aha. resample doesn't do what you thought it did
[19:02:24] <tsoome> sxce cant be redistributed.
[19:03:38] <lblume> lewellyn: It changes the dpi resolution, right?
[19:03:43] *** leito__ has joined #opensolaris
[19:03:56] *** leito_ has quit IRC
[19:03:57] <lblume> It bricked my system, too. Easy DoS.
[19:04:10] <lblume> That Linux feeling I got, hmmmm......
[19:04:38] *** niq has quit IRC
[19:04:51] <ampex> there's nothing wrong with linux
[19:05:12] <lewellyn> lblume: yes it does
[19:05:24] <lewellyn> and i think i have the answer. please hold for the next available representative
[19:06:02] *** syamajala has quit IRC
[19:06:20] <jamesd2> ampex, all OSes suck, they just all suck in different ways, its best to pick the OS that meets your needs and suck less at those things... if there is nothing wrong with linux why are you here?
[19:06:53] <ampex> jamesd2: because I agree with you
[19:06:58] <lblume> ampex: Except that Solaris used to handle such excessive resource demand much better than Linux. Not anymore. Anyway.
[19:07:16] <mes> TomJ: thanks, do you know a place that lists the diffs between SXCE and opensolaris proper?
[19:07:33] *** FireflyST has joined #opensolaris
[19:07:44] <jamesd2> lblume, you mean i can't crash my linux box in 2 seconds by executing a fork bomb but solaris lives through it?
[19:08:24] <TomJ> mes: no, but SXCE is dead now. there won't be any more releases or updates, so you're best off just getting the latest OpenSolaris and sticking with that.
[19:09:06] <TomJ> you can either get Osol 06.09 from the main site, or genunix.org has a b134 ISO which is quite a bit newer (and is a beta or alpha for 10.04 or whatever the nexxt release ends up being called)
[19:09:21] <TomJ> I am using the latter and it seems pretty solid
[19:09:32] <jamesd2> TomJ, unless you want to run sun prepackaged software or other 3rd party software that hasn't been ported to opensolaris yet, and still want new functionality that sxce gives...   sunray server, sgd, sun vdi, to name a few.
[19:09:46] <TomJ> jamesd2: yes that's true
[19:10:23] <lblume> jamesd2: Yes, that's it. At least, Solaris won't start killing random processes until hitting a critical one by mistake.
[19:11:20] <lblume> lewellyn: -resize 123x456! the FM has more details than the CLI ;-)
[19:12:38] *** PhlowX has joined #opensolaris
[19:13:07] <mes> I don't like the Sun prepackaged software (GNU stuff), but what about Elite3d and other components with no source?
[19:13:59] <lblume> Problem with those is not the lack of source, but the lack of Xsun.
[19:14:02] <lewellyn> lblume: got it!
[19:14:16] <lewellyn> gm composite -gravity center -geometry 500x333 "$1/$pic" "basesnap.pnm" "$1/snaps/$pic"
[19:14:43] <galt> lblume: actually, in OOM, that's precisely what Solaris DOES, kill random processes
[19:15:07] *** psychicist has quit IRC
[19:15:10] <galt> I think I've mentioned OOM killin off desktop switcher before
[19:15:31] <mes> I have a SunBlade 2000, been running svn_57 for a while ... lblume: opensolaris uses xorg? and xorg has no drivers for elite3d?
[19:15:47] <lblume> lewellyn: My way is simpler ;-)
[19:16:03] * odyi has never actually seen Solaris go OOM.
[19:16:06] <lblume> galt: Where is that documented?
[19:16:40] <odyi> Now Linux...that does it all the time and oom-killer is like shooting fish is a barrel with a shotgun.
[19:16:44] <galt> lblume: my half-vast experience
[19:16:47] *** nitrile_ has joined #opensolaris
[19:16:54] <lblume> mes: I don't know if it has or not, but that would be the main problem. nVidia is not open, but has drivers for Xorg.
[19:16:59] <jamesd2> mes, you might want up to upgrade, since solaris 10u8 has newer patches than svn_57..
[19:17:24] <lblume> galt: What I've seen is processes dying because they fail to handle a failed malloc()properly. But not Solaris killing them.
[19:17:42] *** nitrile has quit IRC
[19:17:58] <lewellyn> lblume: your way doesn't add a background color and center the picture in the middle
[19:18:06] <tsoome> solaris can go oom quite easily depending on your memory and swap setup. and no, there is no berseking app killing random processes in case of oom.
[19:18:11] <lewellyn> it simply resizes the picture to fit inside a 123x456 box
[19:18:21] <lewellyn> (while maintaining aspect ratio)
[19:18:44] <lblume> lewellyn: You missed the very important ! at the end
[19:19:09] <lewellyn> lblume: that breaks the aspect ratio
[19:19:11] <lewellyn> ;)
[19:19:22] <lewellyn> i WANT the aspect ratio, and i want it to fill the "box"
[19:19:44] <mes> Well, I have two new 147G fibre channel drives, and I am planning to make what will likely the last build out for this machine, or any sun hardware for that matter. Just trying to go in the right direction
[19:19:47] <galt> lblume: quick way to sort of recreate, use FF on a lot of image sites, with a half dozen or so tabs, when the disk starts thrashing, watch as random things go bye-bye
[19:19:56] <lblume> ok, hmm, I'm not sure I understand what you want then :-D
[19:20:05] *** Spencer_tt has joined #opensolaris
[19:20:21] <lblume> galt: I just told you: those random things are just dying because they don't handle failed malloc()
[19:20:33] <lblume> It's not a kernel action, it's applications bugs.
[19:20:43] <galt> in that case, there are a LOT of uggy apps
[19:20:45] <lblume> Those who do handle that don't die.
[19:21:08] <tsoome> if you get NULL from malloc and will attempt to use it, you will get what you deserve
[19:21:09] <Aria> You have no idea how many buggy apps are out there when it comes to no resources.
[19:21:13] <Aria> Out of RAM and out of disk space both
[19:21:27] <lblume> Yes, because usually, an app that does a malloc() does so for a reason, and exit() might be the best way to handle it.
[19:22:17] <lblume> It might not be exactly buggy, but I doubt there are very graceful ways to fget out of that. But the Solaris kernel does not shot random passerbys.
[19:22:34] <lewellyn> lblume: i'll give you some before and after in msg in a moment :)
[19:23:10] <lblume> Not too long? I ought to get back on a regular schedule ;-)
[19:23:55] <jbit> actually i was suprised that some apps do handle failed allocations gracefully
[19:24:15] <lewellyn> lblume: already given ;)
[19:24:31] *** milek_ has joined #opensolaris
[19:24:34] <jbit> image manipulation/viewing apps are usually a good candidate, if you limit them to say 1gbyte and try to load a huge image alot of such apps will just complain ratehr than crashing
[19:25:56] <galt> I wonder how much magic you'd have to do to have your app monitor RSS, and take steps to make sure that it wasn't eating the whole machine
[19:26:21] <lblume> lewellyn: Aha, got it! Sorry, too late for me to understand quickly ;-)
[19:26:22] <jbit> galt: it depends on the app, some apps can easily free memory and give it back to the OS
[19:26:33] <lewellyn> it's also not an easy problem to understand :)
[19:26:56] <lewellyn> i can't say i've ever considered this case before :)
[19:27:06] <lblume> jbit: What kind of app can do that? free() doesn't.
[19:27:12] <lewellyn> (note that graphicsmagick is orders of magnitude faster than imagemagick)
[19:27:28] <lblume> daemons killing subprocesses maybe?
[19:27:35] <jbit> lblume: apps which manage their heap with something slightly more advanced than malloc()/free()
[19:27:39] *** leito__ has quit IRC
[19:28:02] <lblume> jbit: That's why I'm asking, I didn't study programming long enough :-)
[19:28:15] <jbit> lblume: my engine allocates dynamic allocation from bottom of heap upwards, and static level data from top of heap downwards
[19:28:30] <jbit> lblume: so every time i change level i can resize the heap and give the system back stuff if needed
[19:28:35] <jbit> (and, if the OS supports this)
[19:29:15] <RoyK^> jbit: why would you use something else than malloc/free?
[19:29:44] <jbit> RoyK^: so i can get more fine grained control?
[19:29:54] <RoyK^> control over what?
[19:29:55] <galt> RoyK: because you like not having arbitrary vulnerabilities?
[19:30:04] <lblume> jbit: Is that in practical use?
[19:30:08] <jbit> lblume: yes
[19:30:11] <jbit> lblume: most game engines do this
[19:30:39] <jbit> RoyK^: where it's allocated, how it deals with fragmentation, also i can tag stuff for dbugging use, etc
[19:30:41] <lblume> But OS's do support that and are able to take back the freed heap?
[19:30:43] <RoyK^> imho the best way to write stable, secure code is to do it by the book, with standard system calls, used correctly
[19:30:51] <jbit> lblume: yes
[19:31:00] <jbit> RoyK^: malloc/free is not a system call
[19:31:03] <jbit> RoyK^: it is a libc function
[19:31:12] <RoyK^> jbit: still
[19:31:19] <RoyK^> jbit: it's standardised
[19:31:42] <jbit> RoyK^: my allocate function takes six arguments, malloc takes one...
[19:31:56] <RoyK^> manually mmap()ing stuff just to make it more fancy won't make it easier to debug
[19:31:58] <lblume> jbit: Quite interesting.
[19:31:59] <jbit> while i could just wrap malloc this would limit my control
[19:32:01] <galt> so was passwd when kt made his presentation on trust...
[19:32:10] <jbit> RoyK^: that's not what i'm doing though
[19:32:13] <RoyK^> perhaps for you, since you know it, but what if someone needs to take over?
[19:32:20] <jbit> RoyK^: i'm doing it my own way since malloc doesn't give me enough control
[19:32:52] <RoyK^> control over WHAT?
[19:33:13] <jbit> RoyK^: where allocations are placed is hte main gain
[19:33:17] <jbit> RoyK^: as i've said twice now
[19:33:24] <RoyK^> if you start fucking around with manual allocations, it doesn't really help the system as a whole much
[19:33:43] <RoyK^> out of interest - what sort of system is it your making? a new rdbms or something?
[19:33:48] <jbit> RoyK^: if i knwo an allocation will only last a few milliseconds and then get freed i can decide to put it in a different place than a large one
[19:33:53] <jbit> RoyK^: it's a game engine
[19:33:56] *** nevdull has quit IRC
[19:33:59] <tsoome> you can always use malloc from libmapmalloc.
[19:34:39] <jbit> RoyK^: giving the allocator lots of information about hte allocation can help releive fragmentation issues, and can allow you to actually give things back t othe system
[19:34:45] <tsoome> there are like 3 different malloc implementations in solaris
[19:35:07] <jbit> which are both very important factors when you want to use 240mbytes of memory on a system that has 240mbytes of memory :p
[19:35:42] <jbit> memory allocation is not a solved problem, tehre is no one size fits all solution, and if there was it certainly wouldn'T be malloc()/free()
[19:35:45] <tsoome> jbit: which is not solaris, so its no issue.
[19:36:19] <lblume> Well, could be Solaris 8.
[19:36:20] <jbit> tsoome: i'd rather my game process use 500megs of ram ratehr than 700megs (aprox. numbers from actual use cases)
[19:37:24] <lblume> I never realized game developers actually cared about the gamers' machines RAM....
[19:37:44] <jbit> lblume: ones who also write console games do :)
[19:37:48] <jbit> lblume: most pc coders do not
[19:38:02] <lblume> Ah, ok, that makes sense :-D
[19:38:34] <jbit> if by clever monitoring and hinting of the heap allocation algorithm i can fit more assets in ram in one go.. that sounds like a pure win
[19:38:58] <jbit> and while this may seem specialized, i think thinking about allocation and having more advanced systems can help everything
[19:39:14] <lblume> ...except memory makers.
[19:39:26] *** leito__ has joined #opensolaris
[19:39:27] <lblume> You're trying to slow down the economic recovery here.
[19:39:35] <jbit> :P
[19:39:40] *** psychicist has joined #opensolaris
[19:40:05] <lblume> Anyway, thanks for the information, that's quite interesting :-)
[19:40:15] <lblume> Time for sleep now, bye all....
[19:40:23] <jbit> i was trying to find a really nice paper about this, which can explain it all far better than me
[19:40:29] <jbit> but unfortuantly google is failing me :P
[19:40:54] <TomJ> hmm, when you do :scrollback in screen ,it pre-allocates RAM for that amount.  e.g. do :scrollback 1000000 and it takes about 30 seconds and is then using 250MB RAM!
[19:41:08] <TomJ> it never used to do this in the screen I used in Solaris 10
[19:41:10] <TomJ> weird
[19:42:16] *** lblume has quit IRC
[19:42:29] *** PhlowX has quit IRC
[19:48:10] *** leito__ has quit IRC
[19:48:32] *** Aria has quit IRC
[19:50:56] *** leito__ has joined #opensolaris
[19:52:36] *** MikeTLive has joined #opensolaris
[19:53:32] *** andersenep has quit IRC
[19:55:35] <CIA-21> Gangadhar Mylapuram <Gangadhar.M at Sun dot COM>: 6942379 eeprom: getprop statements ending up in bootenv.rc
[19:55:39] *** derchris has quit IRC
[19:55:44] *** derchris has joined #opensolaris
[19:56:29] *** idonotknow_ has quit IRC
[19:56:50] *** iFire has quit IRC
[19:57:34] *** iFire has joined #opensolaris
[19:57:46] *** dalibor has joined #opensolaris
[20:02:05] *** icarus901 has quit IRC
[20:05:20] *** Dagobert has joined #opensolaris
[20:05:53] *** e1kg has quit IRC
[20:06:59] *** OclkdMan has joined #opensolaris
[20:07:14] <lewellyn> TomJ: file a bug! :D
[20:08:50] *** \amethyst has joined #opensolaris
[20:09:29] *** blizzow has joined #opensolaris
[20:16:21] <jbit> has anybody ever got virtualbox's direct3d acceleration working under an opensolaris host?
[20:17:11] <xyyz> how can you mirror 2 controllers with zfs?
[20:17:27] <xyyz> i have 2 controllers handling 4 disks each
[20:17:32] *** gebi has quit IRC
[20:17:35] <OclkdMan> is there an "emacs" pkg for opensolaris?
[20:17:44] <xyyz> and i want to mirror each group of 4 disks
[20:18:11] <Beket> OclkdMan, yes
[20:18:28] <xyyz> what i did was the following: zpool create mirrored-pool raidz disk1 disk2 disk3 disk4 raidz disk5 disk6 disk7 disk8
[20:18:36] <tsoome> you dont mirror controllers, you mirror disks
[20:18:58] *** milek_ has quit IRC
[20:19:21] <OclkdMan> in the official repo?
[20:19:22] <tsoome> its up to you to identify which disks are connected to which controller, you get it from cfgadm -al
[20:19:32] <xyyz> well that much i know
[20:19:49] <xyyz> but my command syntax
[20:19:53] <xyyz> will it do what i want
[20:19:56] <xyyz> or am i off?
[20:20:04] <datadigger> xyyz: I don't think that's a mirror of raidz
[20:20:05] <tsoome> man zpool
[20:20:19] <tsoome> you cant mirror raidz.
[20:20:29] <Beket> OclkdMan, yes
[20:20:38] <OclkdMan> ok thx!
[20:20:40] <xyyz> so then how do i go about doing what i want to do?
[20:20:48] <tsoome> man zpool
[20:20:57] <xyyz> *sigh*
[20:21:02] <xyyz> anyone else have any idea?
[20:21:06] <tsoome> vdev types are all described there
[20:21:38] *** OclkdMan has quit IRC
[20:21:42] <xyyz> man zpool doesn't exactly help a noob...
[20:21:57] *** synegy34 has joined #opensolaris
[20:22:10] <tsoome> have you actually attempted to read it?
[20:22:42] <xyyz> if i didn't read it, how else would i have known the commands i entered
[20:23:02] <tsoome> then how did you miss the part to create an mirror?
[20:23:07] <xyyz> anyways, i dont wanna get into an RTFM debate
[20:23:19] <xyyz> i don't think you're getting what i want to do
[20:23:31] <xyyz> it's more than just a mirror
[20:23:52] <tsoome> you can *only* mirror disks. its also written in that manual.
[20:24:34] <tsoome> so it can only be zpool create tank mirror disk1 disk2 mirror disk3 disk4 and so on
[20:24:35] <xyyz> i want to use raidz to make 2 groups of 4 disks each and then mirror those two groups
[20:24:55] <tsoome> tsoome: you can *only* mirror disks. its also written in that manual.
[20:25:03] <xyyz> but will that handle a drive failure in one enclosure?
[20:27:03] <xyyz> hmmm.... how much of a performance hit will i take if i raidz3 all 8 disks with 1 spare?
[20:27:19] <tsoome> what is the "that" you are referring? if you make mirror disk1 disk2, and disk1 is from one controller and disk2 is from another controller then yes, you are protected from controller failure
[20:27:32] *** leito__ has quit IRC
[20:28:04] <tsoome> how big are those disks?
[20:29:33] <xyyz> that being the configuration....  so if i do zpool create mirrored-pool mirror disk1e1 disk2e1 disk3e1 disk4e1 mirror disk1e2 disk2e2 disk3e2 disk4e2
[20:29:36] <tsoome> 8 disks may give better result with raidz if you wanna have perfomance, but ofc raidz3 will give more protection
[20:30:14] <tsoome> uhm, you wanna create 4 way mirror?:P
[20:30:29] <xyyz> ha... i dont
[20:30:33] <xyyz> i was just about to ask that
[20:30:53] <tsoome> mirror is keyword to define vdev
[20:31:12] <tsoome> in this example you would create mirror device of 4-way mirror
[20:31:34] <xyyz> shyte this is complicated... i guess raidz3 with a spare is the only way to do it :/
[20:31:52] <tsoome> to do what?
[20:32:02] <xyyz> have the redundancy i want
[20:32:49] <xyyz> i want to "keep going" in the event that 1 set of 4 disks disappears because the controller fails or the external enclosure fails
[20:33:46] <tsoome> well, whats wrong with mirror disk1e1 disk1e2 mirror disk2e1 disk2e2 and so on?
[20:33:48] <MikeTLive> you should do a stripe of mirrors.
[20:34:26] <tsoome> in this case you mirror your disks between enclosures and can loose entire enclosure.
[20:34:35] <MikeTLive> then you can afford alternate failures from either controllers and the rebuild is the individual mirrors not the whole stripe.
[20:34:51] <xyyz> hmmm.. i didn't think of it that way....
[20:35:00] <xyyz> that's actually pretty simple and brilliant
[20:35:03] <MikeTLive> if you stripe first then one fail in that stripe kills the stripe.
[20:35:17] <MikeTLive> one fail on the other controller kills that stripe.
[20:35:22] <MikeTLive> and buh-bye
[20:35:26] <xyyz> and i guess that also reduces the overhead
[20:35:33] <tsoome> you cant mirror stripes with zfs;)
[20:35:40] <tsoome> you only can stripe mirrors
[20:35:56] <xyyz> yet? or it's just not something you can do period?
[20:36:13] <MikeTLive> i cant come up with a solid reason to mirror a stripe.
[20:37:33] <Dagobert> Because the controller/raid software can't stripe mirrors :-)
[20:37:43] <tsoome> zpool is an stripe of vdevs. vdev is disk, or mirror or raidz*
[20:38:01] <tsoome> you cant mirror vdev, nor you cant make raidz* from vdev
[20:38:13] <tsoome> as nested vdevs are not supported
[20:38:30] <xyyz> hmmm thanks for explaining the reasoning too
[20:38:54] <tsoome> from man zpool:      Virtual devices cannot be nested, so a mirror or raidz  vir-
[20:38:55] <tsoome>      tual device can only contain files or disks. Mirrors of mir-
[20:38:55] <tsoome>      rors (or other combinations) are not allowed.
[20:39:00] <Dagobert> Funny thing is that SVM looks like mirrors of stripes, but is in fact stripe of mirrors
[20:39:45] <tsoome> Dagobert: yes, thats true.
[20:39:55] <xyyz> SVM?
[20:40:10] <Dagobert> Solaris Volume Manager. Sorry, Solaris...
[20:40:18] <tsoome> solaris volume manager. previously known as solstice disksuite
[20:40:33] <Dagobert> previously known as online disk suite
[20:40:50] <tsoome> :)
[20:41:05] <Dagobert> Hell, ODS was expensive at that time...
[20:42:19] <xyyz> is svm a package or something.... because i don't see /usr/sbin/smc
[20:42:40] <tsoome> you dont wanna have it with opensolaris
[20:43:01] <xyyz> why's that?
[20:43:51] <tsoome> because zfs does provide same functionality and more
[20:44:12] <tsoome> and raidz* is way better than raid5 from svm
[20:44:43] <xyyz> got it
[20:45:38] <tsoome> so, you dont like mirror disk1e1 disk1e2 mirror disk2e1 disk2e2 ?
[20:45:45] <tsoome> or did i miss something?
[20:47:04] <xyyz> no no... that's is sound
[20:47:07] <xyyz> and that's what i did
[20:47:15] <xyyz> much obliged for that
[20:47:49] <tsoome> well, that setup will be flawed if you loose both disks from same mirror pair
[20:47:55] <Dagobert> Is there some project for OSOL that does similar things like imon and dnotify from Linux?
[20:48:16] <xyyz> well let's hope that doesn't happen...
[20:48:33] *** CodeWar has joined #opensolaris
[20:48:46] <tsoome> thats the reason why raidz2 is better than mirror
[20:49:05] <xyyz> but i thought even raidz3 with a spare also isn't entirely redundant
[20:49:18] <tsoome> yep
[20:49:30] <xyyz> :/
[20:49:35] <tsoome> as if you loose 4 disks at once, you have dead pool
[20:49:56] <xyyz> haha... oh man
[20:50:08] <xyyz> alright then raidz3 with TWO spares :)
[20:50:26] <xyyz> that works no?
[20:50:39] <tsoome> well, thats still vulnerable if you loose 4 disks at once.
[20:50:58] <deet> redundant servers
[20:51:04] <xyyz> well what if i have the spares across enclosures
[20:51:06] <deet> always build two of everything, that's the american way
[20:51:10] <xyyz> so each enclosure has it's own spare?
[20:51:36] <tsoome> spare is only there to reduce time of having broken disk in system
[20:51:55] <xyyz> but that way there's data only on 3 drives right?
[20:51:58] <tsoome> disk fails, system will start resilvering to spare
[20:52:20] <xyyz> so if one enclosure drops out
[20:52:20] <tsoome> if you have nore errors, spare is "just" an empty drive
[20:52:36] <tsoome> s/nore/no/
[20:53:23] <xyyz> so then back to square one... there's no way to be completely redundant in my sittuation
[20:53:40] <xyyz> the mirror has a point of failure if both disks across enclosures crap out
[20:54:14] <xyyz> the raidz configurations wont support a 4 disk/enclosure/controller failure
[20:54:15] <tsoome> well. what are odds you can loose entire enclosure?
[20:55:15] <xyyz> not likely... but i dunno
[20:55:41] <tsoome> also, there is still one possible setup, but it needs ... wait, thare are 2 possible setups
[20:55:45] <tsoome> there*
[20:55:48] <xyyz> i'm all ears
[20:55:57] <tsoome> you can make raidz from 2 disks
[20:56:12] <xyyz> how is that different from a mirror though?
[20:56:36] <tsoome> so you can do raidz disk1e1 disk1e2 raidz disk1e1 disk2e2
[20:56:46] <xyyz> if in the very unlikely event that the raidz'd disks across the enclosure fail then i'm still screwed right?
[20:57:05] <xyyz> err enclosures
[20:57:08] <tsoome> yes.
[20:57:32] <xyyz> same problem as the mirrors... :/
[20:57:38] <xyyz> so what's the second way?
[20:57:46] <tsoome> add more enclosures:P
[20:57:54] <xyyz> ha!
[20:58:05] <xyyz> i wish they still made these 3ware sidecars
[20:58:12] <tsoome> then you can use raidz2 on 3 disks:P
[20:58:14] <xyyz> i'd be very interested in getting one :)
[20:58:50] <tsoome> but then again, its always the question how valuable is your data versus downtime.
[20:59:37] <tsoome> i mean, if you create mirror over 2 enclosures and have backup, then you will have your data still even if both disks in mirror will fail
[20:59:57] <xyyz> well, fortunately, this isn't a production environment.  it's a home/learning environment
[21:00:14] <tsoome> its still another question how high are chances you loose entire enclosure
[21:01:08] <tsoome> also, the more redundancy you wanna have, the more you will loose perfomance.
[21:01:12] <xyyz> each enclosure will be on a fairly decent sized UPS
[21:01:37] <tsoome> then prolly go with raidz2 or 3, depending on perfomance you need
[21:01:48] <tsoome> then you can afford to loose 2 random disks
[21:02:02] <xyyz> mirror has better performance than raidz right?
[21:02:10] <xyyz> or are they going to be the same for 2 disks?
[21:02:28] <tsoome> yes, mirror does not need to make parity calculations
[21:03:10] <xyyz> can i add an SSD for cache in an already used disk?
[21:03:14] <tsoome> with raidz* you wanna have decent amount of ram to compensate for speed
[21:03:25] <xyyz> i have 8 gigs on the machine
[21:03:30] <xyyz> that should be enough right?
[21:03:33] <tsoome> yes, you can add cache to existing pool
[21:03:50] <xyyz> does it give a significant performance boost?
[21:04:10] *** phil has quit IRC
[21:04:13] <tsoome> depends on data and usage. in our one DB server we have arc cache about 40GB
[21:05:13] *** Hedonista has quit IRC
[21:05:24] <xyyz> this is what i get
[21:05:32] <tsoome> as always, the more you have the better is the result.
[21:06:06] *** phil has joined #opensolaris
[21:06:23] <xyyz> pfexec zpool add enetwerks-zfs cache c4t0d0
[21:06:32] <xyyz> invalid vdev specification
[21:06:41] <xyyz> use '-f' to override the following errors:
[21:06:47] <tsoome> the less cache, the more physical IO you need to do.
[21:07:06] <xyyz> ...   /dev/dsk/c4t0d0s0 is part of active ZFS pool rpool. Please see zpool(1M).
[21:07:17] <xyyz> this disk is the OS disk
[21:07:25] <xyyz> so can i still use it for the cache
[21:07:30] <xyyz> or will it break everything?
[21:07:52] <tsoome> rpool? you have ssd for rpool?
[21:08:15] *** Yu\2 has quit IRC
[21:08:18] <tsoome> and i would not share disk between pools if i can
[21:09:18] <tsoome> and certainly you cant share disk like that, with this command you did attempt to assign entire disk as cache
[21:09:31] *** Hedonista has joined #opensolaris
[21:09:42] <xyyz> goes to show you how much i dont know
[21:10:13] <tsoome> if you have entire c4t0 allocated to rpool, you cant share it with anything else
[21:10:50] <xyyz> then i'll need an extra ssd if i wanted to do this
[21:10:54] <tsoome> yes.
[21:11:07] <xyyz> man, the costs keep on adding
[21:11:23] <tsoome> sharing disk between 2 pools is possible, but if that disk will go bad, you have impact on both pools
[21:11:31] <tsoome> and thats not good.
[21:12:09] *** Dagobert has quit IRC
[21:13:16] <xyyz> hmmm... so right now the setup that you suggested that i implemented with mirroed single disks across enclosures is the best one available
[21:13:25] <tsoome> well, cache is like extension of arc in ram, its nice to have, can help to get IO load off from physical disks, but is not "must have"
[21:13:49] <CodeWar> why not buy more RAM ..
[21:13:52] <tsoome> you can start without, see how it will go and decide later if you still wanna add it
[21:14:33] <tsoome> and yea, for the price of ssd you probably could get quite an amount of ram
[21:15:03] <CodeWar> PEX SSD seemed really expensive ..
[21:15:25] <tsoome> obviously there is an limit of ram you can insert:)
[21:16:28] <xyyz> well i have 8gb of ram
[21:16:33] <xyyz> in 2gb modules
[21:16:41] <CodeWar> true what size SSDs are popular these days ..48GB?
[21:16:48] <xyyz> going to 4gb a module becomes really expensive
[21:16:53] <xyyz> do they make 48gb versions?
[21:17:03] <xyyz> i always thought they were in multiples of 32gb
[21:17:16] <xyyz> with the exception of a 60 OCZ makes
[21:17:17] <CodeWar> xyyz, how many dimm slots on your chipset?
[21:17:21] <TomJ> never seen a 48GB SSD.  32, 64, 128, 256, is the normal, but there's one company that does 30/60/120, and a couple of other exceptions
[21:17:21] <xyyz> 4
[21:17:25] <TomJ> yeah OCZ that's them
[21:17:38] <xyyz> each with a 2gb module
[21:17:38] <CodeWar> hmm thats less
[21:17:49] <CodeWar> DDR2?
[21:17:53] <xyyz> DDR2
[21:17:57] <TomJ> oh and 160/320 is also available - my Intel SSD is a 160GB
[21:18:14] <xyyz> intel? latest generation?
[21:18:19] <TomJ> yes
[21:18:28] <xyyz> i hear they're blisteringly fast
[21:18:45] <TomJ> mine isn't as fast as some on write.  it's 240MB/s read, 110MB/s write
[21:18:46] <xyyz> i have a kingston in here, which i think is a rebranded intel, but i'm not sure
[21:19:07] <TomJ> I chose it because it was 160GB - my other option was a 128GB Corsair that had 200MB/s write. but I chose the extra 32GB in preference
[21:19:19] <TomJ> not sure I made the right choice actually, I'm not using the extra space yet, not nearly
[21:19:31] <tsoome> for cache you need fast reads only
[21:19:42] <xyyz> that's the thing... people think that the SSD sizes are too small
[21:20:07] <xyyz> but if you keep the OS on it, and then your data standard HDs
[21:20:15] <xyyz> you['ll find that it has more than enough space
[21:20:17] <tsoome> well, if you take ssd as hdd replacement, then yes, they are too small:)
[21:20:35] <TomJ> my SSD is my laptop's only drive, which is why I wanted a bit more space. but if I made the choice again I would go for a faster 128GB
[21:20:39] <xyyz> i have the 60GB OCZs in 2 laptops.... i haven't had a space issue yet
[21:21:04] *** pjfloyd has joined #opensolaris
[21:21:07] <tsoome> i have 500gb hdd in laptop with 276GB of free space:)
[21:21:14] *** Aria has joined #opensolaris
[21:21:24] <CodeWar> well on my 2xxeon 5620 system I have 16 dimm slots ddr3 each with a 1GB stick to save money .. I really dont see SSD helping over more RAM
[21:21:56] <tsoome> could have some ssd and external hdd, but i dont like an idea to carry separate hdd...
[21:22:01] <xyyz> and that xeon's a home system?
[21:22:13] *** Mousk4r1 has quit IRC
[21:22:50] <CodeWar> yep workstation got it at 2 USK .. if I d hand made it woul d have cost 1500 but at the time didnt know how to glue things together
[21:23:10] *** kwinz3 has quit IRC
[21:23:27] <xyyz> that's a pretty amazing setup.... a bit beyond my budget though...  :)
[21:23:42] <xyyz> tsoome... btw.... thanks a really lot for your help with this
[21:23:58] <tsoome> you are welcome
[21:24:52] <xyyz> i'll probably consider adding an smaller ssd when the prices come down in a few months
[21:25:12] <mui> are we there yet are we there yet
[21:25:18] <mui> no, dedup bug still there
[21:25:21] <mui> :(
[21:25:23] <tsoome> lol
[21:25:49] <xyyz> take care all
[21:25:54] *** xyyz has quit IRC
[21:26:19] *** kwinz3 has joined #opensolaris
[21:28:04] *** randw has joined #opensolaris
[21:28:38] *** hajma has joined #opensolaris
[21:28:47] *** stewart has joined #opensolaris
[21:32:54] <CodeWar> http://www.thenerds.net/INTEL.Intel_8_DIMM_PCI_Express_x16_Memory_Expansion_Board.BFCMEM.html?affid=8&ci_src=14110944&ci_sku=BFCMEM^~^INTEL
[21:34:34] *** nikolam_ has joined #opensolaris
[21:38:10] *** nikolam has quit IRC
[21:39:31] <turtle> w
[21:39:34] *** mavhk has joined #opensolaris
[21:40:40] *** mes has quit IRC
[21:44:10] <jbit> CodeWar: want... maybe...
[21:45:14] <jbit> i wonder what it shows up as on the bus? and what most operating systems will do with it
[21:45:40] <CodeWar> an I/O device that is memory mapped? hopefully Intel provides drivers for it
[21:46:19] <jbit> well, the OS should be made aware that it'S slower than normal memory
[21:46:56] <jbit> but even with that knowledge i don't know if an OS can make wise decissions, i think solaris can make decissions about memroy placement when it has lots of CPUs with their own memory controllers... so maybe this can fit into the topography some how
[21:47:14] <CodeWar> I see it as a disk cache not memory replacement
[21:47:52] <CodeWar> but yes NUMA aware OS can make similar decisions if they want to
[21:48:14] <jbit> "Memory Placement Optimization"
[21:48:15] <tsoome> solaris got memory placement technologies quite some time ago;)
[21:49:06] <CodeWar> I m not convinced NUMA DSM is a successful technology for high performance apps.. I can elaborate if required .. the whole business of hiding latency from the programmer is a mistake
[21:49:14] <TomJ> Is there any workaround to rename a ZFS filesystem used by a zone?  I'm getting:  "cannot rename 'rpool/zone/manage2': child dataset with inherited mountpoint is used in a non-global zone"
[21:49:14] <jbit> CodeWar: whame it's not ram backed, would make a nice zil
[21:49:49] <tsoome> TomJ: shut down the zone and unset zoned parameter
[21:49:51] <TomJ> I guess I could down the zone, set the zone's mountpath to something else, and then try - but I guess I'd need to manually hack the zone XML files to do that, otherwise changing zonepath is disallowed
[21:49:59] <TomJ> oh, thanks tsoome!
[21:50:03] <jbit> http://www.ddrdrive.com/ is what you want your zil on :)
[21:50:03] <TomJ> didn't know that was writable
[21:50:15] <tsoome> and you can use zoneadm detach/attach to reset zonepath
[21:51:22] <jbit> $2000 is a bit much for 4gigs though
[21:51:30] <tsoome> zoneadm also got move command, im not sure if it does support zfs rename tho....
[21:51:35] <TomJ> hmm, I'm getting device busy on /zone/manage2, even with the zone shutdown.  Maybe I'll set it not to autoboot and reboot the server, that might be easiest
[21:52:16] *** ivo_ has quit IRC
[21:52:23] <tsoome> make sure you are not in that directory in your shell;)
[21:52:33] <TomJ> yeah I'm not :)
[21:52:42] <TomJ> just going to install lsof before I do the whole server reboot
[21:52:45] <TomJ> see if I can find what it is
[21:53:20] <tsoome> and you may need to manually unmount filesystem(s) in it, or at least, unset zoned parameter, as i already wrote:)
[21:53:38] <TomJ> sure I did the zoned thing as you said
[21:53:46] <TomJ> you didn't mention the unmount - I'll try that too
[21:54:08] <tsoome> if you had that dataset defines in zone config, the zoneadm will set zoned to prevent global zone from accessing it
[21:54:11] <TomJ> no, zfs unmount gives device busy too
[21:54:19] <TomJ> and lsof didn't show any files using that path
[21:54:23] <TomJ> I'll just do a reboot I guess
[21:54:35] *** strato has joined #opensolaris
[21:54:52] *** strato is now known as ivo_
[21:54:57] <TomJ> ah no wait
[21:55:12] <TomJ> I turned zoned off at the wrong level.  I turned it off for rpool/zone/manage2 and it needs to be done on rpool/zone/manage2/ROOT
[21:56:05] *** CodeWar has quit IRC
[21:56:06] *** bdrewery has quit IRC
[21:56:16] *** hsp has joined #opensolaris
[21:56:35] *** ma3x has quit IRC
[21:56:40] <TomJ> hmm, unsetting zoned hasn't allowed me to use rename. it still knows it's part of a zone
[21:59:31] <TomJ> argh, still can't rename it even after unsetting zoned, detaching the zone, and manually editing the zone.xml and the index file to remove reference to that filesystem
[21:59:35] <TomJ> I'll try a host reboot
[21:59:41] <tsoome> :)
[22:00:23] *** nikolam_ is now known as nikolam
[22:03:05] *** bdrewery has joined #opensolaris
[22:03:33] *** bdrewery is now known as Guest89670
[22:03:43] *** synegy34 has quit IRC
[22:09:15] *** ma3x has joined #opensolaris
[22:09:17] *** ma3x has joined #opensolaris
[22:15:26] *** dalibor has quit IRC
[22:15:43] *** eleusis has quit IRC
[22:17:07] *** eleusis has joined #opensolaris
[22:19:47] *** niq has joined #opensolaris
[22:19:49] *** raichoo has joined #opensolaris
[22:24:16] *** MikeTLive has quit IRC
[22:25:11] *** MikeTLive has joined #opensolaris
[22:27:24] *** CodeWar has joined #opensolaris
[22:27:47] *** andersenep has joined #opensolaris
[22:28:17] <CodeWar> wth ... are they molesting Sun employees @ Oracle or what .. now  Goslings out :-(
[22:29:38] *** winstonw has joined #opensolaris
[22:29:42] *** MikeTLive has quit IRC
[22:30:44] <jamesd2> CodeWar, if they are, everyone is being quiet, just a constant flow of blogs leaving  blogs.sun.com with no one saying why.. and lots of people leaving.
[22:31:55] <CodeWar> jamesd2,  which would be the professional thing to do of course
[22:32:43] <jamesd2> CodeWar, yes but with the number of defections you would think someone would say something.. not everyone is professional 100% of the time.
[22:34:02] <CodeWar> I have a theory but it would be impolite and harsh on the Sun employees so I ll keep it with me :-(
[22:35:33] <CodeWar> anyways someone like Gosling leaving has an impact on Java followers .. cant wait for the crazy rumor mills to start grinding this :-)
[22:35:51] <pjfloyd> this is irc, why hold back on being impolite and harsh?
[22:36:03] <Beket> ++pjfloyd
[22:36:04] <Beket> :P
[22:36:44] *** bahumbug has joined #opensolaris
[22:39:21] *** niq has quit IRC
[22:39:31] *** bahumbug is now known as niq
[22:40:01] <lewellyn> CodeWar: they have
[22:40:10] <lewellyn> it just hasn't made it into an article yet :P
[22:40:29] <CodeWar> conspiracy theories
[22:40:32] <lewellyn> yup
[22:40:53] <lewellyn> of course, there's also the "everything's fine" view at the opposite end of the spectrum
[22:41:34] <lewellyn> "for all we know, he had been thinking about leaving for a while, and with a great offer right now, and this being a good time to transition, he did so"
[22:42:16] *** Yu\2 has joined #opensolaris
[22:42:50] <CodeWar> haha twitter is going nuts .. i hope they get extra machines/backup this weekend
[22:46:28] *** Beket has quit IRC
[22:46:57] <CodeWar> suffice to say I ve been amazed reading Sun blogs.. folks talk so openly about IP and future products.. how does that make business sense
[22:47:25] <CodeWar> and I never quite figured why they started open sourcing like crazy either.. even open sourced SPARC ... even if you cant sell machines you can sell IP look at transmeta
[22:48:50] <tsoome> its all about IP and shared knowledge and technologies.
[22:49:35] <tsoome> if you can share, you will be able to sell systems as well.
[22:50:12] <CodeWar> huh?
[22:50:50] <lewellyn> CodeWar: companies like minimizing vendors
[22:51:09] <lewellyn> that's one of the best things about oracle+sun
[22:51:27] <lewellyn> it oracle doesn't fuck up badly, they'll reverse the ibm trend quickly
[22:51:54] <tsoome> look how science works. every result is shared, and result is that you will get more and more new discoveries and achievements
[22:51:59] <CodeWar> I dont doubt Oracles business strategies its Suns ex Opensource everything that always surprised me
[22:52:36] <CodeWar> tsoome,  I disagree with your line of thinking so badly that trying to get my thoughts over to the other side is not even going to be possible over a text medium :-)
[22:52:46] <lewellyn> CodeWar: 1) they can't close the sources. 2) larry likes to one-up ibm, and the sheer number of oracle open-source projects lets him.
[22:53:55] <CodeWar> ok let me try .. you hire 3 real smart guys they work on afuture product for 3 years.. theres 10 IPs involved.. not patent yet why would I want to open source it
[22:54:06] <richlowe> You don't.
[22:54:12] <tsoome> well, ofc, you can try to work as typical 20. century company - close up all, sell your closed solution and tie your customer from hands and legs into your product.
[22:54:23] <richlowe> in the cases I know well, the patent was *always* filed before the code and important docs became visible
[22:54:28] <richlowe> for the sake of the validity of the patent.
[22:54:42] <richlowe> perhaps that was just used as an excuse, but used as an excuse it has been.
[22:54:45] <CodeWar> richlowe, I m no patent lawyer but patent filed is not patent granted
[22:55:33] *** pjfloyd has quit IRC
[22:57:13] *** purserj has quit IRC
[22:57:39] *** purserj has joined #opensolaris
[22:57:46] <CodeWar> to argue on the side of closed systems not being bad for business let me ask this ... how many internals do you find on JRockit/ blogs about nitty gritty details? Is it open source? No right? does that make it less competitive? does that make it less usable? IIUC its got quite a decent penetration most of the top benchmarks are using jrockit not hotspot
[22:59:05] <tsoome> well. if they have not made sparc consortium there would have no fujitsu sparc around and sun would have been long dead.
[22:59:22] <CodeWar> NDA?
[22:59:35] <nikolam> CodeWar, point of open development as I see it is that every damn one who install, use it, send bugs, talk about is.. etc .. not to mention writing code is contributing.
[23:01:27] *** winstonw is now known as CATS
[23:01:47] *** CATS is now known as CATSS
[23:02:42] *** CoolMa has joined #opensolaris
[23:04:14] <CodeWar> you dont want idiots contributing to the core OS / kernel on a daily basis .. what works for Linux doesnt ahve to work for Solaris .. you need masses for bug fixes /spot fixes not large component devel or perhaps I m missing something by a generation here
[23:04:56] <duckinator> i agree with CodeWar ^^
[23:04:57] <tsoome> thats very true, you cant have an complete anarchy
[23:05:23] <jbit> most large opensource projects are quite strict
[23:05:25] <tsoome> but then again, its like natural selection process you see in nature
[23:05:35] <duckinator> after working very hard on multiple oses...and rewriting one completely because some douche trashed the source while "fixing" it... i dont think you want any old idiot to access it....
[23:05:43] <tsoome> if you are bad, you wont stay around
[23:05:44] <duckinator> especially not without a mediator somewhere, but that's just a waste of time imo
[23:06:05] <jbit> tsoome: unfortuantly, it's usually the loudest who stay around, not the best
[23:06:13] <jbit> tsoome: and the loud ones will often drive away your good programmers
[23:06:16] <tsoome> thats true as well
[23:06:18] <duckinator> just find people who you can trust at least some, and use git..problem solved ^^
[23:06:22] <duckinator> i drive off the loud ones
[23:06:22] <duckinator> :D
[23:07:16] <duckinator> blocking their access to anything involved with dev does wonders :)
[23:07:24] <tsoome> but then again, if you think you are the smartest guy around to decide what should get in and what not, who can quarantee you really are and you really are not making mistakes?;)
[23:07:41] <CodeWar> tsoome, the fact that you own the business and that if you f* up you lose money
[23:07:54] <tsoome> yes, and wont be around;)
[23:08:06] <CodeWar> I m not arguing against open source .. but saying a company the size of Sun with heck lot of IP doesnt need to do what Linux does
[23:08:08] <jbit> tsoome: alot of devs spend far too much time "making sure mistakes don't happen"
[23:08:22] <jbit> tsoome: when they should really be spending time making sure things don't go to hell when they do
[23:08:36] <duckinator> from my experience, what works best - at least for smaller operating system projects - is having 5-10 people who trust each other "in charge", and have multiple coders who help out here and there if need-be, have the 5-10 main people check everything
[23:08:59] <duckinator> of course those numbers are based on the fact that my os is only ~4.6k lines long
[23:09:10] <duckinator> (of code)
[23:09:10] <CodeWar> oh yes .. thats how it is in our company too ... 4 people thats it .. everybody else bug fix or communicate but ownership is clear and business driven
[23:09:18] <lewellyn> i was going to say, you need 5-10 testers for building on alone :P
[23:09:25] <jbit> duckinator: indeed, also you really really need to make sure people know who's in charge
[23:09:44] <jbit> duckinator: and even if you think they're wrong you should follow what they're saying
[23:10:03] <duckinator> jbit: yea, i learned that first-hand...that's why i rewrote dux (my first os) completely, once i got rid of the more idiotic devs :\ they completely trashed the source when i turned my back for a week while being out of town
[23:10:04] <jbit> there is a time for debate and discussion and a tiem for getting shit done
[23:10:31] <duckinator> if you cant trust them to get shit done without being idiots when you walk away, dont let them in to start with
[23:11:37] <CodeWar> to play devils advocate .. based on what I see on lkml.org the advantage Linux has is an army of bug fixers for free
[23:11:56] <CodeWar> for a copmany with closed source triage/debug/fix/review/checkin is a huge bottleneck
[23:12:02] <jbit> google did a nice talk on open source project managment... hit on some great topics such as an awesome coder with a poisonous attitude will do much more damage than a bad coder who has a good attitude
[23:12:49] <lewellyn> Requirements
[23:12:51] <lewellyn> * Patience, and some Coffee™.
[23:13:00] * lewellyn thinks he'll get along fine with this thing!
[23:13:00] <duckinator> fwiw, to clear up any potential misunderstandings, i'm only 17, and my os is not run by anything resembling a company - just a small group of coders around my age ^^
[23:13:06] <tsoome> person with good attitude knows his limits and will accept fixes.
[23:13:18] <jbit> tsoome: nod
[23:13:28] <jbit> bad coders who knwo they are bad coders are a useful resource
[23:13:35] <jbit> good coders who think they are the second coming are not
[23:14:44] <tsoome> and above all, the ideas are important, once idea is accepted, there will be enough people to implement it
[23:15:29] <CodeWar> jbit, link?
[23:16:09] <jbit> CodeWar: http://video.google.com/videoplay?docid=-4216011961522818645
[23:16:16] *** cervello has joined #opensolaris
[23:16:32] *** niq has quit IRC
[23:16:48] <jbit> i think that's the one anyway
[23:25:41] *** echelog has joined #opensolaris
[23:26:18] *** CodeWar has quit IRC
[23:28:15] *** Yu\2 has quit IRC
[23:29:17] *** randw has quit IRC
[23:30:20] *** balrog has quit IRC
[23:31:00] *** randw has joined #opensolaris
[23:33:15] *** CoolMa has quit IRC
[23:34:28] *** jthunder has joined #opensolaris
[23:36:01] *** bdrewery has joined #opensolaris
[23:36:30] *** bdrewery is now known as Guest29377
[23:38:00] *** Guest89670 has quit IRC
[23:39:30] <Spencer_tt> jbit: I thought coders who can get things done are good enough - egos need money and other bits to manage
[23:41:27] *** piwi has joined #opensolaris
[23:42:09] *** justnow has joined #opensolaris
[23:43:50] *** ibseo has joined #opensolaris
[23:47:01] <jbit> Spencer_tt: well, that's kinda my point, you don't want coders with egos :)
[23:47:58] <jbit> you want coders who can go "i think you're wrong, i've said why, but i'm going to continue to work your way anyway otherwise nothing will get done"
[23:50:00] *** raichoo has quit IRC
[23:50:35] <Spencer_tt> yeah I'll look at that a practical advice which works in coding environments, even students know how to get things done once they have the confidence to tackle a few challenges minus the ego ;)
[23:51:52] *** dalibor has joined #opensolaris
[23:55:19] *** jamesd2 has quit IRC
[23:56:52] *** piwi has quit IRC
[23:58:04] *** sergiusens has quit IRC
[23:59:52] *** sergiusens has joined #opensolaris

   April 10, 2010  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >