[00:00:03] <alanc> X version changes!
[00:00:04] <timeless> sh versioning changes! :)
[00:00:06] <Gman> slowhog_home: back
[00:00:12] <gdamore> ksh93? :-)
[00:00:25] <alanc> oh yeah, ksh version changes - those might be first
[00:00:54] <alanc> and GNOME version changes should be here in nv_53
[00:01:01] <timeless> jamesd: but seriously
[00:01:06] <timeless> can i get nexenta as a zone?
[00:01:27] <jamesd> ask in nexenta's mailing lists or possibly there mailing list
[00:02:13] <gdamore> heh.  I just submitted bug #1000 in our database.  :-)  I feel special.
[00:02:20] <Symmetria> hrm, anyone know where I get a working libstdc++
[00:02:24] <Symmetria> for solaris 10 x86
[00:02:37] <timeless> usually SFW
[00:02:39] <slowhog_home> hi Gman
[00:02:45] <timeless> are you trying to build something or run something?
[00:02:51] <gdamore> bug #1000 was hat we used /usr/j2se to build wifitool, instead of /usr/java
[00:02:51] <slowhog_home> do you know how to get a metacity core dump?
[00:03:06] <Symmetria> timeless trying to build something
[00:03:09] <gdamore> kill -SEGV ? :-)
[00:03:09] <Symmetria> lftp specifically
[00:03:21] <timeless> using which compiler?
[00:03:27] <Symmetria> have tried both
[00:03:49] <Symmetria> heh will install the package from the site
[00:03:54] <Gman> slowhog_home: i'd probably try and dbx it remotely
[00:04:13] <Gman> since if it crashes locally and you're in dbx, you'll lose keyboard/mouse focus
[00:04:35] <alanc> could turn on global core dumps with coreadm to dump to /var/core or somewhere like that
[00:04:38] <timeless> gman: Xvnc?
[00:04:57] *** steamR has joined #opensolaris
[00:05:00] <Gman> or what alanc said that i didn't know how to do :)
[00:05:11] <Gman> timeless: yeah, that would work too
[00:05:13] <slowhog_home> I do have global core dumps enabled, let me check if the core dump is there
[00:05:22] <timeless> although for some definition, that's remote
[00:05:29] <timeless> coreadm is neat
[00:05:31] <alanc> when you need to catch X server core dumps, you learn about coreadm very very quickly
[00:05:34] <slowhog_home> no, no /var/core
[00:05:43] * timeless has /var/crash/cores
[00:06:05] <timeless> <500mb of cores
[00:06:14] <timeless> of course, one's a 2gb gzip'd core :)
[00:08:14] <slowhog_home> hu, no, I didn't have global core dump enabled.
[00:08:29] * slowhog_home goes to enable global core dum
[00:09:27] * quasi wonders if tomorrow will be when sun announces honeycomb - http://blogs.sun.com/mary/entry/the_air_is_humming
[00:09:36] <gdamore> schily, are you here?
[00:11:12] <gdamore> what is honeycomb?
[00:12:22] <quasi> http://www.sun.com/storagetek/honeycomb/
[00:13:09] <gdamore> interesting.  sounds like they forgot about auspex though. :-)
[00:13:09] <OnkelSchorsch_> hm. not much info there :(
[00:13:45] <slowhog_home> alanc: If I start metacity from failsafe mode, isn
[00:13:57] <quasi> OnkelSchorsch_: there's more if you search for it
[00:14:00] <slowhog_home> isn't that core goes to per-process?
[00:14:22] <slowhog_home> I did have that enabled, but didn't see a core dump
[00:14:34] *** piwi has joined #opensolaris
[00:14:35] <alanc> if you turn on global core dumps it should go to both
[00:14:36] <OnkelSchorsch_> quasi, on the sun site? or do you mean those news-site articles?
[00:14:49] <alanc> I have no idea why it wouldn't also be going to the per-process location
[00:14:56] *** stefanjo has quit IRC
[00:15:04] <quasi> OnkelSchorsch_: both
[00:15:26] <slowhog_home> got it. the .bashrc is not called for the failsafe session
[00:15:35] <OnkelSchorsch_> hm. quasi, haven't found any in-depth material
[00:15:35] <slowhog_home> which is where I did the coreadm
[00:16:03] <gisburn> slowhog_home: AFAIK */.bashrc is called for inteactive shells.
[00:16:07] <quasi> OnkelSchorsch_: http://www.sun.com/emrkt/innercircle/newsletter/0606edchoice_nontrad.html - but insane details probably has to wait for an official launch
[00:16:08] <gisburn> er
[00:16:12] <gisburn> ~/.bashrc
[00:16:26] <gisburn> slowhog_home: the login shell may not be interactive
[00:17:02] <OnkelSchorsch_> quasi, thanks.
[00:17:43] <quasi> OnkelSchorsch_: if you search a bit more, you'll find pictures with a fair number of discs
[00:17:50] *** pip` has joined #opensolaris
[00:18:06] <OnkelSchorsch_> yep. I've seen those
[00:18:23] <quasi> ah, there it was - http://research.sun.com/minds/2005-0628/
[00:18:24] <OnkelSchorsch_> I did a search on it a couple of month ago and also one, earlier today
[00:19:40] *** ShadowHntr has joined #opensolaris
[00:19:41] <Symmetria> question, with the stuff downloaded from blastwave, if I download all the dependancies etc, is there any way to make pkgadd install all the dependancies automagically when I install something that needs one
[00:19:46] <Symmetria> if I have all the files there
[00:19:49] *** micbro has joined #opensolaris
[00:19:57] <twincest> symm: you're not using pkg-get?
[00:20:04] <ShadowHntr> Symmetria: the pkg-get script does that automagically.
[00:20:07] <slowhog_home> gisburn: you are correct. I don't remember exactly why I decided bashrc is a better place, maybe for remote ssh sessions
[00:20:12] <Symmetria> twincest heh, I should be :p where do I grab that?
[00:20:22] <ShadowHntr> Symmetria: www.blastwave.org
[00:20:27] <twincest> read the docs on blastwave.org
[00:20:30] <Symmetria> aahh its on there
[00:20:31] <Symmetria> sweet
[00:20:32] <ShadowHntr> http://www.blastwave.org/pkg-get.html
[00:20:33] <Symmetria> :)
[00:21:15] *** Mazon is now known as mazon
[00:21:52] <slowhog_home> t@null (l@1) program terminated by signal SEGV (no mapping at the fault address)
[00:21:53] *** hile_ has joined #opensolaris
[00:21:53] <slowhog_home> 0xcf99544b: FcPatternPosition+0x006b:   movl     (%eax,%edx,8),%eax
[00:21:54] <slowhog_home> (dbx) where
[00:21:56] <slowhog_home> =>[1] FcPatternPosition(0x82a1b98, 0x81b6d58), at 0xcf99544b
[00:22:45] <slowhog_home> dbx metacity with a core dump in the failsafe term won't work, dbx segfaut itself
[00:23:17] <slowhog_home> I have to do it in a ssh session. Really don't know why.
[00:23:40] <Gman> slowhog_home: that's fontconfig
[00:23:45] <alanc> crap, that's fontconfig - you're screwed now
[00:23:53] <alanc> no one left here who knows that
[00:24:04] * Symmetria mirrors blastwave real quick
[00:24:05] <Gman> heh
[00:24:29] *** hile_ has quit IRC
[00:24:30] <slowhog_home> http://pastebin.com/807953
[00:24:59] <alanc> (I'm not joking - our fontconfig's guy last day was last Friday)
[00:25:16] <alanc> err, move the 's one word to the right
[00:25:39] <Symmetria> anyone got any idea how big that blastwave archive is?
[00:25:43] *** steamR has quit IRC
[00:26:00] <quasi> Symmetria: dclarke might know ;)
[00:27:01] <jamesd> Symmetria, x86 and sparc are about 4GB each..
[00:27:07] <jamesd> they fit on a dvd
[00:27:26] <Gman> slowhog_home: there was a fontconfig issue that people were experiencing a build or so ago - that might well be a dup
[00:27:35] <Gman> though i'm not a fontconfig expert either :)
[00:27:51] * slowhog_home should learn not to do an upgrade on laptop used for work
[00:28:20] <Gman> check to make sure the fontconfig smf service is running
[00:28:21] <alanc> I already tried passing it off as a dup of that bug
[00:28:22] <Symmetria> heh ok this will be a pretty quick mirror then
[00:28:23] <Gman> for a start
[00:28:32] <stevel> those are the best kinds of laptops to upgrade
[00:28:41] <stevel> because then you can justify "fixing" it as work :-)
[00:29:07] *** tjh has joined #opensolaris
[00:29:31] *** stefanjo has joined #opensolaris
[00:29:36] <slowhog_home> Gman: fc-cache is
[00:29:48] <slowhog_home> there is no pattern match fontconfig
[00:29:57] *** koolniczka has left #opensolaris
[00:30:10] <Gman> slowhog_home: hrm, that's the one i was thinking of
[00:30:50] <slowhog_home> stevel: I would agree with you, if I am in OpenSolaris group
[00:31:09] *** koolniczka has joined #opensolaris
[00:31:10] <slowhog_home> ;-)
[00:31:23] *** koolniczka has left #opensolaris
[00:32:21] <Symmetria> wheee 3gigs mirrored so far
[00:32:24] * Symmetria snores and waits
[00:34:39] <timeless> what's /opt/csw/lib and why would someone expect me to have it?
[00:34:47] <richlowe> timeless: Blastwave.
[00:35:06] <timeless> ah, thanks
[00:39:42] <tjh> Hi, I'm a newbie to Opensolaris. I have been using Solaris 9 & 10 a little bit. If I download Solaris Express Community Release build 49, can I use it as it is or do I have to start downloading and building more stuff like ON as soon SXCR is installed?
[00:39:45] *** peteh has quit IRC
[00:42:43] <alanc> you can use it as is
[00:42:58] <alanc> it's a full binary release of Solaris
[00:42:59] <gisburn> stevel: ping!
[00:43:06] <Symmetria> heh btw, if you guys ever get any south africans asking about blastwave in here other than me, you can tell em mirror.ac.za now carries it (I'll get it added to the mirror list shortly)
[00:43:11] <stevel> gisburn: pong
[00:43:18] <gisburn> stevel: when will be next intermediate source snapshot be released ?
[00:43:25] <richlowe> tomorrow.
[00:43:26] <gisburn> stevel: e.g. B50.5
[00:43:33] <gisburn> ok
[00:43:38] <gisburn> thanks! :-)
[00:43:39] <richlowe> unless stevel slacks.
[00:43:43] <Symmetria> heh, well, it will in a second: [Blastwave mirror running - current inbound: 80.72Mbit/second]
[00:43:44] <Symmetria> heh
[00:43:53] <stevel> gisburn: nightly deliveries are every tuesday
[00:43:54] <Stric> tjh: do note that official security fixes aren't released for such.. /me is having that problem with nspr & sxcr48 right now..
[00:44:16] * gisburn misses the links in the /topic
[00:44:21] * gisburn mourns them
[00:44:28] <stevel> why?
[00:44:28] <gisburn> !summon linksintopic
[00:44:39] <alanc> security fixes for SXCR are "upgrade to new build"
[00:45:19] <gisburn> stevel: because you can click them in many browsers and the other solutions are digging, digging, digging, digging, digging, digging, digging, digging, digging, found, click...
[00:45:32] <richlowe> or remembering "every tuesday"
[00:45:37] <richlowe> and dlc.sun.com/osol/on/downloads
[00:45:45] <stevel> or just rejoin
[00:46:33] <alanc> is there a msg you can send to ChanServ to have it send you the current greeting message?
[00:46:42] <stevel> alanc: not from what i could see
[00:47:13] <richlowe> I thought 'info #channel' included that.
[00:47:23] <gdamore> does nautilus use cdrecord or does it use cdrw?  can it burn data dvds?
[00:47:29] <stevel> richlowe: nope :(
[00:47:32] <richlowe> except it just says [#opensolaris is private]
[00:49:10] *** rmorse has joined #opensolaris
[00:50:08] <tjh> So one way to upgrade would be to download a new SXCR and wipe the hard disk and install the new SXCR? A quicker way would be to use the build tools and upgrade the old SXCR? Is the result the same?
[00:50:25] <Gman> you can use live upgrade either
[00:50:53] <Gman> which basically installs another version in a different part of the disk while you're running the older version
[00:51:00] <alanc> or just download the SXCR and use the upgrade option
[00:51:03] <oxygene> tjh: building won't give you _all_ the upgrades.. solaris is still more than O/N - but you can try the upgrade option of the installer or liveupgrade
[00:53:43] <Symmetria> heh btw, blastwave is a *lot* bigger than 8gigs now, it mighta been 4gig each a while back but heh Im already over 10 and climbing :)
[00:54:11] *** anthony79 has joined #opensolaris
[00:54:53] *** anthony79 has quit IRC
[00:55:28] <tjh> Stric: what is nspr? And why can't you use this live upgrade?
[00:56:45] <Stric> tjh: you can use live upgrade.. but you can't apply "regular patches"
[00:56:59] <slowhog_home> Gman: is the JDS svn alive/
[00:57:09] <alanc> so it's a 6 CD download instead of a meg or two download
[00:57:41] <slowhog_home> I think maybe I should try to build fontconfig with debug symbol
[00:58:32] <slowhog_home> maye there is something missing in the core dump, I think it does provide debug symbol
[00:59:03] <Gman> slowhog_home: later this week hopefully
[00:59:07] <alanc> hmm, how much hate mail would I get if I used gcc instead of Sun Studio for the offical Xorg 7.2 builds?
[00:59:36] <alanc> slowhog_home: was that 2 separate conversations?   fontconfig isn't in JDS svn
[01:00:24] <gisburn> alanc: I guess the decreased performance on SPARC would be a huge problem.
[01:00:26] <slowhog_home> alanc: I don't know that. :-P, so what consolidation has fontconfig? sfw?
[01:01:01] <alanc> gisburn: you can't get much of a decrease from "WON'T RUN"   - 0 is not less then 0
[01:01:07] <alanc> slowhog_home: X
[01:01:28] <slowhog_home> hehe, so you are the one I should keep bothering. ;-D
[01:01:42] <alanc> I just build it, I don't know how it works
[01:01:46] <gisburn> alanc: I guess it is more like |NaN| vs. |0|
[01:01:53] <gisburn> alanc: which problem do you have ?
[01:02:28] <alanc> the 40% performance loss on x86 from not being able to use MMX/SSE with Sun Studio since it doesn't recognize the code written for gcc
[01:02:44] <alanc> (in the alpha blending operations, not overall performance)
[01:02:49] *** micbro has quit IRC
[01:02:59] <gisburn> alanc: XRENDER or Mesa ?
[01:03:05] <alanc> render
[01:03:09] <gisburn> ouch
[01:03:16] <alanc> Without MMX:
[01:03:16] <alanc>   10000 trep @   4.2304 msec (   236.0/sec): Fill 300x300 aa trap with 4 bit alpha
[01:03:20] <alanc> With MMX:
[01:03:20] <alanc>   10000 trep @   3.1511 msec (   317.0/sec): Fill 300x300 aa trap with 4 bit alpha
[01:03:29] <gisburn> alanc: drop support for render and suggest to use Mesa instead ? =:-)
[01:03:30] *** nwf has quit IRC
[01:03:49] <alanc> mesa's asm is gcc only as well isn't it?
[01:04:02] <g4lt-mordant> what is this MMX you speak of, is that a kludge that proper processors don't need?
[01:04:17] <alanc> and the entire JDS team would come lynch me if I dropped Render support 8-)
[01:04:37] <gisburn> alanc: I think a while ago there was assembler code for Forte in mesa
[01:04:40] <alanc> g4lt-mordant: it's a feature built into all AMD64 processors, and very similar to the VIS kludge in SPARC processors
[01:05:03] <gisburn> alanc: then let the JDS team fix that problem - they want RENDER ? Then let them fix it.
[01:05:04] <twincest> alanc: can't you build x86 with gcc and sparc with studio?
[01:05:22] *** jacotton has joined #opensolaris
[01:05:29] <alanc> sure, if I ever actually build Xorg for SPARC
[01:05:36] <tjh> This is important basic stuff. I don't think it is mentioned very clearly on the home page. Or I just don't know where to look.
[01:07:22] <gisburn> Mhhh... searching for "terror" in bugs.opensolaris.org generates some intesting hits
[01:07:31] * gisburn searches for "bin laden"
[01:07:41] * boyd wonders why on earth you would search for terror
[01:07:46] <boyd> What about Komodo
[01:07:54] <Stric> boyd: looking for weapons of mass destruction perhaps?
[01:08:08] <boyd> I don't think he'll find any :)
[01:08:11] <slowhog_home> alanc: would you send me the debug build? or can I simply get the fontconfig 2.3.2 and build it?
[01:08:20] <tjh> So, what is nightly and mercurial?
[01:08:33] <gisburn> "bin laden" seems to be very popular at sun - 11301 hits.
[01:08:46] <alanc> I'd have to figure out how to make a debug build - we don't normally do that
[01:08:48] *** mlh has quit IRC
[01:09:07] <boyd> gisburn: Yeah, because "bin" would get so few matches
[01:09:22] <Gman> tjh: nightly source code drops, mercurial source code management tool
[01:09:39] <gisburn> 208 hits for "Komodo dragon"
[01:09:49] * sparc-kly|WORK is back (gone 117:37:06)
[01:09:56] * boyd wonders how many of those are bugs that gisburn logged
[01:09:57] <gisburn> boyd: so sun is in league with the evil ?
[01:10:06] * gisburn hides...
[01:10:16] <slowhog_home> alanc: ok, i just found the latest stable is 2.4, but I guess it won't be wise to try 2.4?
[01:10:31] <gisburn> 40 hits for "cooking"
[01:10:34] <alanc> you can try 2.4, but it won't match what's in Solaris now
[01:10:55] <alanc> Solaris has 2.3.2 + Sun patches which were offered upstream but not accepted
[01:11:02] <slowhog_home> Fontconfig version 2.4 is API and ABI compatible with the 2.3 release
[01:11:32] <gisburn> 25 hits for "goose"
[01:12:14] <gisburn> 5 hits for "buffalo"
[01:12:47] <gisburn> 0 for "giraffe"
[01:12:50] <gisburn> ;-(
[01:12:56] <Gman> ok, we get it
[01:13:14] <tjh> Is it possible to mount an EXT3 or ReiserFS filesystem in Solaris?
[01:13:36] <gisburn> tjh: someone here started to port ReiserFS3
[01:13:45] <jengelh> heh
[01:13:54] <slowhog_home> alanc: where can I find the source files used by Sun if I don't have access to SWAN?
[01:13:56] <jengelh> now that Novell moves away from reiser3, it's becoming the crossplatform standard :p
[01:14:03] <alanc> opensolaris.org
[01:14:04] *** mlh has joined #opensolaris
[01:14:21] <slowhog_home> whole X consolidation?
[01:15:13] <alanc> yep
[01:15:16] <gisburn> Gman: did you never looked at the problems of "Gnaw" not displaying correctly in gnome-terminal when the utf-8 mode was used ?
[01:15:26] <alanc> don't have it separately packages
[01:16:10] <Gman> gisburn: nope
[01:16:46] <tjh> Then what about EXT2? Or is there any way to share a files or a filesystem with Linux and Solaris?
[01:17:01] <jengelh> NFS...
[01:17:08] <jengelh> iso9660
[01:17:33] <jengelh> ufs I think it not too readwrite supported on linux
[01:17:44] <gisburn> Gman: do you have time to look at this soon ? I see similar problems with japanese manual pages not formatted correcly...
[01:17:48] <boyd> FAT
[01:17:50] <slowhog_home> the fontconfig seems to segfault consistently, xchat segfault at the same place as well
[01:17:56] <mlh> maybe someone could port sun's ufs and run it under linux's fuse
[01:18:03] <mlh> c'mon hop to it!
[01:18:12] <jengelh> ufs? haha.
[01:18:18] <jengelh> don't try - save the effort for ZFS
[01:18:21] <mlh> yeah fat is it
[01:18:29] <slowhog_home> share common bt up to pango_context_get_metrics
[01:18:38] <mlh> slowhog_home: no!
[01:19:43] <gdamore> hmmm.... nautilus doesn't work well on b49 for burning dvds.
[01:19:50] <slowhog_home> mlh: you are saying? ;-)
[01:20:09] <gdamore> it complains that i need to insert a disk with at least XXX MiB free.  (~3900)
[01:20:21] <gdamore> i put a blank dvd+r in the drive.
[01:20:26] <tjh> Can a hard disk be iso9660? I thought it was for CDs?
[01:20:52] *** coffman has quit IRC
[01:21:05] <gdamore> oh wait. never mind, i'm an idjit.  the drive is a cdrw, but only dvd rom.  argh.
[01:21:12] <twincest> well, a CD can be UFS
[01:21:22] *** benr has joined #opensolaris
[01:21:23] *** ChanServ sets mode: +o benr
[01:21:44] <jamesd> hi benr
[01:21:47] <gdamore> i don't see why you couldn't put an iso9660 fs on a hard disk partition, but it might not be performant
[01:22:03] <mlh> tjh: anything can be iso9660.  but it doesn't really help too much as it's not read/write
[01:22:05] <Stric> gdamore: you can.. but you can't write anything more to it..
[01:22:18] <mlh> whats the variant of iso9660 that's read/write again?
[01:22:24] <benr> hey
[01:22:29] <gdamore> true.  iso9660 is write-once, isn't it?
[01:22:50] <Stric> depends on how you see it.. it's "build the filesystem once"
[01:23:10] <mlh> ah:  udf
[01:23:43] <mlh> http://www.osta.org/specs/
[01:23:46] <slowhog_home> between b35 to b50, one fix is related to fontconfig. 5095888 libfreetype/libfontconfig should access name table with using UCS2/UTF-16BE
[01:24:00] <mlh> slowhog_home: you again? :-)
[01:24:02] <gisburn> does anyone remeber whether the blade1500 supports the XVR-1000 ?
[01:24:21] <gdamore> i believe it does.
[01:24:30] <alanc> no - XVR-1000 is UPA, SB1500 has no UPA slot
[01:24:45] <gdamore> i misread that as XVR-100.  doh.
[01:25:12] <alanc> yeah, XVR-100 (PCI Radeon 7000-based) works in SB1500
[01:25:13] <gisburn> alanc: because the taco chip provides no UPA port, right ?
[01:25:24] <gdamore> i think with sb1500 you have to use pci, or zulu (safari bus)
[01:25:27] <alanc> I don't know
[01:25:46] <gisburn> alanc: I am asking about XVR-1000, maybe the greasted gfx card sun ever made (except the XVR-4000) ...
[01:25:47] <alanc> zulu won't fit in SB1500 - it can only go into V880z
[01:25:56] <alanc> SB1500 is PCI only
[01:25:59] <gisburn> alanc: V890 works, too.
[01:26:04] <gdamore> safari bus, IIUC, is totally different from UPA.
[01:26:29] <mlh> apparently solaris supports udf v1.5.  I wonder how compatible the sun and linux implementations are.  not at all the bitter twisted inner self says
[01:26:32] <gdamore> i thought for sure it would fit.  there was this whole deal about sb1000 having an extra "expansion chassis" for zulu.  i'm amazed that it won't work in sb1500
[01:26:34] <alanc> yes - it's the backplane in the V880 series - you replace a CPU board with a video board
[01:26:58] <alanc> gdamore: that was the original plan, before the Zulu plan got cut way back
[01:27:13] <gdamore> ah.  so they never delivered it.  i haven't followed sparc graphics too closely.
[01:27:35] <alanc> not that we ever sold many Zulu (XVR-4000) boards - most people would rather have a car than a graphics board for $25k
[01:27:54] <gdamore> heh.  well, i thought the whole idea was rather outlandish.
[01:27:56] <slowhog_home> mlh: yup, i got a laptop screwed after upgrade frm b35 to b50, need to get that fixed. :-(
[01:28:02] <gisburn> At least three were sold in germany.
[01:28:10] <mlh> last link and I'll stop talking to myself: http://en.wikipedia.org/wiki/Universal_Disk_Format
[01:28:11] <gdamore> but i seem to recall one of the advertising points being "you could render Toy Story in real-time". :-)
[01:28:12] <gisburn> And I know two of the owners :-)
[01:28:42] <gdamore> heh.
[01:28:43] <alanc> XVR-1000 & XVR-4000 were the end of the line for Sun-designed graphics boards - everything since is purely OEM'ed from ATI & 3Dlabs
[01:28:48] <gdamore> UDF should be standard.
[01:29:09] <gdamore> and XVR-2500 is end of the line from 3Dlabs.
[01:29:15] <alanc> yep
[01:29:23] <gdamore> i wonder WTF sun is going to do when XVR-2500 stock runs out.
[01:29:55] <gdamore> there is no 3D support for Radeon hardware in Solaris right now.  And Nvidia is x86 only.
[01:30:38] *** rmorse has quit IRC
[01:30:41] <gisburn> gdamore: re: xvr-2500 - pray for a wonder ? They have eol'ed all the engineers who can build new gfx cards...
[01:30:50] <gisburn> gdamore: but there is one slim hope for sun
[01:30:55] <gdamore> (we're looking at XVR-600 parts for possible use on our high end "laptops".   but we have no long term roadmap.)
[01:30:56] <gisburn> gdamore: at least in theory
[01:31:14] <gdamore> ATI?
[01:31:41] <gisburn> gdamore: some group here in germany is building OpenGL-compatible realtime raytracing FPGAs
[01:31:51] <gdamore> that would be nice.
[01:32:34] <gisburn> gdamore: sun could harvest the students and their work, plug a MAJC in front of the stuff and would have a scaleable gfx card system which makes the xvr-4000 look like a toy.
[01:32:55] <gdamore> There is matrox i suppose.  but they seem to have abandoned the 3D market mostly , as well
[01:32:56] *** neoxed_ has quit IRC
[01:33:08] <gisburn> gdamore: right now they can outperform the xvr-4000 with 8 fpgas
[01:33:13] <Auralis> to bad about 3dlabs, they had the best pro cards
[01:33:35] <gdamore> 3dlabs is not doing any 3d work anymore.  they're targeting cellphones now. :-)
[01:33:36] <alanc> the odds of Sun shipping any graphics card not OEM'ed are somewhere south of slim
[01:33:58] <gisburn> alanc: I know.
[01:34:00] <alanc> http://www.tacc.utexas.edu/research/users/features/remotevis.php is the closest to anything original being worked on by Sun
[01:34:05] <gdamore> well, if someone built a card around suitable hardware, and provided drives, it could work.
[01:34:10] *** nachox has joined #opensolaris
[01:34:11] <alanc> (in the graphics field that is)
[01:34:52] <gisburn> alanc: I just say if sun wants to get back in the buisiness there would be a quick and efficient way to do so - and kick Nvidia&&ATI in a place where it hurts
[01:35:03] <alanc> Sun doesn't want to
[01:35:06] <gdamore> this is actually kind of criminal.  we have major customers who want SPARC and want OpenGL, and can pay big $$, but now we think there is no solution for them.
[01:35:09] <gisburn> alanc: I know.
[01:35:15] <gdamore> (big customers like the US military)
[01:35:21] <tjh> osta means buy in Finnish. ;)
[01:35:24] <gisburn> alanc: but we all know that they may change some day.
[01:35:40] *** neoxed has joined #OpenSolaris
[01:36:02] <alanc> there's is far too little of the old SPARC graphics group left to even think about custom hardware, much less have a chance of being able to build it
[01:36:17] <gdamore> if your german friends have a design for a PCIe board (or even PCI) using those FPGAs, and want to make drivers for Solaris for them, I can find a market for them really fast.
[01:36:26] *** Tpenta has quit IRC
[01:36:37] *** Tpent1 is now known as Tpenta
[01:36:49] *** ChanServ sets mode: +o Tpenta
[01:36:55] *** stevel has quit IRC
[01:36:57] <gdamore> as far as I can tell, the few people left in the SPARC graphics group are far more interested in trying to avoid getting canned than doing anything actually productive.
[01:37:10] *** Tpent1 has joined #opensolaris
[01:37:11] *** MadMethod has joined #opensolaris
[01:37:18] <gdamore> i met with some of them recently.  lets just say I was "not impressed".
[01:37:42] *** pip` has quit IRC
[01:38:01] <alanc> could try escalating to their management - they report to John Fowler now that SPARC & x86 hardware groups have been merged
[01:38:24] <gisburn> gdamore: right now they have a HT bridge and some weired hardware to plugin the FPGAs to that bridge - and much CPU time is wasted to shuffle the data and keep the pipelines busy - this is why I am suggesting to place a MAJC in front of the mess
[01:38:36] <gdamore> they wanted Tadpole to help provide them justification for keeping around Xsun, and didn't want to supply us any of the information we needed about Sun's OpenGL stuff.  (Stuff we needed to provide acceleration compatibility for XVR-100.  which we would have given _back_ to Sun.)
[01:38:59] <gisburn> gdamore: and I suggest to ditch the FPGAs and create real silicon (for which the students have no money for)
[01:39:10] <gdamore> alanc: its frankly not worth the effort to pursue further.
[01:39:28] <gdamore> that would help with power consumption as well.
[01:39:43] <gdamore> are they looking for investors?
[01:40:05] <gisburn> ugh
[01:40:07] <gisburn> er
[01:40:08] <gisburn> uhm
[01:40:12] <gisburn> these are students
[01:40:13] <gisburn> guess it
[01:40:15] <gisburn> :-)
[01:40:49] <Auralis> gisburn: the openrt stuff?
[01:41:08] <gdamore> lets just say if they could provide good OpenGL for Solaris, I could probably arrange for them to get either fully employed or have their work purchased outright.
[01:41:56] <gisburn> Auralis: yes
[01:42:07] *** jsubl2 has joined #opensolaris
[01:42:11] <Auralis> gisburn: that stuff is great
[01:42:22] <gdamore> but i need more than just ray tracing.  actually i need other parts of opengl worse.
[01:42:25] <gisburn> gdamore: their point is high quality
[01:42:43] <Auralis> 1bn polys interactive on the screen baby :)
[01:42:46] <gisburn> gdamore: not always performance... but then their aim at realtime, too.
[01:42:49] *** deather_ has joined #opensolaris
[01:43:21] <gisburn> gdamore: see http://www.openrt.de/Gallery/OliverDeussen_Sunflowers/Images/sunflowers_3.jpg
[01:43:32] *** hile_ has joined #opensolaris
[01:43:35] <gisburn> gdamore: watch the plant move in the wind - in realtime.
[01:43:57] <gisburn> gdamore: if they would get a dedicated frontend cpu (like MAJC) their stuff would rock.
[01:44:00] <Tpenta> !seen sommerfeld
[01:44:00] <Drone> sommerfeld is currently online in #opensolaris and last spoke on Mon 16 Oct 2006 15:56 GMT, saying 'twincest: no, the implementation was replaced in nevada'.
[01:44:40] <gisburn> gdamore: wait... http://www.openrt.de/Gallery/OliverDeussen_Sunflowers/Images/sunflowers_4.jpg is more impressive.
[01:45:15] <gdamore> lots of triangles, for sure. :-)
[01:45:27] <Auralis> the real impressice stuff is seeing that sunflower scene move in realtime
[01:45:35] <gisburn> yes
[01:45:38] <Auralis> gdamore: that scene is about 1bn polys
[01:45:40] <gdamore> i wonder what the power consumption for something like that is.
[01:45:49] <gisburn> gdamore: not much
[01:46:21] <gisburn> gdamore: the only real problem is the bottleneck caused by the "missing" frontend cpu
[01:46:21] <gdamore> we have some interesting power budgets in our mobile units. :-)  and PCI limits you to 25W over the bus anyway.
[01:46:33] <gdamore> i think that could be solved pretty easily.
[01:46:46] <alanc> just plug in a bunch of hard disk power connectors like the high-end nvidia cards
[01:46:47] *** krozinov has joined #opensolaris
[01:46:56] <gisburn> gdamore: note they use pre-2002 hardware for their stuff
[01:47:07] <gisburn> (at least the fpga stuff is)
[01:47:07] <alanc> of course, in a laptop, you'ld drain the battery pretty fast that way
[01:47:44] <gdamore> yes.  but we have some _big_ batteries. ;-)
[01:47:53] <gisburn> alanc: erm... throw $200000 at these guys to update their hardware and you'll get like a dual-XVR-4000 on y chip.
[01:47:58] <gdamore> and then our dual us3i systems have an external 300W power supply. :-)
[01:48:25] *** Method has quit IRC
[01:48:28] *** ShadowHntr is now known as maverickbna
[01:48:40] <nachox> i wonder if there are nvidia cards that consume more power than a regular modern cpu
[01:48:49] <gdamore> i think i'm going to talk to my management about these guys.  this is a market where we have a need, and getting some hardware out that would crank this would be cool.
[01:48:54] *** maverickbna is now known as ShadowHntr
[01:49:11] <gdamore> nachox: absolutely.  regular modern cpus are often far less than the 25W power budget of a PCI card.
[01:49:33] <gdamore> and nvidia and ATI boards often require _external_ direct power supply access, because they need more than 25W.
[01:49:42] <gisburn> gdamore: yes, but remeber my comments. and don't expect single-CPU silicon without investing lots of money in this direction
[01:49:43] <Auralis> nachox: standard cunsomer gfx card use about 40 watts currently, highend cards a lot more
[01:49:59] <gdamore> in fact, I know that even the 2500 has ~60W consumption.
[01:50:05] <gdamore> (XVR-2500 that is)
[01:50:18] <nachox> wow
[01:50:33] <gdamore> the XVR-600 uses 23W.
[01:50:34] <Auralis> my xvr-1000 drains 70 watts
[01:50:45] <Stric> next gen nvidia/ati is heading towards 200-250W
[01:50:48] <gisburn> gdamore: using FPGAs is not really usefull for a customer and the missing frontend chip would be a problem, too - it consumes lots of CPU power and dramatically limits the performace of the card.
[01:50:58] <alanc> gisburn: I have no money to throw at anyone
[01:51:11] <gdamore> we're talking General Dynamics here.  _we_ might have money to throw at this problem.
[01:51:22] *** solarisjon has joined #opensolaris
[01:51:50] *** solarisjon has left #opensolaris
[01:51:51] <nachox> hmm, what about the OGP? you need more than that right?
[01:51:56] <gdamore> of course, whether anyone else would ever see the benefit outside of GD/Tadpole hardware is entirely a different question.
[01:54:13] *** delewis has joined #opensolaris
[01:54:40] <slowhog_home> alanc: Now I think of it, the font name might be the issue. I have asian fonts comes with NV installed
[01:55:12] <slowhog_home> alanc: how can I verify/work around that issue?
[01:55:26] <alanc> find the fonts and rm them
[01:55:36] <alanc> or upgrade to nv_51 X pacakges
[01:56:25] <alanc>  	6477136- Application crashed in FcHideFont if non-alphabetical name exists
[01:59:06] *** me2_lol has joined #opensolaris
[02:00:05] *** deather has quit IRC
[02:00:25] *** dho has quit IRC
[02:00:25] <slowhog_home> hmm, FcHideFont is not in the backtrace
[02:02:30] <slowhog_home> firefox segfault at same location called from FcFontMatch
[02:02:45] *** saltmiser has quit IRC
[02:03:01] *** apa has quit IRC
[02:05:15] *** me2_lol has quit IRC
[02:12:43] <timeless> slowhog_home: can i see a pastebin url for your stack trace?
[02:13:13] <slowhog_home> for metacity or firefox?
[02:13:36] <slowhog_home> http://pastebin.com/807953 for metacity
[02:13:49] *** lloy0076 has joined #opensolaris
[02:14:38] <timeless> pastebin.mozilla.org ?
[02:14:47] <timeless> pastebin.com/pastebin.ca have been really slow of late
[02:15:08] *** alanc is now known as alanc-away
[02:16:27] <slowhog_home> http://pastebin.mozilla.org/874 for firefox
[02:17:48] <slowhog_home> http://pastebin.mozilla.org/875 for metacity
[02:18:27] * timeless looks for a cross reference that includes Fc*
[02:18:56] *** Kernel86_ has quit IRC
[02:19:00] <timeless> i don't seem to have one :(
[02:19:01] *** Kernel86|Laptop has joined #OpenSolaris
[02:21:16] *** eokyere has left #opensolaris
[02:23:24] *** MindChild has joined #OpenSolaris
[02:23:27] <gisburn> Is the solaris installer opensource ?
[02:23:44] <gisburn> gdamore: ping!
[02:23:47] *** steleman has quit IRC
[02:23:56] *** steleman has joined #opensolaris
[02:24:13] <MindChild> So I got this Ultra5 and it seems to have a "PC" card in it... 400 mhz x86 processor, video/sound/USB onboard... what is this for?
[02:24:19] <boyd> gisburn: Not as far as I know.
[02:24:24] <boyd> We need a new one anyway
[02:24:35] <gisburn> boyd: are there any plans to make it opensource ?
[02:24:38] <gisburn> boyd: why ?
[02:24:47] *** tjh has quit IRC
[02:25:00] <Triskelios> MindChild: that's a SunPCi card, you can run an x86 OS on it
[02:25:02] *** gtc_ has quit IRC
[02:25:29] <MindChild> Triskelios: Is there a special piece of software I gotta use, or is there buttloads of choices?
[02:26:29] <gisburn> Triskelios: correction:: you can run old windows versions on it.
[02:26:40] <gisburn> Triskelios: Linux is not supported
[02:26:45] <Triskelios> gisburn: we've run Linux on ours
[02:27:21] <gisburn> Triskelios: ouch
[02:27:31] * gisburn feels sorry for Triskelios
[02:28:28] <MindChild> I mean, is there something that comes with solaris or do I have to buy somethign
[02:28:29] <Triskelios> hey they're like tiny, relatively low power little boxes
[02:28:55] <Triskelios> MindChild: the supporting software for Solaris is downloadable for free
[02:28:56] <gisburn> boyd: why ?
[02:28:59] <Auralis> MindChild: just need to download the sunpci drivers from sun, they are free
[02:29:00] <Gman> steleman: haven't forgotten about your mail, just still thinking about it
[02:29:03] * gisburn pokes boyd
[02:29:06] <MindChild> ok, great
[02:29:07] <MindChild> thanks
[02:29:08] <Triskelios> I think it was last released for Solaris 9, but works fine on 10
[02:29:21] <steleman> hi Gman :-)
[02:29:27] <steleman> its ok i dont think it's an emergency
[02:29:53] <nachox> gisburn: because the current one is not exactly user friendly compared to that of modern linux distros afaik
[02:29:59] <Gman> steleman: no, but didn't want you to feel ignored ;)
[02:30:05] <boyd> gisburn: Sorry.. I was called away. To integrate with the new packaging system we need :)
[02:30:12] * boyd ducks
[02:30:18] <steleman> Gman lol ok i dont feel ignored
[02:30:19] <gisburn> ouch
[02:30:19] <richlowe> gisburn: the Installer is not yet open.
[02:30:33] <gisburn> boyd: really... does sun plan a new package system ?
[02:30:34] <nachox> boyd: :P
[02:30:37] <richlowe> gisburn: it's on the roadmap, and I think it's listed there as targetting sometime fairly soon.
[02:31:06] <boyd> gisburn: I said we need one. I don't know what sun plan
[02:31:33] <richlowe> You can't replace the current package system.
[02:31:38] <gisburn> ok... then I file two bugs, "want option to configure xntpd at install time" and "want optiion to move root's home dir to /root on user request"
[02:31:40] <richlowe> ok, well, obviously I can't speak for Sun either.
[02:31:49] <richlowe> but you can't do it, and they shouldn't.
[02:31:52] *** PerterB has quit IRC
[02:32:09] <boyd> richlowe: I don't think there is a need to replace it actually, I think it needs a layer on top of it
[02:32:12] <richlowe> you could extend it to add the features necessitated by a tool like apt.
[02:32:18] <boyd> (and enhancements to support said layer)
[02:32:20] <richlowe> you could fix the crappy performance.
[02:32:33] <richlowe> you may even be able to replace it in a way that retains compatibility (maintaining both package DB's sounds ugly though)
[02:32:34] <gisburn> richlowe: fixing the crappy performace is quite easy
[02:32:37] <nachox> i'm guessing the package management utilities are constrained by the same backwards compatibility issues that affect the rest of solaris
[02:32:37] <richlowe> but yanking it out just won't happen.
[02:32:51] <richlowe> nachox: not just the utilities, but the format of the packages.
[02:33:04] <richlowe> what's the point of having *binaries* from 2.5.1 still work if the packages containing them no longer install?
[02:33:22] <gisburn> heh
[02:33:30] <steleman> hi gisburn :-)
[02:33:39] <gisburn> steleman: Hi!
[02:33:52] <nachox> richlowe: makes sense :)
[02:33:57] <boyd> But supporting old packages does not preclude a new packaging system. witness, e.g. "alien" which allows rpms on debian using some kind of translation/wrapping
[02:34:13] <rydis> 2.5.1? Heck, T (Yale Scheme), compiled on SunOS 4, still runs fine on Solaris 10.
[02:34:18] <Auralis> in thory you just need a pkgtranslator to the new format
[02:34:27] <richlowe> you could in theory adjust things so that both systems use the same underlying data.
[02:34:28] <nachox> i'd love to have something like synaptic to handle sysv packages though
[02:34:34] <richlowe> but really, what people seem to want is an apt-a-like.
[02:34:46] <richlowe> package creation isn't particularly fun, but it's not as bad as it could be either.
[02:34:55] <richlowe> and there's no reason apt, or something like it, couldn't work over what we have already.
[02:35:00] <Gman> apt is only as good as the repository to back it
[02:35:03] <richlowe> the "apt means .deb" thing is a falacy.
[02:35:06] <nachox> thanks but no thanks, i dont like apt
[02:35:10] <richlowe> but one that persists for some reason.
[02:35:23] *** PerterB has joined #opensolaris
[02:36:06] <nachox> well, what i really dont like is debian's policy of a lot of very small packages, you need 4 packages to have a working cups
[02:36:20] <boyd> Why do you care?
[02:36:27] <boyd> The deps bring them in
[02:36:37] <boyd> Normally the separation is to allow optional functionality
[02:37:01] <boyd> such as client/server separation in pgsql. (not sure about what happens with cups)
[02:37:07] <nachox> i'm just a weird user i guess :)
[02:37:30] <boyd> Also, smaller packages allow smaller updates
[02:37:45] <Gman> amen to that :)
[02:37:59] <richlowe> which is part of why a piecemeal-upgrade of express doesn't make sense.
[02:38:08] <richlowe> smaller packages *could* allow that, in theory.
[02:38:14] <rydis> So /that's/ why I keep downloading X fonts over and over again on the Debian box. ;)
[02:38:18] <richlowe> (right now, if you look at package versions in the install images, nearly everything changes)
[02:38:25] <boyd> richlowe: I'm not sure how you do apt-like stuff without adding versioning to the underlying packages and ditching patches
[02:38:27] <richlowe> so your "small automatic upgrade" would still be a gig and a half.
[02:38:42] <nachox> hehe
[02:38:48] <Gman> you are looking at a development version of solaris though
[02:39:01] <richlowe> Gman: Yeah, as boyd said, patches complicate releases.
[02:39:09] <Gman> yeah
[02:39:10] <richlowe> you either refresh an entire package, which is sane with smaller packages.
[02:39:11] <boyd> Yeah, but surely 1.5G of stuff doesn't change every 2 weeks :)
[02:39:17] <richlowe> or you have to work out the patch v. package stuff.
[02:39:24] <Gman> the patch mechanism is pretty stretched
[02:39:31] <richlowe> boyd: not functionally, but effectively.
[02:39:36] <richlowe> boyd: every binary in ON "changes".
[02:39:38] <boyd> patch dependency is a freaking nightmare
[02:40:17] <sommerfeld> right, new build timestamp
[02:40:22] <Gman> it's also because we're not just delivering 'patches' anymore
[02:40:37] <Gman> rather huge chunks of new stuff in updates
[02:40:57] <richlowe> I think making software installation in general better is an important thing to do.
[02:41:04] <richlowe> but I don't think it's anything like as simple as is assumed.
[02:41:33] <Gman> yeah, tim marsland mentioned that as his personal priority in a recent all hands
[02:41:44] * Gman imagines it's pretty darn difficult
[02:41:53] <boyd> I assume that it's ridiculously hard... but I think it's time for a serious revamp
[02:45:20] * nachox is happy about hal being close to integration
[02:46:08] <boyd> Of the top of your head, kdoes anyone know if the "dbus shouldn't run in NG zones" bug is fixed
[02:46:57] <timeless> what's NG?
[02:47:02] <boyd> non-global
[02:47:09] * timeless was just going to guess that
[02:47:36] <ProfMikey> hmm
[02:51:05] *** agliodbs has quit IRC
[02:52:14] *** brendang has joined #opensolaris
[02:52:47] <ProfMikey> good night
[02:52:54] <Gman> nachox: i think it's integrated actually
[02:53:07] <Gman> artem got an award for his efforts
[02:53:11] <nachox> it should be there in vermilion b52
[02:53:23] <Gman> [which i think is a good thing given the pain he went through]
[02:53:33] <Gman> yeah, 2.16 planned for b53 though
[02:53:58] <boyd> Hi brendang!
[02:57:57] *** krozinov has quit IRC
[03:03:37] *** sommerfeld has quit IRC
[03:04:57] <twincest> does anone know off hand if studio has an equivalent to gcc __builtin_expect? (gcc exception to provide branch-prediction information to the compiler)
[03:05:13] *** krozinov has joined #opensolaris
[03:06:39] <gisburn> twincest: yes
[03:06:45] <gisburn> twincest: #pragma blabla
[03:07:00] *** jsubl2 has quit IRC
[03:07:12] <gisburn> twincest: timeless may know more
[03:07:21] <timeless> not offhand
[03:07:32] <twincest> that's okay, i can look it up if it does exist
[03:07:34] <twincest> thanks :)
[03:07:53] <gisburn> alanc-away: yt ?
[03:09:20] <timeless> http://mxr-test.landfill.bugzilla.org/mxr-test/seamonkey/source/xpcom/base/nscore.h#509
[03:09:29] <timeless> if mozilla knew about it, the bit would be there
[03:11:45] *** yongsun has joined #opensolaris
[03:18:06] *** nwf has joined #opensolaris
[03:19:32] * sparc-kly|WORK is away: Go to home :)
[03:24:48] <timeless> what does it mean to not inherit-pkg-dir?
[03:26:47] <timeless> #set up the zone normally, no inherit-pkg-dirs though
[03:29:29] *** yanyh has joined #opensolaris
[03:29:55] <delewis> timeless: by not inheriting the pkg-dir all packages installed in that pkg-dir will be copied rather than symlinked in the zone
[03:30:16] *** yanyh has quit IRC
[03:39:28] *** jsubl2 has joined #opensolaris
[03:41:42] *** Gman has quit IRC
[03:42:01] *** Gman has joined #opensolaris
[03:42:57] *** piwi has quit IRC
[03:49:44] *** nwf has quit IRC
[03:50:05] *** vmhobbes has quit IRC
[03:50:26] *** nwf has joined #opensolaris
[03:57:18] *** karrotx has joined #opensolaris
[03:58:10] *** bengtf has quit IRC
[03:59:08] <gisburn> Does anyone know how the "SCM Management " @ www.opensolaris.org works ?
[04:06:53] *** lloy0076 has left #opensolaris
[04:07:28] *** MindChild has quit IRC
[04:09:05] *** Tpent1 has quit IRC
[04:11:47] *** nyati has quit IRC
[04:35:20] *** karrotx has quit IRC
[04:35:23] *** karrotx has joined #opensolaris
[04:36:42] *** gisburn has quit IRC
[04:37:54] <brendang> Hi boyd
[04:38:01] <jamesd> hi brendang
[04:38:10] <brendang> Hi James
[04:40:20] <jamesd> brendang, i guess working at sun has kept you very busy
[04:41:00] <brendang> very busy catching up on regular sun things - how to do builds properly etc
[04:41:17] <brendang> I still need to install irssi somewhere
[04:41:25] <brendang> although I now have screen - so I'm getting there
[04:41:57] <brendang> gotta run for muni. I'll try to log in more often again :)
[04:42:00] *** brendang has quit IRC
[04:51:47] *** hyde has joined #opensolaris
[04:51:51] <boyd> Sheesh
[04:51:59] <boyd> :)
[04:53:12] <hyde> I am using OpenSSH_4.3p2 on solaris 10, everytime after login via ssh, it gives out 2 lines for "Sun Microsystems Inc.   SunOS 5.10      Generic January 2005", how to remove one line in this case?
[04:53:23] *** jsubl2 has left #opensolaris
[04:54:58] <boyd> I'd guess that you want to edit /etc/ssh/sshd_config, find PrintMotd and change it to "no", then svcadm refresh ssh
[04:56:19] <boyd> Oh, sorry.. OpenSSH... it's config file will be somewhere else
[04:56:56] <hyde> boyd: thanks it works. it's config file is at /usr/local/etc/sshd_config
[04:57:27] <boyd> Ouyt of interest, why are you using OpenSSH?
[04:57:48] <hyde> It was suggested by the developer, he just wants to use the latest version
[04:58:31] * boyd shrugs. Just wondering, since most people are happy with the SunSSH
[05:01:47] <Tpenta> sun ssh also works correctly with Sun PAM and auditing
[05:02:08] <boyd> and Kerberos
[05:02:27] <hyde> I want to use SunSSH as well, now I have to use this openSSH, and I am having problem configuring it with Sun PAM.
[05:02:34] <boyd> But I wish it would be updated to have some of the features of at least the client side of OpenSSH
[05:03:23] <hyde> I can't get this openSSH to work with Sun PAM for the last several hours.
[05:04:14] *** nwf has quit IRC
[05:04:15] <hyde> for ldap authentication
[05:05:24] <boyd> I don't know if it's possible to get OpenSSH doing that. May be time to ask the developer whay exactly he needs from OpenSSH. There may be another way
[05:07:20] <hyde> yes. but I googled for a while, it seems someone got it to work before, but not me. I will put in a few more hours on this, if still not possible, I will have to ask the developer
[05:10:06] *** jlc has joined #opensolaris
[05:14:07] *** halton has joined #opensolaris
[05:24:34] <hyde> btw, I just got openssh to work with Sun PAM. The key is to set this one:
[05:25:08] <hyde> ChallengeResponseAuthentication no
[05:28:07] *** Sieghard has joined #opensolaris
[05:34:11] <hyde> for rhel, I can mount with acl option in order to use setfacl, is this the same for solaris 10 in order to use setfacl?
[05:35:01] <delewis> hyde: no
[05:35:10] <delewis> UFS/ZFS support ACLs by default
[05:35:17] <delewis> no need to mount them with some arbitrary option
[05:35:59] <hyde> thanks delewis
[05:55:05] *** OnkelSchorsch_ has left #opensolaris
[05:56:14] *** laca has quit IRC
[05:59:29] *** nyati has joined #opensolaris
[05:59:54] *** libkeiser has joined #opensolaris
[06:00:30] *** qbit has joined #opensolaris
[06:05:13] <hyde> exit
[06:15:50] *** hyde has quit IRC
[06:19:16] *** besonen_mobile_ has quit IRC
[06:19:21] *** solarisjon has joined #opensolaris
[06:19:27] *** solarisjon has left #opensolaris
[06:24:36] *** jlc has quit IRC
[06:36:24] *** nachox has quit IRC
[06:38:05] *** Sieghard has quit IRC
[06:41:52] *** gtc has joined #opensolaris
[06:42:15] *** unixconsole has left #opensolaris
[06:43:23] *** sahafeez has joined #opensolaris
[06:45:18] *** sahafeez has quit IRC
[06:48:25] *** ShadowHntr has quit IRC
[07:00:22] *** benr has quit IRC
[07:35:02] <sickness> http://www.sun.com/2006-1017/feature/curtain.jsp <- anyone knows what this should be? =)
[07:35:33] <Error_404> my guess is something to do with honeycomb
[07:35:46] <sickness> tnx
[07:38:49] <Error_404> remember, however, that i have no idea what i'm talking about
[07:39:11] <sickness> oh, tnx, I've no idea too but I'm just curious :)
[07:39:14] <Error_404> and have never so much as been within 2000 miles of sun's campus
[07:45:04] *** ShadowHntr has joined #opensolaris
[07:50:13] *** sparc-kly__ has joined #opensolaris
[07:50:27] *** GigaE has joined #opensolaris
[07:54:30] *** nyati has quit IRC
[07:57:50] *** bengtf has joined #opensolaris
[08:08:19] <asyd> \_o<
[08:08:25] <trygvis> morning o/
[08:11:37] *** sparc-kly_ has quit IRC
[08:11:38] *** dunc has quit IRC
[08:18:48] *** lplatypus has quit IRC
[08:25:09] *** lplatypus has joined #opensolaris
[08:28:17] *** slowhog_home has quit IRC
[08:28:50] <boyd> Of course. I download nexenta, and then I see that there's a new release minutes from appearing :(
[08:28:57] <asyd> :)
[08:29:13] <timeless> boyd: hey, since you're playiing w/ nexenta, maybe you could help?
[08:29:17] <timeless> http://pastebin.mozilla.org/877 http://pastebin.mozilla.org/878
[08:29:36] <timeless> i'm trying to get nexenta into a zone, and one of the packages is burping :(
[08:29:52] <boyd> Well, I'm *not* playing with it, since my download failed. I think they broke it while preping the new rev
[08:30:16] <timeless> what do you mean broke?
[08:30:19] <boyd> Strangely the reason I'm fetching it is to try to put it in a zone though. Are you following me?
[08:30:23] <timeless> as in broke the pacakging?
[08:30:32] <timeless> not intentionally
[08:30:33] <boyd> timeless: The download failed
[08:30:45] <timeless> boyd: which download?
[08:30:56] <boyd> Of nexenta.
[08:31:01] <timeless> yours?
[08:31:17] * timeless sighs
[08:31:28] <boyd> yes. mine.
[08:31:49] *** gtc has quit IRC
[08:31:52] <boyd> Maybe the new release has caused some packages to be in a state of flux.
[08:31:57] <boyd> See the top news item here http://www.gnusolaris.org/gswiki
[08:32:35] *** Auralis has quit IRC
[08:34:48] *** Auralis has joined #opensolaris
[08:34:55] <timeless> i'm trying to follow http://www.gnusolaris.org/gswiki/Nexenta_Zones_under_opensolaris
[08:36:04] <timeless> fwiw my get_nexenta.sh is http://pastebin.mozilla.org/879
[08:36:54] <boyd> It was wierd. it got nearly to the end, then stopped and started again. Even though my d/l manager should have resumed.
[08:37:40] <boyd> timeless: Someone was in here a while ago who developed that procedure... I think there was some problem then. lemme look back
[08:37:54] <timeless> thanks
[08:38:40] <timeless> anyway, w/ the script changed to include /opt/sfw/bin in PATH (to find wget)
[08:38:51] <timeless> i actually did manage to pull a number of packages
[08:39:14] <trygvis> try giving it /usr/sfw instead of /opt/csw
[08:39:32] <timeless> oh, you're here too
[08:39:37] <timeless> boy, that makes life so much easier :)
[08:40:06] <trygvis> :D
[08:40:15] <boyd> It was "Justin" who was talking about it....
[08:40:37] <timeless> trygvis: no change
[08:41:26] <timeless> so, i have a install/root/debootstrap/debootstrap.log
[08:42:35] *** ShadowHntr has quit IRC
[08:42:59] <timeless> i think the main problem is that http://apt.gnusolaris.org/dists/elatte-unstable/main/binary-solaris-i386/Packages doesn't have Package: base-config
[08:43:28] <timeless> i presume it should if i'm expecting this process to "work"
[08:43:32] <timeless> does that seem "reasonable"?
[08:43:39] <trygvis> yah
[08:43:54] <boyd> Nah... looks like stevel was having a different problem
[08:44:14] <timeless> ok, so, where could i buy a base-config? :)
[08:44:21] <richlowe> boyd: what'd he break?
[08:44:24] <richlowe> Oh.
[08:44:31] <trygvis> timeless: it might be in the stable tree
[08:44:32] * richlowe acquires context, goes into hiding
[08:45:12] <boyd> richlowe: Heh... you do what I do... you search for context like dynamically scoped variables :)
[08:47:33] *** Cyrille has joined #opensolaris
[08:47:42] <timeless> are there mirrors for apt.gnuesolaris?
[08:47:50] <timeless> s/gnue/gnu/
[08:47:54] <timeless> (too much zork?)
[08:48:14] <trygvis> hehe
[08:48:20] <trygvis> let me know if you find a mirror
[08:49:16] <timeless> http://apt.gnusolaris.org/dists/elatte-testing/main/binary-solaris-i386/Packages doesn't list Package: base-config either
[08:49:37] <trygvis> try s,stable,unstable,
[08:49:57] <timeless> http://apt.gnusolaris.org/dists/elatte-unstable/ doesn't have Pacakges
[08:50:02] <timeless> or Packages for that matter
[08:50:47] <timeless> all this just so that i can use apt to get sources
[08:50:58] <timeless> i'm pretty sure i'm already *using* an apt which means this is really silly :)
[08:52:37] <timeless> trygvis: do you happen to have nexenta running?
[08:53:01] <trygvis> yep
[08:53:27] <trygvis> but I only have base-files
[08:53:38] <timeless> ...
[08:54:04] <timeless> what's the simplest way to force apt to ignore a missing file?
[08:54:48] <trygvis> not possble, I'm sure the script is forcing it to be install
[08:54:54] <trygvis> s,forcing,requesting,
[08:54:57] <trygvis> try changing the script?
[08:55:25] <timeless> given that Packages references it dozens of times, that seems futile
[08:55:42] <timeless> how would i make an "empty" package named base-config?
[08:56:03] <trygvis> hm, there are a few utils to make meta packages but I just can't remember their names
[09:05:28] *** Dar_HOME is now known as Dar
[09:09:14] *** rachel has joined #opensolaris
[09:09:19] *** bengtf has quit IRC
[09:09:35] *** SymmDC has joined #opensolaris
[09:09:58] *** SymmDC is now known as SymmWork
[09:19:02] *** mazon is now known as Mazon
[09:21:37] *** Symmetria has quit IRC
[09:31:31] *** sparc-kly_ has joined #opensolaris
[09:39:17] *** sparc-kly__ has quit IRC
[09:43:59] *** damienc has joined #opensolaris
[09:55:53] *** sahafeez has joined #opensolaris
[10:00:03] *** benr has joined #opensolaris
[10:00:32] *** ChanServ sets mode: +o benr
[10:10:05] *** Gman has quit IRC
[10:13:53] <GigaE> when building a module containing a call to gettimeofday I get the following error:  implicit function declaration: gettimeofday()
[10:14:14] <Cyrille> that's usually just a warning.
[10:14:19] <GigaE> but i do have #include <sys/time.h> and #include <time.h>
[10:15:13] <SymmWork> hrm dumb question, but how do I configure multiple ips on an interface and save it so that it boots like that (I understand how to configure a single ip using the hostname.interface in etc, but what about multiple ips
[10:15:16] <GigaE> Cyrille, should I use it with gettimeofday(&tv1) or gettimeofday(&tv1,NULL) ? any idea?
[10:15:54] <Cyrille> man gettimeofday says &tv1, NULL
[10:17:31] <Cyrille> multiple IP on a single interface?
[10:17:38] <GigaE> Cyrille, thnks; I assume this is true also to a kernel modul
[10:18:17] <SymmWork> Cyrille yeah as in ip aliases basically
[10:18:31] <SymmWork> so you have 196.21.20.2 20.3 20.4 20.5 all on your e1000 card
[10:25:41] <libkeiser> SymmWork: i can't remember the exact syntax in the hostname.* file, but I know they explain how to do it in the old sun blueprints document on IPMP
[10:26:52] *** bengtf has joined #opensolaris
[10:27:02] <SymmWork> yeah I found it
[10:27:09] <Cyrille> SymmWork: man ifconfig tells all about that (addif and other bits)
[10:27:14] <SymmWork> Im gonna have to do this with a new init script though, because the method of doing this really really doesnt scale
[10:27:28] <SymmWork> you can add hostname.interface:1 :2 :3
[10:27:29] <SymmWork> in etc
[10:27:38] <SymmWork> but errr when you gotta put 300 interfaces
[10:27:44] <SymmWork> I mean addresses
[10:27:46] <SymmWork> that gets messy
[10:27:47] <SymmWork> :p
[10:27:58] <SymmWork> so a new rc script it is :)
[10:33:04] *** sahafeez has quit IRC
[10:35:38] *** yongsun has quit IRC
[10:48:53] *** AbeFroman has quit IRC
[10:55:09] <Vanuatoo> Come back at 12:00 pm PST to experience an  extraordinary simple idea.
[10:55:16] <Vanuatoo> how PST is translated to GMT?
[10:56:17] <nightswim> pacificic, pakistan or phillipines standard time?
[10:56:22] <LeftWing> SymmWork: You can also put addif commands in /etc/hostname.* files from memory, without creating 300 files.
[10:56:35] <nightswim> pacific st is utc-7
[10:56:42] <nightswim> uuhm, -8 even
[10:57:49] <Cyrille> http://www.timezoneconverter.com/cgi-bin/tzc.tzc
[10:58:26] *** Gman has joined #opensolaris
[10:58:38] <SymmWork> hrm
[10:58:49] <SymmWork> LeftWing Im busy writing a nice init script :)
[10:59:02] <SymmWork> just trying to work out the proper syntax for removeif for the stop part of the script
[10:59:06] <LeftWing> heh
[10:59:17] <LeftWing> You could always just use one of the supported methods. ;P
[10:59:51] *** Dar is now known as Dar_AFK
[11:00:48] <SymmWork> LeftWing *shrug* this will make its a hell of a lot easier for me to maintain afterwards, basically what Im doing is relying on /etc/ipaliases.interfacename which has a list of ip addresses, netmasks and potentially other info in it
[11:01:03] <SymmWork> can just throw whatever in there and my script will do its own thing off a list :p
[11:01:05] <SymmWork> nice and simple
[11:01:06] <SymmWork> :p
[11:01:22] <LeftWing> You realise that that's practically what the nice and simple existing scripts do?
[11:01:59] <SymmWork> LeftWing /etc/hostname.e1000g I was told could only have one address in it and I needed to add :1 :2 :3 :4 files to add the aliases
[11:02:11] <SymmWork> *shrug* and thats what Im reading on the web
[11:02:19] <LeftWing> I'm fairly certain you've been misinformed.
[11:02:28] <Cyrille> so if it's written on the intarwebs it must be true.
[11:03:00] <SymmWork> Cyrille *shrug* then I'll try the other method and see what happens
[11:03:30] <LeftWing> If you throw new lines into your hostname.e1000g of the form "addif <blah>" it should work.
[11:03:52] <SymmWork> leftwing how will the shutdown script to stop the networking handle that?
[11:04:16] <LeftWing> It'll bring down the box? =P
[11:05:02] <SymmWork> errr heh, thats the other reason Im writing my script, so I can do a stop on the virtual interfaces
[11:07:01] *** Mazon is now known as mazon
[11:07:29] *** kimc has joined #opensolaris
[11:13:24] *** AbeFroman has joined #opensolaris
[11:16:21] <kimc> good morning
[11:18:10] <PerterB> 'owdo
[11:33:34] *** eokyere has joined #opensolaris
[11:35:21] <SymmWork> sweet, heh, simple init script that reads a file called /etc/ip-aliases, adds virtual interfaces and can also shut them all down with a stop command, and works with cidr masking IF the ip aliases are in that file or not
[11:35:31] <SymmWork> oops I mean if the ip aliases have a cidr mask on them or not
[11:37:07] <kimc> nice
[11:38:08] *** coffman has joined #opensolaris
[11:38:38] <SymmWork> heh will pastebin the script in a secon
[11:39:57] <SymmWork> http://pastebin.ca/206210
[11:40:00] <SymmWork> there ya go if anyone wants it
[11:40:09] <SymmWork> very simple :)
[11:40:56] <SymmWork> heh now to make it handle vlans as well
[11:47:29] *** Griff has joined #opensolaris
[11:47:30] *** Griffous has quit IRC
[12:04:44] *** MattMan has joined #opensolaris
[12:13:26] *** Gman has quit IRC
[12:13:55] *** Gman has joined #opensolaris
[12:14:40] <SymmWork> heh, got my vlan start script working nicely, now just integrating my virtual-ip add script to be able to handle vlan interfaces nicely
[12:14:59] *** benr has quit IRC
[12:28:33] *** jcea has left #opensolaris
[12:29:38] *** lplatypus has quit IRC
[12:44:19] <SymmWork> bang :p all done
[12:45:10] <timeless> fwiw http://www.debianhelp.org/node/1169 mentions base-config
[12:49:15] <SymmWork> http://pastebin.ca/206295
[12:49:19] <SymmWork> for anyone who is interested
[12:49:33] <SymmWork> heh, nice clean way to handle a ton of vlans and a ton of ip addresses together
[12:49:46] <SymmWork> well, heh, I find it kinda clean and simple :)
[12:52:34] *** yongsun has joined #opensolaris
[12:52:49] *** stefanjo has quit IRC
[12:53:37] *** Ireul has joined #opensolaris
[12:54:10] <jteo> moin.
[12:54:19] <SymmWork> lo jteo :)
[12:54:31] <jteo> SymmWork: hello.
[12:54:46] <SymmWork> heh Im enjoying solaris now :p Its actually motivating me to do things I havent done in years, like a little bit of scripting and stuff
[12:55:09] * SymmWork never finds time for this stuff any more :( *sigh* I kinda miss my serious techie days before moving up the ladder
[12:55:41] <SymmWork> heh my mirror server now *nicely* and *cleanly* configures all its interfaces :)
[12:55:50] <SymmWork> and brings them down nicely as well
[12:55:51] <SymmWork> :p
[12:57:29] <jteo> SymmWork: the joys of scripting. ;)
[12:57:59] <SymmWork> heh yeah, like I say, just dont find much time anymore, usually end up delegating it to people further down the chain
[12:58:46] *** besonen_mobile has joined #opensolaris
[13:00:19] *** stefanjo has joined #opensolaris
[13:04:29] *** yongsun has joined #opensolaris
[13:07:05] *** coffman has quit IRC
[13:09:14] <GigaE> question about gettimeofday() in solaris : I am getting strange results for 2 consecutive gettimeofday() calls
[13:09:46] <GigaE> I am printing it with cmn_err(CE_CONT,"tv2.tv_sec  %ul \n", tv2.tv_sec);
[13:10:22] <GigaE> and the diffence is seconds is more than 3000 seconds from the previous call !
[13:10:31] <GigaE> but only 2-3 seconds elapsed !
[13:11:48] <jteo> blame the scheduler.
[13:12:10] <GigaE> is there something which should be done between 2 consecutive calls to gettimeofday ?
[13:12:16] <GigaE> in linux it works !
[13:12:35] <GigaE> jteo, really, any idea what should I do ?
[13:13:30] <SymmWork> heh GigaE this isnt linux though, heh, thats always driven me crazy when people go "It works in blah", I can handle people saying "Whats the equivelant way to do Y if I would do X in linux to make Y happen", but "It works in linux!" drives me nuts :)
[13:14:08] <jteo> GigaE: i'm guessing it's the scheduler. anyway, what are you trying to accomplish?
[13:14:11] <GigaE> well , blaming the scheduler will not solve my problem...
[13:14:26] <timeless> use dtrace instead of calling gettimeofday yourself?
[13:14:33] <jteo> timeless: precisely.
[13:15:17] <timeless> reinventing some minimal portion of dtrace (and badly), when dtrace is available really makes no sense
[13:15:25] <timeless> you're on one of the best platforms available
[13:15:48] <GigaE> I saw that in the kernel there is usage of gettimeofday
[13:17:07] <SymmWork> heh finally ready to put solaris 10 into production, everything working 100% from ftp daemon to http daemon to vlans to virtual ips to iSCSI san's
[13:17:08] <SymmWork> yay
[13:17:31] <jteo> GigaE: what are you trying to accomplish?
[13:17:35] * SymmWork schedules time for getting rid of linux when he gets back from presenting stupid papers at stupid conferences in senegal
[13:17:54] <GigaE> jteo, performance metrics
[13:18:06] <jteo> GigaE: measure it using dtrace.
[13:18:07] <jmcp> GigaE: instead of gettimeofday() use gethrtime()
[13:18:07] <GigaE> jteo, I am sure it can be done in dtrace
[13:18:37] <jteo> jmcp: ah!
[13:18:41] <GigaE> But it will take me time to learn it
[13:19:13] <GigaE> unless someone can tell me if it is straight away: how to measure entry and exit from a method in a module I wrote
[13:19:23] <GigaE> jmcp, I will try
[13:19:40] <jmcp> dtrace is the correct tool for that purpose
[13:20:46] *** yarihm has joined #OpenSolaris
[13:20:57] * timeless tries to figure out how to change a zone from 'configured' to 'installed'
[13:23:16] *** phips has quit IRC
[13:23:36] <jteo> GigaE: dtrace -n 'pid$target:a.out:myfunc:entry { self->flag=timestamp}' -n 'pid$target:a.out:myfunc:return /self->flag/ {trace(timestamp-self->flag); self->flag=0}'
[13:23:47] *** phips has joined #opensolaris
[13:23:51] <jteo> on and add on a '-c <your program>'
[13:23:56] <GigaE> jteo, thnks !!! I will try it
[13:24:00] <timeless> jteo: are you using a cookbook or writing these by hand?
[13:24:03] <SymmWork> hrm whats the fastest way to generate a 2gig file of random crap?
[13:24:06] <jteo> timeless: by hand.
[13:24:17] <twincest> symm: dd
[13:24:21] <jteo> SymmWork: download something off a P2P app.
[13:24:23] <jteo> ;)
[13:25:07] <GigaE> the pid in a case of a kernel module - is it the module id ?
[13:25:17] <GigaE> I am working with a kernel module...
[13:25:32] <timeless> symmork: i have some nice xpcshell core dumps that are 2g :)
[13:26:24] <jteo> GigaE: even simple then. it should be something like "fbt:modulename:myfunc:entry"
[13:26:47] <jteo> GigaE: go take a read at the Dtrace Guide. it's easy to pick up.
[13:27:00] <GigaE> jteo, thnks
[13:27:59] <timeless> ouch
[13:28:21] <timeless> my nexenta zone transitioned to maintenance mode
[13:30:38] <SymmWork> hrm, anyone here recommend a good way to benchmark disk speed in solaris?
[13:32:03] <SymmWork> cause generating files off dd seems crap slow :p it only wrote 2gigs in 2 minutes and 55 seconds
[13:32:48] <jmcp> SymmWork: don't bother with dd, it's single-threaded
[13:32:57] <jmcp> find iozone or vxbench instead
[13:33:15] * jteo takes notes.
[13:33:44] *** lloy0076 has joined #opensolaris
[13:35:26] <quasi> or filebench?
[13:36:10] <SymmWork> heh installing iozone now quick
[13:39:22] <lloy0076> When a snapshot is taken using "fssnap(1M)", I take it that one might possibly get away with a backing storage file that is less than the original file size...that is the backing storage file only keeps "changes" and isn't a full copy of what you've snap shotted...
[13:39:42] <lloy0076> I meant "file system size" not file size.
[13:40:16] *** yongsun has quit IRC
[13:43:29] <SymmWork> heh
[13:43:38] <SymmWork> DAMN iSCSI HBA offload makes a *HUGE* difference
[13:43:51] <SymmWork> iSCSI straight running iozone with the SAN linked over normal IP = 70% cpu utilization
[13:43:58] <SymmWork> iSCSI on an HBA offload = 6% cpu
[13:44:08] <SymmWork> huge difference in speed as well
[13:44:21] <quasi> for the same amount of io?
[13:44:26] <SymmWork> yeah
[13:44:32] <quasi> not bad
[13:44:43] <quasi> which hba?
[13:44:50] <SymmWork> thats the qlogic 4050C
[13:45:04] <SymmWork> apool                  3.8T   2.1G   3.8T     1%    /mirror/iscsi-san-1
[13:45:09] <SymmWork> :)
[13:45:41] <SymmWork> thats the test one, I've got 3 live ones that will go onto the final mirror.ac.za machine that are identical
[13:45:45] <SymmWork> on 3 seperate HBA's
[13:46:42] *** lloy0076 has left #opensolaris
[13:46:58] <SymmWork> heh this san makes a bloody noise with all its fans
[13:50:09] *** salamanders has joined #opensolaris
[14:02:38] *** Dar_AFK is now known as Dar
[14:17:18] * timeless chuckles
[14:17:23] <timeless> somehow i don't think this is correct:
[14:17:26] <timeless> +++ /export/home/zones/solaris/nexenta/install/root/bin/gzexe   Tue Sep 19 11:03:37 2006
[14:17:33] <timeless> +# On Ultrix, /bin/bash is too buggy, change the first line to: #!/bin/bash5
[14:18:33] *** axisys has joined #opensolaris
[14:19:14] <jmcp> timeless: have you ever used Ultrix?
[14:19:20] <timeless> no
[14:19:25] <timeless> have you ever used /bin/bash5 ?
[14:20:00] <jmcp> yeah
[14:20:03] <jmcp> a long time ago
[14:20:04] <jmcp> it wasn't pleasant
[14:20:11] <timeless> bash5 or ultrix?
[14:20:25] <axisys> if u ought to pick a OS to run BIND that can handle 8000 qps and can manage spam using IPF and can manage anycast using bgpd/zebra.. which one would u pick? sol 10 u2 or Freebsd 6.x?
[14:20:55] <jmcp> timeless: ultrix
[14:20:59] <axisys> 8000 may grow to 10000.. btw the hardware is X2100
[14:21:02] <PerterB> Ultrix was pretty unpleasant
[14:21:19] <jmcp> axisys: I'd be going sol1ou2 - I have more confidence in the networking stack
[14:21:34] <jmcp> which - initially at least - would appear to be where most of your load would come from
[14:22:24] <axisys> we have been doing some testing between the two OS before a major upgrade all cross the country.. i am leaning towards sol10u2 as well
[14:22:38] <timeless> jmcp: you understand my problem w/ that line, right?
[14:22:48] <timeless> i don't mind something complaining that ultrix sucks
[14:22:55] <timeless> you sonds justified and i certainly believe you
[14:23:03] <timeless> s/onds/ound/
[14:23:08] <axisys> some argues that bgpd/zebra is better suited in freebsd
[14:23:26] <axisys> i personally not sure how well it stacks up in sol 10 u2
[14:23:31] <jmcp> timeless: it's an interesting conundrum - how much support do you want to leave in for an OS that is *very* dead and buried?
[14:23:51] <jmcp> axisys: do a pro/con table for both, and add weights
[14:23:55] <timeless> jmcp: oh, i just don't like global search+replaces
[14:24:07] <jmcp> timeless: they can be dangerous, sure
[14:24:14] <timeless> it's like when collabnet forked bugzilla
[14:24:23] <timeless> they ran s/bug/issue/g
[14:24:35] <axisys> i kinda.. let me see if i can make it available w/o disclosing sensitive stuff
[14:24:36] <timeless> but there was a story about Zarro boogs found
[14:24:45] <timeless> and they kinda messed up the story w/ their s/r
[14:24:45] <axisys> jmcp: one sec
[14:25:01] <jmcp> timeless: haha :)
[14:26:13] <SymmWork> heh I'll find out in a bit how well quagga/bgpd handles in solaris 10, I need it for announcing AS112 stuff back to my routers
[14:26:15] <SymmWork> along with bind
[14:27:21] <SymmWork> I currently do that under linux, its another thing I wanna migrate (bgp session running to my router that announces the route-server address of the server used to handle RFC1918 addresses so that all RFC1918 resolve requests dont go out onto the internet)
[14:27:33] <SymmWork> and then a bind instance to quickly reject those
[14:28:24] <axisys> jmcp: we have mostly graphs.. one key note is FreeBSD using less CPU than Sol 10 u2
[14:28:39] <jmcp> axisys: exactly same hardware?
[14:28:44] <jmcp> mrtg sorta thing>?
[14:29:29] <axisys> jmcp: exactly same hardware
[14:29:36] <axisys> jmcp: mrtg kind a
[14:29:43] <jmcp> ok, and was freebsd using all the available cpus?
[14:30:09] <axisys> jmcp: pcap stats shows freebsd can take highe traffic
[14:30:25] <jmcp> have you done any tuning of either network stack?
[14:30:27] <axisys> freebsd was using single threads
[14:30:43] <axisys> multi thread freebsd on bind perform bad
[14:30:50] *** nwf has joined #opensolaris
[14:30:59] <axisys> not tuning on sol 10 u2 side.. just out-of-stock
[14:31:20] <jmcp> ok, so be fair to both stacks and tune them, then re-run your tests
[14:31:23] <axisys> there were quite a bit of work done on dgp traffic injection process
[14:31:36] <axisys> s/dgp/bgp/
[14:32:17] <axisys> jmcp: sol 10 u2 does most change dynamic ?! so no tuning done
[14:32:32] <jmcp> you can use ndd, generally
[14:32:41] <jmcp> and as far as I recall it's dynamic
[14:32:54] <jmcp> I'll find the right ref for you ....
[14:33:32] <jmcp> http://docs.sun.com/app/docs/doc/817-0404 Solaris Tunable Parameters Reference Manual
[14:34:12] <asyd> hmm I guess log of "Bus fault" messages when I run probe-scsi-all  is not a good thing ? ;p
[14:34:30] <jmcp> asyd: setenv auto-boot?=false ; reset-all ; probe-scsi-all
[14:34:40] <asyd> ok
[14:34:41] <jmcp> s/;/[enter key]
[14:36:52] <axisys> jmcp: check this out http://silenceisdefeat.org/~iqbala/Screenshot.png
[14:37:12] *** apa has joined #opensolaris
[14:38:04] <axisys> jmcp: thnx for that doc btw
[14:38:09] <jmcp> you're welcome
[14:38:20] <jmcp> have you got a better resolution image? I'd like to read the axes a bit more closely
[14:39:27] <axisys> jmcp: let me see
[14:45:02] *** Griff has quit IRC
[14:45:07] * jmcp creates a kitchensink zone
[14:45:14] <timeless> jmcp: yum
[14:45:26] <timeless> http://www.mozilla.org/docs/web-developer/samples/kitchensink.xml
[14:45:29] <jmcp> it's gonna be my dns, jet and apache/tomcat/roller server
[14:45:34] * jmcp chuckels
[14:45:44] <jmcp> that's a yokel-y chuckle :)
[14:47:00] <timeless> nexenta# shutdown
[14:47:13] * timeless sighs
[14:47:23] <timeless> i have to confirm Do you want to continue?
[14:47:34] <jmcp> no!!!!!
[14:47:45] <timeless> Do you want to continue? (y or n):   Broadcast Message from root (console) on nexenta Tue Oct 17 05:32:03...
[14:47:46] <timeless> False Alarm:  The system nexenta will not be brought down.
[14:47:47] * jmcp utters the movietone scream of despair
[14:47:51] <jmcp> :)
[14:49:08] <timeless> svc.startd: The system is coming down for administration.  Please wait.svc.startd: Killing user processes: Requesting System Maintenance Mode
[14:49:09] <timeless> (See /lib/svc/share/README for more information.)
[14:49:39] *** laca has joined #opensolaris
[14:50:49] <jmcp> it's aka single user mode when booted from the known-good grub archive, *and* aka single user mode when booted without any services running
[14:51:12] <timeless> yeah, i understand what it would be
[14:51:19] * jmcp feels the floorboards underfoot while the u20m2 does the initial svccfg import for kitchensink
[14:51:20] <jmcp> hmmmm
[14:52:34] *** LeftWing has quit IRC
[14:52:44] *** LeftWing has joined #OpenSolaris
[14:53:04] <timeless> but shouldn't the docs actually include items based on how they're referenced?
[14:53:10] * jmcp mutters
[15:00:47] <axisys> jmcp: pulling the image out of pdf was lot of work.. thanks to pdfimages tool
[15:00:56] <jmcp> wow
[15:01:34] <axisys> jmcp: here are four graphs http://silenceisdefeat.org/~iqbala/solaris-cpu.jpg , solaris-pcap.jpg, freebsd-pcap.jpg, freebsd-cpu.jpg
[15:01:40] <jmcp> thanks
[15:02:26] <jmcp> that's certainly an interesting spike in solaris-cpu between 80 and 97 seconds
[15:02:37] <jmcp> what do the colours represent?
[15:02:47] <Cyrille> different wavelengths?
[15:02:49] <Cyrille> sorry.
[15:02:52] * jmcp snorts
[15:02:56] <jmcp> nice on Cyrille :)
[15:03:15] *** lloy0076 has joined #opensolaris
[15:05:48] <axisys> jmcp: that i can tell u in a sec
[15:07:33] *** lloy0076 has quit IRC
[15:08:04] <axisys> jmcp: orange - user, purple - system, green - idle cpu usage
[15:08:19] <jmcp> cool
[15:08:26] <asyd> jmcp: thanks for the reset-all
[15:08:31] <asyd> working better now
[15:08:34] <jmcp> asyd: you're welcome :)
[15:08:37] <jmcp> yeah, it will do
[15:08:51] <asyd> but I wonder why I need to do that on a new box
[15:09:21] <jmcp> asyd: do what? reset-all? because that's part of what OBP requires
[15:09:36] <jmcp> it's been that way for more than a decade
[15:09:41] <asyd> ok :)
[15:09:44] <axisys> jmcp: cyan - pkts in , blue - queries in, green - responses, purple - total load of PCAP
[15:10:14] *** laca is now known as lacaAFK
[15:10:42] <jmcp> axisys: for the pcap results, both freebsd and solaris look really close to me - near enough as makes no effective difference
[15:11:31] <jmcp> for the cpu usage .... both OSes are spending the majority of their time in user+idle - system is constant-ish
[15:12:00] <jmcp> so really it comes down to "why is the %user vs %idle being looked at as the metric?"
[15:12:07] <axisys> jmcp: that is kind a our conclusion is .. hehe
[15:12:51] <jmcp> I guess if you could massively ramp up the load - perhaps to 10x what you've measured already? - then you might begin to see a statistically significant difference
[15:12:52] * jmcp shrugs
[15:12:53] <axisys> jmcp: too see how much more it will be able to take.. as a future growth
[15:17:50] <jteo> why oh why do i always pick seemingly simple problems...
[15:18:02] <jmcp> jteo: you're too trusting
[15:18:22] *** bengtf has quit IRC
[15:19:21] <jteo> jmcp: true. i suspect i think groupwise is simple too. ;)
[15:19:46] <jmcp> jteo: ooooh ... that's a bit below the belt
[15:19:59] <jteo> jmcp: heh. ;)
[15:20:17] <quasi> froupwise is the devil! ;)
[15:20:18] <jmcp> jteo: on gropewise, apparently my current employed is upgrading to v7.1 shortly
[15:20:34] <jteo> jmcp: and that's a new level of hell?
[15:21:00] <jmcp> I don't know. that's part of the problem
[15:23:43] <jteo> jmcp: better the devil you know...
[15:23:48] * jmcp glares
[15:24:12] *** Kernel86 has quit IRC
[15:24:20] <eugene> jmcp: hey good news
[15:24:35] <jmcp> eugene: yeah?
[15:24:40] <eugene> jmcp: there is 80% chance that we will be eating dinner in Sydney next January :)
[15:24:57] <jmcp> neat!
[15:24:57] *** eboutilier has joined #opensolaris
[15:25:02] *** eboutilier has left #opensolaris
[15:25:02] <jmcp> I'll just check my weekends ...
[15:25:12] <eugene> jmcp: :) I should be there on 14 to 21 Jan
[15:25:19] <jmcp> ok
[15:25:50] <jmcp> I have a 30th to go to in Brisbane on the 19/20th, but apart from that we should definitely make plans to catch up
[15:26:03] <eugene> jmcp: awesome
[15:26:12] <eugene> jmcp: if you are attending linux.conf.au, that's even better :P
[15:27:07] <jmcp> probably won't be, but you never know
[15:27:28] <eugene> no worries
[15:29:13] *** nwf has quit IRC
[15:36:48] *** karrotx has quit IRC
[15:39:03] <SymmWork> heh I'll be in canberra in december
[15:39:08] <SymmWork> mauritius in november
[15:39:10] <SymmWork> senegal next week
[15:39:25] <SymmWork> new york in april, estonia in may
[15:39:28] <SymmWork> heh gah it never stops
[15:39:36] <SymmWork> brb
[15:40:36] <axisys> jmcp: u think sol10u2 may shine if we ramp up the load?
[15:42:07] <jmcp> axisys: yes
[15:42:28] <axisys> jmcp: what are default parameters (there are tons in that tuning doc) that u may think should be tuned on sol side
[15:42:49] <axisys> jmcp: stuffs related to network stack ?
[15:42:51] <jmcp> I'd be looking at the hiwatermarks for tcp
[15:42:55] <jmcp> yeah, of course
[15:43:03] <axisys> jmcp: gotcha.. anything else?
[15:43:10] <jmcp> I'd also look at the time/close wait connection times
[15:43:19] <jmcp> apart from that ... dunno. I'm not a network engineer
[15:43:29] <axisys> jmcp: note taken
[15:43:42] <jmcp> I'm a storage kinda guy :)
[15:43:45] *** besonen_mobile has quit IRC
[15:44:20] <axisys> jmcp: talking about storage we had few issues lately w/ 5310
[15:44:27] * jmcp runs away screaming
[15:44:37] *** bengtf has joined #opensolaris
[15:45:31] <axisys> lol
[15:45:38] <eugene> SymmWork: i know how that feels. i have been flying a lot too
[15:46:04] <axisys> jmcp: sun made us switch to 5310 from netapp.. :-(
[15:46:14] <jmcp> axisys: *sun* made you do that?
[15:46:50] <axisys> jmcp: well *cough* kind a .. w/ the help of one of our "decision maker"
[15:46:58] * jmcp grins
[15:47:30] <jteo> manager?
[15:47:40] <quasi> jmcp: they must have really twisted the arm of that "decision maker" by being so much cheaper ;)
[15:47:41] <axisys> jmcp: bingo!
[15:47:58] <axisys> quasi: hehe
[15:48:27] *** glagasse has joined #opensolaris
[15:48:38] * quasi hears "rumors" that the 5310/5320s go for about half the price of comparable netapp
[15:48:39] <axisys> now there will be purchase of 5320
[15:49:07] <axisys> quasi: plus support cost loooot cheaper
[15:49:08] <AbeFroman> you very rarely get more than you paid for
[15:49:13] <AbeFroman> (to paraphrase)
[15:49:30] <axisys> AbeFroman: unfortunately so
[15:50:07] *** lloy0076 has joined #opensolaris
[15:50:27] <eugene> i dislike amber lights in my room. it makes me sleepy
[15:51:35] <quasi> AbeFroman: in this case I'd expect netapps are quite a bit more expensive because they're on special build hardware - sun 5320 is more or less just an x4200 with a small display on the front and different software
[15:53:16] <axisys> quasi: my 5310 vol got corrupted twice in 16 days.. and had to reboot twice before it took fsck
[15:54:46] <axisys> even the 5310 cluster has issues
[15:55:08] <AbeFroman> I don't have experience with either.  that's just a refinement that of "you get what you pay for" that I used whe bosses want to buy cheap crap.
[15:55:55] *** mrdeviant has joined #opensolaris
[15:55:58] <axisys> i can ping/https/ssh/ftp a head but cannot nfs mount from it.. found out network port went orange on the switch.. i guess cluster partner only does ping test.. so failover was not working.. what a mess
[15:57:22] <quasi> sounds a bit like someone having put the 5320 in the rack without bothering to get a real config done
[15:59:20] <axisys> quasi: check this out if i unplug the network from one head .. the other head takes over in 60 sec.. however SUN set all those up .. but like u said w/o doing real config
[16:01:03] *** lacaAFK is now known as lacaMtg
[16:04:10] *** GigaE has quit IRC
[16:04:41] *** medar has quit IRC
[16:05:25] *** lloy0076 has quit IRC
[16:07:31] *** hile_ has quit IRC
[16:11:19] *** eboutilier has joined #opensolaris
[16:11:30] <quasi> axisys: I'd probably have a boss doing a whole lot of yelling if that was the case
[16:13:03] <axisys> quasi: we are actually .. but end of the day we keep it w/ some discounts on future purchase
[16:13:33] <quasi> axisys: oh, I'd be yelling to have things fixed
[16:14:11] <quasi> axisys: I rarely call vendors and tell them they can pick up their junk at the back door ;)
[16:14:52] <axisys> not sure how much u all used 5310 .. but i do not think a cluster parter would never takeover if the other head sees orange light in the netwrok switch .. if ping/https/ssh/ftp still responds but nfs not
[16:29:25] *** Dar is now known as Dar_HOME
[16:30:15] *** mazon is now known as Mazon
[16:30:33] *** jcea has joined #opensolaris
[16:40:25] *** loke has joined #opensolaris
[16:41:59] *** halton is now known as haltonAFK
[16:49:31] *** lacaMtg is now known as laca
[16:52:36] *** alanc has joined #opensolaris
[17:01:06] *** besonen_mobile has joined #opensolaris
[17:07:31] *** schily_ has joined #opensolaris
[17:14:05] <gdamore> good morning *
[17:15:46] *** schily has quit IRC
[17:18:07] <jteo> gdamore: morning.
[17:18:58] <gdamore> it looks like i might actually get a signed SCA so I can start contributing some of of my stuff.  wooo hoo.
[17:20:54] <jteo> gdamore: kudos. :)
[17:21:40] <jteo> i'm having fun deciphering zfs code. gotta love...abstraction.
[17:32:52] <timeless> heh
[17:33:11] <timeless> so, can you get copy on write for cp and svn?
[17:33:46] <timeless> svn's impl of file copies is actually quite evil and will defeat every imaginable file system api :(
[17:34:19] <timeless> (it's actually apr, not really svn)
[17:39:19] *** dunc has joined #opensolaris
[17:41:34] *** sommerfeld has joined #opensolaris
[17:41:34] *** ChanServ sets mode: +o sommerfeld
[17:45:23] <icon> meh
[17:45:26] <icon> svn irritates me greatly
[17:45:41] <icon> because file revisions were just soo useless...
[17:45:53] <icon> not to mention ghetto branching/tagging
[17:45:56] <icon> *sigh*
[17:46:24] <timeless> svn's tagging behavior *really* pisses me off
[17:46:32] <icon> yup
[17:46:35] <timeless> that's the main reason i picked ZFS
[17:46:54] <timeless> i'm hoping someone (probably me) can implement zero size copy on write for copies
[17:47:08] <timeless> then i'll reexecute the svn repo checkout version at a time
[17:47:12] <timeless> and get a huge space win
[17:47:25] <icon> meh
[17:47:42] <icon> i applaud trying to do repository revisions... but it is the entire f'n repository irrespective of what module you are working with
[17:47:46] <timeless> imagine having a tag a week foor each of say 100 projects
[17:47:51] <icon> MKS maintains file revisions and project revisions... whats so hard about that?
[17:47:56] <timeless> for say 2 years
[17:48:22] <icon> subversion forces you to tool *everything*
[17:48:44] <icon> at least with cvs you knew which file revision was the previous
[17:48:59] <timeless> svn doesn't even have goodtools
[17:49:05] <icon> nope
[17:49:12] <timeless> w/ cvs i have bonsai+mxr
[17:49:15] <icon> not to mention to server side properties... thats f'n retarded
[17:49:20] <icon> s/to/no
[17:49:31] <icon> because keyword expansion should be the users' responsibility!
[17:49:44] <timeless> svn destroys mxr (building an xref takes 3 1/2hrs, + 3 1/2 hrs )
[17:49:48] *** qbit has left #opensolaris
[17:49:52] <icon> ouch
[17:50:03] <timeless> because i really want to cross reference every single stupid tag
[17:50:20] <timeless> the alternative is *manually* trying to figure out what a "trunk" is
[17:50:26] <icon> subversion makes me want to find a small fillipino child and punch him in the face... repeatedly
[17:50:30] <timeless> and no, people don't consistently name or place their trunks
[17:50:53] <timeless> i'd rather svn be taken out andburned at a stake or two
[17:51:11] <icon> heh, svn makes me want to use cvs for heavy development again
[17:51:12] <timeless> heck, some people don't bother having trunks at all :(
[17:51:22] <timeless> yep
[17:51:33] <icon> i love the fact there is no standard layout... you are 'supposed' to have trunk/ tags/ and branches/
[17:51:34] *** sommerfeld has quit IRC
[17:51:38] <icon> but come on
[17:51:40] <icon> thats just lazy
[17:51:54] <timeless> i have some places where there's
[17:51:57] *** stevel has joined #opensolaris
[17:51:57] *** ChanServ sets mode: +o stevel
[17:52:07] <timeless> foopy/trunk foopy/barpy/trunk foopy/barpy/whacky/trunk
[17:52:16] <timeless> some places that have foopy/foopy too
[17:52:19] *** Symmetria has joined #opensolaris
[17:52:30] <timeless> i.e. where someone really did choose to copy a directory into itself or similarly
[17:53:03] <icon> yech
[17:53:35] <icon> dunno, as much as people rag on cvs/rcs/sccs, they got some things very right
[17:54:06] <icon> honestly, if you took cvs and did project level versioning, atomic commits, and directory versioning, youd have something pretty decent
[17:54:33] <icon> but no. they had to rip off perforce and then promptly screw it beyond all reasonable use
[17:54:47] <timeless> i believe you actually could have atomic commits if you wanted it
[17:54:48] <jteo> then don't use it.
[17:55:00] <icon> jteo: i dont.
[17:55:03] <timeless> jteo: employers or random people in groups ick
[17:55:10] <jteo> mmm.
[17:55:15] <timeless> not everyone works on their own small project
[17:55:30] *** Cyrille has quit IRC
[17:55:37] * timeless wonders if jteo works for sun
[17:55:37] *** sommerfeld has joined #opensolaris
[17:55:37] *** ChanServ sets mode: +o sommerfeld
[17:55:40] <icon> jteo: unlike most users that grouse about svn, ive used it pretty extensively server and user side.
[17:55:46] <timeless> if you did, i'm fairly certain you wouldn't have much input
[17:55:47] <jteo> timeless: i don't.
[17:56:03] <jteo> icon: fair enough. ;)
[17:56:10] <timeless> sun uses teamware or something? :)
[17:56:17] <icon> mercurial for opensol
[17:56:42] <timeless> how is that? it's been mentioned a couple of times, but i've never played w/ it
[17:56:48] <jteo> timeless: teamware. shifting to mercurial.
[17:56:51] <icon> couldnt tell you :)
[17:59:54] * icon grumbles about idiotic use of non-blocking io
[18:00:49] <icon> as much as i apprecia java.util.Process, the decision to make the streams block while the object is used asynchronously is just retarded
[18:00:56] <icon> s/apprecia/appreciate/
[18:01:18] <icon> sorry, java.lang :P
[18:02:11] *** bengtf has quit IRC
[18:05:04] *** bengtf has joined #opensolaris
[18:09:44] *** ShadowHntr has joined #opensolaris
[18:11:08] *** damienc has quit IRC
[18:13:36] *** Poptartz has joined #opensolaris
[18:14:01] <Poptartz> SXCR b50 due when?
[18:15:00] <quasi> Poptartz: http://dlc.sun.com/osol/on/downloads/b50/
[18:15:26] <quasi> Poptartz: since 10/12/2006
[18:16:17] <Poptartz> Oh..not the CD/DVDs though..still b49
[18:16:52] <Poptartz> sol-nv-b49-sparc-dvd-iso-a.zip
[18:16:57] <quasi> ah, you didn't say
[18:17:26] <quasi> in a couple of weeks probably
[18:17:54] <Poptartz> ok
[18:17:58] <quasi> or you could just bfu to b50
[18:18:01] *** eboutilier has quit IRC
[18:20:34] <Poptartz> oh..no kinda needed the CD/DVD. bfu takes too long!
[18:24:49] *** mdj has joined #opensolaris
[18:26:53] <PerterB> Using SVM soft partitions... How can I tell how much free space is left on the metadevice to allocate to new partitions?
[18:27:47] <asyd> hmm I have a little script
[18:27:55] *** mkhl has joined #opensolaris
[18:28:02] <asyd> try http://asyd.net/scripts/svm-stats.pl PerterB
[18:28:06] <PerterB> ta :)
[18:29:04] <PerterB> needs a bit of massaging to work with metasets, but thanks :)
[18:31:09] *** salamanders has quit IRC
[18:32:15] <PerterB> you'd think there would be a simpler way
[18:33:07] * quasi thinks there is
[18:33:22] <PerterB> go on...
[18:34:17] * quasi can't remember details - but there's a huge document about svm on docs. ;)
[18:34:41] *** yarihm has quit IRC
[18:36:56] <PerterB> there is, but it's pretty skimpy regarding soft partitions
[18:38:14] <quasi> true
[18:42:52] *** gtc has joined #opensolaris
[18:43:20] *** MattMan has quit IRC
[18:44:50] <PerterB> metastat -c would be useful, if this wasn't solaris 9 :)
[18:49:59] <axisys> my v20z keep rebooting like a yo yo
[18:50:27] <axisys> it shows this in messages file Oct 17 05:09:22 host.net genunix: [ID 540533 kern.notice] ^MSunOS Release 5.10 Version Generic 64-bit
[18:50:44] <axisys> it rebooted this many times already http://www.rafb.net/paste/results/An2DTO25.html
[18:51:29] <axisys> SUN wants to upgrade the SP thinkng that is the cause.. I think there is someother issue.. what anyone else thinks?
[18:54:01] <Stric> why don't you want to update the SP?
[18:57:03] *** mkhl has quit IRC
[18:59:39] <twincest> i had that problem with an x4100
[18:59:43] <twincest> upgrading the SP fixed it :)
[18:59:52] <twincest> of course that's a completely different system, but..
[19:00:55] <quasi> and a completely different sp
[19:04:20] <quasi> but other than that, there's really no harm in upgrading the sp - sure, it takes a reboot, but is is not like that's going to be much different than the current situation ;)
[19:05:09] *** mkhl has joined #opensolaris
[19:06:13] <Stric> hm.. on v240s etc, it doesn't take a reboot..
[19:07:09] <Stric> (yeah, I know.. different.. but why would the SP need to kill the OS?)
[19:08:45] *** yjsong has joined #opensolaris
[19:09:02] <axisys> just not sure how so many rebootes linkd to SP upgrade.. i am following up w/ the upgrade
[19:10:06] <quasi> since the sp can reboot the system, that's not entirely unlikely
[19:16:50] *** koolniczka has joined #opensolaris
[19:17:15] *** solarisjon has joined #opensolaris
[19:17:48] *** yjsong has quit IRC
[19:24:54] <gdamore> !seen alanc
[19:24:58] <Drone> alanc is currently online in #opensolaris and last spoke on Mon 16 Oct 2006 23:57 GMT, saying ' 	6477136- Application crashed in FcHideFont if non-alphabetical name exists'.
[19:30:46] *** kFuQ has quit IRC
[19:31:24] *** kFuQ has joined #opensolaris
[19:31:49] <axisys> quasi: good point
[19:37:06] *** Symmetria has quit IRC
[19:41:51] *** loke has quit IRC
[19:46:15] * timeless sighs
[19:46:28] <timeless> do zones have some setup phase that includes rewriting their contents?
[19:47:08] <timeless> and is it ok to mix nexenta from snv_50 w/ opensolaris from snv_49 ?
[19:47:19] *** bengtf has quit IRC
[19:50:10] *** svoboda has quit IRC
[19:51:27] *** eboutilier has joined #opensolaris
[19:53:42] *** svoboda has joined #opensolaris
[19:58:56] *** bengtf has joined #opensolaris
[20:03:00] *** gtc has quit IRC
[20:07:33] *** klocze1 has joined #opensolaris
[20:10:11] *** B|nTaRa has joined #opensolaris
[20:10:36] *** gtc has joined #opensolaris
[20:17:46] *** kloczek has joined #opensolaris
[20:20:26] *** gtc has quit IRC
[20:27:03] *** eboutilier has quit IRC
[20:28:56] *** deather_ is now known as deather
[20:39:32] *** dduvall has quit IRC
[20:41:59] *** dduvall has joined #opensolaris
[20:44:28] *** lisppaste3 has joined #opensolaris
[20:44:31] *** sahafeez has joined #opensolaris
[20:49:58] *** Kernel86 has joined #OpenSolaris
[20:50:10] *** quellhorst has left #opensolaris
[20:52:21] *** apa has quit IRC
[20:57:45] <Poptartz> Replacing 3Dlabs XVR-2500 with the Nvidia Quadro FX 3500 may be the long term solution.
[20:59:25] <gdamore> Nvidia would have to provide drivers for SPARC.  And Fcode.  Good luck.
[20:59:56] <Poptartz> ;>
[21:00:39] *** solaris-user has joined #opensolaris
[21:00:45] <gdamore> The biggest problem is that the SPARC graphics "team" doesn't want to play.
[21:01:02] *** OnkelSchorsch has joined #opensolaris
[21:02:03] <solaris-user> px.3b1.org <-- brazilian Opensolaris repository
[21:02:05] <Poptartz> ATI FireGL 7x00 -series suffice?? This would at least give SPARC machines another high-end offering.
[21:03:00] <gdamore> sparc is challenged.  there is a binary ABI used by 3rd party OpenGL apps.  It isn't what you'd "expect".  To get any SPARC OpenGL application, you need the SPARC libGL sources.
[21:03:37] <gdamore> we actually had a "stock" OpenGL running on Radeon 9250 on sparc.  it worked, except legacy OpenGL apps built with Sun's libGL don't work
[21:04:47] <gdamore> we actually even offered to help fund a port to Radeon hardware.  again, the SPARC graphics "team" is not interested.  we even offered to do the work ourselves and give the sources back to Sun.  No dice.
[21:05:28] <Poptartz> So the Sun OpenGL SDK (or sources) are not available through NDA?
[21:05:53] <gdamore> Nope.
[21:06:11] <gdamore> We have the 2D drivers thru NDA, but not OpenGL.  And no SDK or other support.
[21:06:32] <gdamore> As you might tell, the OpenGL/SPARC graphics people are not my favorite people right now.
[21:07:59] <delewis> gdamore: join the club.
[21:08:28] <gdamore> i don't understand how they can stay employed.  they are no longer offering any value to Sun, and are actually costing sun real money.
[21:08:47] <Poptartz> I'd think the OpenGL part would be OEM'd so that someone like 3Dlabs can provide the SDK as well as the drivers.
[21:08:48] <delewis> I've got an SB1000, and it's unfortunate that I'll doubtfully ever be able to use Xvideo on it (unless I throw in an XVR-2500 or whatever is the target for the Xorg driver port)
[21:09:38] <Auralis> delewis: unlikly, since newer cards will be pci-e only
[21:09:57] <gdamore> XVR-2500 won't be supported by Xorg, I think.  Only XVR-100.
[21:10:22] <gdamore> there is a PCIe 3Dlabs chip that Sun will be going to.  maybe it is the XVR-2500, i'm not sure.
[21:10:28] <delewis> gdamore: there was mention of an Xorg driver port for the new cards (which is the XVR-600, XVR-2500, etc.)
[21:10:29] <gdamore> but after that, Sun is out of options.
[21:10:30] <Auralis> the xvr-2500 should be supported by 3dlabs own drivers for x.org since its a realizm 500
[21:10:35] <Poptartz> Would Xorg consider ATI FireGL 7x00 series??
[21:10:39] <delewis> XVR-600 was never confirmed, but I heard the XVR-2500 mentioned a couple of times.
[21:10:59] <gdamore> okay, well, sure, if there are Xorg drivers for wildcat, then it will work.
[21:11:15] <Auralis> Poptartz: except that ati opengl is crap across the board
[21:11:15] <delewis> I wouldn't mind if Nvidia provided Sun a solution -- GeForces with Fcodes :-)
[21:11:50] <gdamore> not likely to happen.  Sun is too tied up with AMD right now, I think.
[21:11:51] *** gtc has joined #opensolaris
[21:12:26] *** OnkelSchorsch has quit IRC
[21:12:37] <delewis> well, if there's not an adequate 3D solution, the SPARC workstation line's future suddenly seems much more finite.
[21:13:05] <gdamore> that's a real problem.
[21:13:56] <gdamore> obviously we have a vested interest in addressing that concern.
[21:14:01] <delewis> gdamore: for the consumer, perhaps not -- consider some of the commercial SPARC binary translators available, nowadays.
[21:14:13] <delewis> for us SPARC enthusiasts, though -- yes, it is a *huge* problem.
[21:14:43] <Poptartz> What about the mobile laptops...I'd think Sun would want a embedded graphics chip so spearhead that market as well.
[21:14:54] <delewis> gdamore: out of the 5,000 employees Sun canned, it's a shame a certain, unnamed department didn't get canned, as well. :-)
[21:15:01] <delewis> Poptartz: Sun got out of the mobile laptop industry
[21:15:04] <delewis> the Ultra 3 was cancelled.
[21:15:09] <gdamore> yes.  i'm trying to make that point.
[21:15:12] <Poptartz> What!
[21:15:17] <Auralis> without a good gfx sparc solution, sparc desktops are dead, period
[21:15:19] <delewis> Poptartz: it was very quiet
[21:15:24] <delewis> I'm apparently the only person that actually noticed
[21:15:26] <gdamore> i hadn't heard that.
[21:15:37] <delewis> as I was going to purchase a SPARC portable solution
[21:15:41] <gdamore> the ultra 3 was an oem'd solution anyway.
[21:15:42] <clee> all of the alternative archs are going away.
[21:15:46] <delewis> though, I've decided to go the Acer Ferrari route
[21:15:49] <delewis> gdamore: yeah
[21:15:57] <gdamore> you can still buy an ultra 3 from us. :-)
[21:16:07] <clee> alpha died long ago, then powerpc, and now sparc
[21:16:08] <delewis> low-end Ultra 3 was a Tadpole of some sort, and the high-end was a Naturetech Meso
[21:16:09] <gdamore> except we'll call it a "viper" or "sparcle". :-)
[21:16:14] <delewis> gdamore: I've already called in a few quotes
[21:16:21] <delewis> just not practical from a hobbyist perspective
[21:16:24] <gdamore> yes, but we've got better high end than nature, too.
[21:16:27] <delewis> I was aiming at the $3.5k range
[21:16:27] <gdamore> true.
[21:16:45] <gdamore> most of our customers are government/military, with a need, and the budgets to match
[21:16:46] <delewis> and I gave a Naturetech VAR, NextCom, and Tadpole a call
[21:16:51] <delewis> all returned quotes around the $5k range.
[21:17:02] *** solaris-user has left #opensolaris
[21:17:10] <delewis> for an UltraSPARC-IIe with 512MB of memory :-)
[21:17:23] <gdamore> i think i could have got you into a $3k laptop, but it would be a 650MHz US-IIi, not a 1GHz US3i.
[21:17:32] <Auralis> well, i do know for certain that my next desktop will not be a sparc
[21:17:37] <gdamore> The US3i systems are all up over $5k
[21:17:38] <delewis> gdamore: that's what I was aiming for -- just a USII
[21:17:44] <delewis> I knew the USIII was going to impossible
[21:17:48] <delewis> but I had some hope for the USII
[21:18:08] <gdamore> hmm... sounds like you weren't talking to the right people.  we have offerings starting at $2995.
[21:18:26] * delewis wishes SPARC portable solution providers would *publish* their prices
[21:18:31] <delewis> see, I didn't know that :-)
[21:18:43] <delewis> everything is cloak and dagger -- call and get a quote
[21:19:25] <gdamore> the problem has to do with some rules regarding the US government.  you can't publish a price cheaper than what you sell to the military for.  and sometimes the military requirements for testing, certification, etc. are huge enough that they _should_ pay more than commercial customers.
[21:19:32] <Poptartz> Would be nice to have a laptop with SPARC and maybe a Nvidia Go chip
[21:19:57] <delewis> gdamore: ah
[21:20:19] <delewis> http://riemann.solnetworks.net/~dlewis/quote1.pdf
[21:20:20] <gdamore> Truth be told, as much as I love SPARC and hate x86, the SPARC systems make terrible mobile systems unless you _need_ sparc.  SPARC was never designed for mobile considerations like power conservation.
[21:20:21] <delewis> that was the NextCom quote
[21:20:29] <delewis> I don't have the Naturetech and Tadpole quotes on hand at the moment
[21:20:31] <delewis> but they were similar
[21:20:57] <gdamore> what?!  we don't even _offer_ a 400MHz unit.  We start at 500MHz.
[21:21:04] <delewis> gdamore: yeah
[21:21:15] <delewis> at that point, I just gave up, and went to us.acer.com :-)
[21:21:47] <Poptartz> Glad you mentioned that..so PowerPC is a dud as well.... not much to choose from nowadays.
[21:21:52] <gdamore> don't blame you.  but again, you should have been able to get a cheap quote from us.  if you can't, please let me know who you are talking to so I can straighten them out.
[21:22:34] <gdamore> Heh.  they're claiming 3D acceleration of their XVR graphics.  I don't think so.:-)
[21:22:42] <delewis> yeah :-)
[21:23:10] <delewis> gdamore: regardless, I appreciate the thought, but for a hobbyist SPARC portable solutions just aren't practical and that's a *real* shame.
[21:23:32] <gdamore> Modern SPARC systems aren't targetted at the hobbyist.  Not enough volume.
[21:23:33] <Poptartz> So true.
[21:23:52] <delewis> gdamore: indeed.
[21:24:16] <gdamore> Believe me, if you understood the costs with the custom production on these units, you'd know why.  When you build 1M units like dell does, you can get some good economy of scale.  Its hard to do that with runs of ~1k units.
[21:24:28] <delewis> of course, I don't know many times I've heard of someone within the OpenSolaris community trying to go the SPARC portable solution route and get turned away because of pricing or bad representation from the company.
[21:24:48] <gdamore> If we're to blame for that, I'd like to know.
[21:24:49] <delewis> Rich Teer and benr both were considering here (benr blogged about his experience with a Tadpole rep)
[21:25:14] <gdamore> Nature has a really bad reputation for being "challenged" wrt honesty, but I would hope we don't have those problems.
[21:25:32] <gdamore> Do you have a URL for benr's blog?
[21:25:52] <jamesd> cuddletech.com
[21:25:54] <delewis> http://72.14.209.104/search?q=cache:xDuvj5B0UEUJ:www.cuddletech.com/blog/pivot/entry.php%3Fid%3D104+ben+rockwood+tadpole&hl=en&gl=us&ct=clnk&cd=1&client=firefox-a
[21:25:57] <delewis> only google cache :-(
[21:26:04] <delewis> his blog seems to be down at the moment
[21:26:14] <Auralis> at my last job we had a naturetech 777s, damn, that thing was shit, loud, hot, cheap build, heavy, the batt died every half year and so on, oh and slow as shit
[21:27:06] <gdamore> I've heard other people say less than complimentary things about the Nature products.  They also don't contribute to the community, or back to Sun.
[21:27:43] <gdamore> Anyway, thanks for the log, I am goingto forward this on to mgmt, because if Dennis is not doing his job, then we need to understand why.
[21:27:58] <g4lt-mordant> yeah, naturetech has had sparc wifi drivers for years, are they in ON?
[21:28:16] <gdamore> no.
[21:28:31] <gdamore> they licensed their wifi drivers from some other company.  can't recall the name offhand.
[21:28:35] <g4lt-mordant> of course so has tadpole, so go figure
[21:28:42] <gdamore> they have ~0 in house software expertise.
[21:28:59] <gdamore> I wrote Tadpole's wifi drivers.  So blame me. :-)
[21:29:39] <Poptartz> gdamore - does Tadpole have a high-end graphics solution for SPARC laptops?
[21:29:39] <gdamore> I'm still trying to get our stuff (fixes for pcpw for example) committed back to OpenSolaris.
[21:29:42] <delewis> gdamore: no problem, glad I could be inform someone about the current SPARC portable situation from the consumer perspective.
[21:32:25] <gdamore> Poptartz: we are still figuring out what to do about graphics.  Most of our units have m64's but we also use XVR-100.  We experimented with R9250, but we have the same compatibility problems I've already mentioned.
[21:33:04] <gdamore> basically, _nobody_ can do decent sparc graphics without getting access to Sun's libGL internals
[21:33:20] <delewis> it's amazing that even Tadpole can't get a response from the SPARC graphics team.
[21:33:22] <gdamore> (at least not without breaking the Sun OpenGL ABI)
[21:33:46] <Poptartz> :<
[21:34:13] <gdamore> Now you know why I want the idiots that are preventing _anyone_ from making forward progress to get their asses fired.
[21:34:22] <delewis> gdamore: indeed.
[21:34:54] <delewis> send Jonathan an email :-)
[21:34:56] <gdamore> We have also looked at 3Dlabs parts, in the interim, but even the XVR-600 is not conducive for laptop integration due to power and real-estate requirements (and also heat dissipation)
[21:35:06] *** brandon_ has joined #opensolaris
[21:35:30] <Poptartz> 3Dlabs mentioned they have embedded solutions..any ones come to mind?
[21:37:03] <brandon_> is it possible to create a zpool over disk partitions residing on different systems over a network?
[21:37:13] <Poptartz> I'd think even the Permedia would get some support through Sun.
[21:37:29] <Poptartz> They had the drivers a long time ago.
[21:38:47] <gdamore> we have looked very, very closely at 3dlabs.  we're still in conversations with them.  but right now we're finding it hard to match product requirements to offerings.
[21:39:06] *** xzilla_ has quit IRC
[21:39:26] <gdamore> and ultimately, if it isn't supported by a stock Sun driver, we are going to have the same compatibility problems.  Thank you again, SPARC graphics fuckups.
[21:39:58] <sahafeez> can someone just go and slap the sparc graphic people on the head.
[21:40:26] <Poptartz> How many of them are left?
[21:40:39] <delewis> Poptartz: I'm sure that's classified :-)
[21:40:44] <gdamore> only a few.
[21:40:52] <Poptartz> ;>
[21:40:56] <gdamore> but they are very worried (and rightly so) about their jobs.
[21:41:09] *** xzilla has joined #opensolaris
[21:41:14] <gdamore> all decisions about sparc graphics seem to be based on how it will help them stay employed.
[21:41:36] <delewis> pleasing the OpenSolaris community might be a start.
[21:41:45] <Poptartz> OpenGL 2.1 maybe?
[21:42:14] <delewis> not to mention all the SPARC portable vendors :-)
[21:43:35] <Poptartz> Would tadpole be interetsde in designing the hardware if they had engineers partnering with the graphic companies to write the drivers? If possible?
[21:43:40] <noyb> so what do you guys think of project black box?
[21:43:43] <Poptartz> 'interested'
[21:43:45] <gdamore> i think so, yes.
[21:44:11] <delewis> noyb: seems like a PR stunt.
[21:44:18] <Poptartz> Hmm... see...there are a few jobs opening up everyday!
[21:44:42] <gdamore> the problem is again, we're stuck with the sparc graphics people, who are stonewalling us.  fucking idiots.
[21:44:47] <noyb> delewis: you don't think sun is serious about it then?
[21:44:50] <gdamore> (sorry for the language)
[21:45:26] <delewis> gdamore: you have every right to be irritated. Anyone with a SPARC workstation has similar feelings at the moment, including myself.
[21:46:05] <Poptartz> You have a good point as some of us may want an Ultra 20 over and Ultra 25 and should know these things.
[21:46:16] <gdamore> this could cost us like $300M.  we would _throw_ money at them if they would just work with us.
[21:46:18] <noyb> maybe they're stonewalling because they don't like to be verbally assaulted.  offer your solution to the problems, not your venom.
[21:46:20] <Stric> it feels weird when a sunray1 beats the crap out of a blade1000 w/ c3d in graphics performance :P
[21:46:31] <delewis> gdamore: that's unbelievable.
[21:46:54] <gdamore> we've asked for a source license.  price no object.  no interest from them.
[21:47:17] <delewis> and this is just libGL right?
[21:47:22] <delewis> no hardware specs
[21:47:32] <gdamore> noyb: believe me, the venom is just a last resort, because i've tried _everything_ else.
[21:47:36] <gdamore> yes, just libGL.
[21:47:43] <delewis> wow.
[21:48:18] <noyb> go higher and higher in the mgmt chain until you get the right answer to your proposal of 'no limit $$' for source.
[21:49:04] <Poptartz> Micorsoft dealt with this about 10 years ago..they let the ISV provide the drivers for OpenGL...while they worked on the DirectX extract with is now having OpenGL layered on top again..
[21:49:10] <gdamore> believe me, i think we're trying.  but its really hard.  there has been some "dishonesty" about why they "can't" release libGL.
[21:49:52] <delewis> gdamore: so, basically, they *could* release libGL (via source license and NDA), but they're too lazy/unmotivated, etc.?
[21:49:55] <gdamore> i don't think we've engaged Jonathan yet.  but he might be next.
[21:50:02] <gdamore> yes.
[21:50:19] * delewis sigh
[21:50:32] <delewis> if getting libGL out of them is this difficult, I imagine hardware specifications are next to impossible.
[21:50:49] <gdamore> i mean, it might require a 3 way agreement with 3Dlabs or something, but we are having conversations with 3Dlabs, and I think they would friggin' _jump_ at the opportunity to get some licensing revenue from us.
[21:51:30] <gdamore> for graphics chips, yes.  i think corporate espionage is the only way anyone is ever going to see them.
[21:51:59] <delewis> no Xvideo for my Elite3D then :-(
[21:52:19] <gdamore> not ever likely.
[21:52:48] <gdamore> if you don't like it, write a blog and post it to jonathan.
[21:53:11] <gdamore> i have to be careful, because my company still has positive relationships with other parts of Sun that we don't want to sour.
[21:53:24] <delewis> that's really what I or anyone else needs to do that's still interested in SPARC workstations.
[21:53:54] <gdamore> the problem is, I'm not sure anyone at _sun_ is still interested in sparc workstations any more.
[21:54:00] <noyb> gdamore: so far, in the last 10 minutes anyway... you haven't been very careful.
[21:54:03] <delewis> gdamore: indeed.
[21:54:04] <Poptartz> But you have a valid point... some people need good graphic workstations.
[21:54:13] <delewis> the situation might've been different about a year ago, but with all the AMD hype..
[21:54:20] <Poptartz> Especially MCAD/MCAE
[21:54:21] <gdamore> there is no public log of this, i hope. :-)
[21:54:40] <delewis> gdamore: you're lucky, 'fish' isn't in here at the moment
[21:54:59] <noyb> uh... I log everything.  Every day.   :)
[21:55:14] <delewis> you don't have it in a public_html dir, do you? :-)
[21:55:24] * Stric logs * too, for grep usage only
[21:55:25] <gdamore> plausible deniability.  "i didn't know it was getting recorded for posterity" :-)
[21:55:43] <noyb> not anymore...   ;)
[21:55:52] <gdamore> thanks, noyb.
[21:56:34] *** bengtf has quit IRC
[21:57:13] * gdamore wonders if noyb is really jonathan's anonymous handle.
[21:57:44] <delewis> cat irclog.txt | sed s/gdamore/foo/ | mail jonathan at sun dot com :-)
[21:57:48] <g4lt-mordant> gdamore, you have some competition in portable computing: http://www.sun.com/emrkt/blackbox/story.jsp
[21:57:51] <g4lt-mordant> ;P
[21:58:21] <gdamore> heh.  :-)
[21:58:30] *** alanc-away is now known as alanc_work
[21:58:38] <gdamore> we fit about a dozen of our units into a shelter on a humvee.
[21:59:00] <gdamore> i also think soldiers are going to have a hard time carrying a black box on their back. ;-)
[21:59:11] <alanc_work> there's one sitting in the Sun MPK parking lot right now with tours being given to the press and employees - just came in from there
[21:59:15] <alanc_work> looks cool
[21:59:19] <g4lt-mordant> gdamore, US could do it nicely ;P
[21:59:23] <g4lt-mordant> USN even
[22:00:06] <gdamore> its actually a great idea.  i wonder how many are getting shipped to iraq?
[22:01:31] <noyb> gdamore: put your mind at ease.  I'm not him, I'm just a nobody at Sun.
[22:01:35] <alanc_work> they do have one painted camoflauge colors
[22:01:41] *** Tpenta has quit IRC
[22:01:50] <gdamore> heh.
[22:01:59] <alanc_work> jonathan's out with the blackbox giving the press tours right now anyway
[22:02:23] <Auralis> lets face it, sparc graphics are dead as it is currently, and with them the sparc workstations will went away as well
[22:02:27] <delewis> I wonder what pricing is on one :-)
[22:02:30] <gdamore> just add power and bandwidth.  i wonder if there is going to be a cottage industry of folks building matching generators in containers.
[22:02:31] *** Kush- has joined #opensolaris
[22:03:31] <gdamore> noyb: if you have the ability to thump someone in the sparc graphics group on the head, please do so.  if you work in the sparc graphics group, please try to figure out how to fix the problems i've identified.  if you want to contact me in e-mail, let me know.
[22:03:34] <alanc_work> just imagine the bandwidth of a container full of thumpers - latency is a killer, but you could still probably get a petabyte from LA to SF faster by truck than net
[22:03:47] <gdamore> heh.
[22:04:21] <gdamore> i should do the math.  just a sec.
[22:05:04] <gdamore> 34.7Mbps.
[22:05:18] <gdamore> assuming 8 hour travel time.
[22:05:33] <gdamore> correction.  34.7Gbps.
[22:05:39] <delewis> hah
[22:05:53] <delewis> few orders difference of magnitude there :-)
[22:05:57] <Auralis> now take a 3000 container cargo ship new york to london :)
[22:06:37] <gdamore> heh.  but if you have to ship less than the full payload (internal fragmentation?) its going to be _really_ expensive.
[22:07:03] <gdamore> what's the travel time from new york to london via container ship?
[22:07:23] <Auralis> no clue, 10 days?
[22:07:30] <gdamore> sounds too short to me.
[22:07:44] <Auralis> lets say 14 days
[22:07:55] <gdamore> but for redundant data centers, black box would be _cool_.
[22:08:06] <gdamore> i wonder what the cost on these is like.  ~$1M?
[22:09:03] <gdamore> 2.48Tbps.
[22:09:26] <gdamore> assuming 3000 containers and 14 days to deliver.
[22:09:37] <gdamore> (and each container holding a petabyte)
[22:10:03] <gdamore> (what is the next order of magnitude after petabyte?)
[22:10:12] <AbeFroman> exa
[22:10:42] <gdamore> so who needs data packets with an MTU of 3 exabytes?
[22:11:48] <gdamore> hmmm.... my calculations above are off by a factor of 8.  i forgot to convert bytes -> bits.
[22:12:16] *** Andrew____ has joined #opensolaris
[22:13:28] <alanc_work> it's just a prototype being shown today - they're not ready to start selling yet, so I don't think they've set pricing
[22:13:44] <gdamore> i like the picture of the black box on the moon.  i _think_ the size is maybe not well suited for delivery via rocket. :-)
[22:13:53] <sommerfeld> i thought it was mars
[22:13:59] <gdamore> i meant mars.
[22:14:21] <alanc_work> but it could fit in the backyard for the startup who outgrows their founder's garage
[22:14:23] *** Mazon is now known as mazon
[22:14:27] <Poptartz> Alanc <- is it competitive to IBM Blue Gene (scientific/military/DoD purposes)
[22:14:34] <sommerfeld> i think the airbag landing system would need work to handle a container that size.
[22:14:38] <Poptartz> This 'Black Box'
[22:14:39] <gdamore> heh.
[22:14:39] <alanc_work> I don't know IBM Blue Gene
[22:15:09] <Poptartz> AlanC <- Supercomputer competitive.
[22:15:12] <alanc_work> http://www.sun.com/blackbox explains black box
[22:15:17] <alanc_work> black box is not a computer
[22:15:22] <Poptartz> ah
[22:15:22] <gdamore> i remember when starcat (SF15K) was limited due to sizing of crates they could ship on an airplane.  I guess black box is the answer to that problem.
[22:15:27] <alanc_work> it's a portable (luggable?) data center
[22:15:30] <Poptartz> 250 T1000s ?
[22:15:40] *** coffman has joined #opensolaris
[22:15:53] <alanc_work> though they said if you filled it with the new 4600 M2's it would be in the Top 200 supercomputer list
[22:16:17] <alanc_work> it's a standard sized shipping container with a mini-data-center inside
[22:16:19] *** dunc has quit IRC
[22:16:23] <g4lt-mordant> never underestimatre the bandwidth of a bob hope class carrying a full load of blackboxes
[22:16:39] <alanc_work> you decide what goes in the racks - all CPU's, all disks, or whatever mix you need
[22:16:46] <gdamore> i guess this is just a bunch of racks, power, and network infrastructure in a container.  interesting.
[22:16:53] <Poptartz> Ah..the black box is mainly the mobile unit to house a datacenter.
[22:17:01] <alanc_work> yes
[22:17:38] <gdamore> the problem is, you still have to get power and network _to_ the black box.
[22:17:41] <alanc_work> for when you need a short-term data center somewhere (disaster recovery?  invasion?) or need to expand quicker than constructing a new one (just dump it in the parking lot)
[22:18:05] <Poptartz> Cool concept..kinda like mobile datacenters for road shows.
[22:18:13] <Error_404> i had that idea a while ago... fill a TFE with computers & call it a computer room
[22:18:20] <Poptartz> Military can use that.
[22:18:27] * gdamore imagines backing up to someone elses' black box with a forklift.  a million dollar heist. :-)
[22:18:36] <Poptartz> :D
[22:19:08] <gdamore> as i said, i wouldn't be surprised if these get very popular in places like iraq.
[22:19:13] <Poptartz> I think I'll keep track of this project.
[22:19:27] <sommerfeld> gdamore: I think the military's been doing this sort of thing for a while, actually.
[22:19:39] <gdamore> though they still need more mobility than even that.  forward observers,  forward command posts, etc. mobile ATC, etc.
[22:19:40] <alanc_work> from what I heard, it sounded like the most innovative part was the water cooling system
[22:19:55] <alanc_work> though that may have just been the easiest to explain 8-)
[22:20:03] <Poptartz> gdamore <- Real world testing - ya know?
[22:20:32] <gdamore> hmm... I think the HOA in my community might complain if I set one of these on my driveway....
[22:20:45] <Kush-> this black box is nothing new
[22:20:48] <Poptartz> Mobile ISP
[22:20:53] <Kush-> from what i've read
[22:21:36] <Poptartz> No it isn't... USAF had something like this many years ago...maybe not as prett though... ;>
[22:21:46] <sahafeez> it seems more of a gimmick then anything
[22:21:56] <Kush-> http://www.vmunix.com/mark/blog/archives/2005/11/21/google-and-slashdot-laughs/
[22:22:17] <Kush-> not a gimmick but good to see Sun trying it out too
[22:23:00] <Poptartz> McDonnell Douglas/Northrop has test units which is similar in concept to the idea.
[22:23:32] <Poptartz> (or had, 15+ years ago)
[22:24:06] <gdamore> well, in retrospect it seems obvious.   Sun may just be the first person to try to get some market play with it.
[22:24:16] <gdamore> s/person/company/
[22:25:48] <Error_404> umm... in the pictures sun made...
[22:26:02] <Error_404> why does the blackbox have a recycle logo on it?
[22:26:12] <AbeFroman> it's all recycleable
[22:26:18] <Error_404> "when you're done, don't just trash it.... sort the glass from the metal, yo"
[22:26:38] <Error_404> "and the lead? just ship it to india"
[22:27:32] <gdamore> heh.  i wonder if BlackBox is RoHS compliant.
[22:28:11] <Poptartz> Well Black Box is a very nice concept even for companies just wanting a 'cold site' backup.
[22:29:05] <Poptartz> I was looking for the typical military certifications for environmentals.
[22:29:08] <Error_404> and perfect for stocking up your fallout shelter/ mobile command post
[22:29:42] <Poptartz> might raise the price tag though.
[22:30:07] *** Kernel86_ has joined #OpenSolaris
[22:30:28] <Poptartz> Error <- mobile command post... I can see some movies out of this one. Solve the graphics problem first though!
[22:30:33] *** Kernel86|Laptop has quit IRC
[22:30:42] <Auralis> bah, add a kitchen and bed, geek dream home :)
[22:31:04] <Error_404> too noisy
[22:31:09] <sommerfeld> gdamore: so the "forklift" had me imagining what happens if the forks get shoved through the side of the container..
[22:31:24] <Auralis> actually, drop the kitchen, bed is enough and place is next to a McD or asia shop
[22:31:28] <Error_404> i wouldn't want to live in a place that requires hearing protection
[22:31:33] <oxygene> sommerfeld: cooked and fried driver, probably
[22:31:58] *** bengtf has joined #opensolaris
[22:32:37] <Poptartz> AlanC... toss in the idea for Universities like UCSD!
[22:33:06] <Poptartz> (or even small colleges looking for a datacenter on the go!)
[22:33:17] <Error_404> or a spamhouse from international waters!
[22:33:27] <Poptartz> :>
[22:34:26] <AbeFroman> even a werewolf deserves legal council
[22:34:42] *** nwf has joined #opensolaris
[22:37:52] *** brandon_ has quit IRC
[22:39:03] <Poptartz> Honestly, I think Sun has a viable option...yet they do need to talk to the developers and everyday users as well to make supoport of these special projects more pleasing... OpenSolaris is great...now we need more access to hardware that is well supported as well.
[22:39:12] <Poptartz> DDK/SDK
[22:39:50] <Poptartz> Low cost hardware..only x86 at the moment.
[22:40:16] <Error_404> what are you going on about?
[22:42:03] *** gtc has quit IRC
[22:42:06] <Poptartz> Just stating that we seems to see all this great stuff coming out...but have the same issues like the previous comments on graphics on SPARC hardware which has gone on for a few years now...
[22:42:31] <Poptartz> I'll get off that soapbox as we already went down that road long enough.
[22:42:54] <Poptartz> ;>
[22:44:21] <sickness> so what is all this virtualization fuzz about?!?
[22:44:35] <sickness> only to say that esx is going to be certified?!?!?
[22:44:56] <sickness> shit, I don't give a shit about esx running on linux, I want the damn vmware-server thing for solaris HOST >:)
[22:46:08] <sahafeez> just cannot get into the vm thing. i like a server per app
[22:46:09] <Auralis> what i realy want from virtualization is, install a base os that controls the machine, like solaris, then add some more OSes in virtuals, like a wintendo for gaming etc, but with full hardware gfx accel and stuff. and then i want to be able to a) display single apps on my base os dsktop or switch the display over completly to the other os
[22:46:15] <alanc_work> http://www.valleywag.com/tech/jonathan-schwartz/caption-contest-what-is-jon-ponytail-schwartz-thinking-208245.php
[22:48:59] <sommerfeld> gdamore: looks like ~1 week NY->London via container ship
[22:49:21] <gdamore> NY->London is faster than LA->NY, I think.
[22:49:33] <gdamore> (No panama canal. :-)
[22:49:34] <jengelh> both as fast
[22:49:54] <richlowe> ny->london and ny->la are about the same I'm told.
[22:50:02] * richlowe hasn't been further west than chicago, for just that reason. ;)
[22:50:08] <gdamore> really?  that's surprising. :-)
[22:50:17] <jengelh> Mathematically, it always takes the same amount of time to travel the same distance when speed is constant.
[22:50:44] <gdamore> sickness: yes, I want vmware for Solaris host too.  I posted on Andy Tucker's blog, but never heard anything else.
[22:50:51] <sommerfeld> http://www.maerskline.com/appmanager/ will give you transit times
[22:51:19] <jengelh> of course you have to consider the time *in flight*, not when you depart/arrive
[22:52:13] <gdamore> i'd have thought that a) the distance going south thru the canal would be a lot longer, b) the time to get thru the canal itself would have added an extra couple of days.  but i'm largely shipping ignorant. :-)
[22:52:20] *** glagasse has quit IRC
[22:52:58] <gdamore> sommerfield
[22:53:01] <gdamore> doh, sorry.
[22:53:27] <sickness> gdamore: yeah, what makes me more angry is that ESX is nothing but linux at last, so if only xen would run windows unmodified (I'm looking forward to VT etc support...) we could ad least virtualize over solaris+ufs/zfs instead of the poor ext3 with $random_crappy_controller_driver_which_works_only_with_redhat7.2_and_kernel_2.4 :/
[22:53:46] *** dunc has joined #opensolaris
[22:53:48] <gdamore> sommerfeld: if you can do anything to poke at the sparc graphics people (maybe report general feelings) please do so.  i'm looking for anyone who can help us get a roadmap for graphics on sparc systems.
[22:54:51] <gdamore> i just want to occasionally boot a VM with NetBSD or Linux in it. :-) I'd also like to be able to run a VM with newer OpenSolaris builds.  (My desktop U20 still runs S10 03/05.)
[22:55:02] *** sahafeez_ has joined #opensolaris
[22:55:36] <Auralis> the local domain stuff for the T1 looks interesting
[22:55:46] <sickness> well xen should be ok for netbsd or linux
[22:56:15] *** sahafeez_ has quit IRC
[22:56:16] <sickness> I wouldn't mind qemu+kqemu too for that task
[22:56:38] <sommerfeld> i have no contacts with the sparc graphics folks.
[22:56:39] <gdamore> xen requires me to reinstall. :-(
[22:57:04] <sickness> heh :/
[22:57:08] <sickness> maybe qemu doesn't ;)
[22:57:15] <gdamore> (my point was that maybe someone who knew someone could kick it upstairs.  Sun mgmt needs to get a clue about the problem.)
[22:57:38] <gdamore> qemu doesn't even _compile_ on Solaris x86.
[22:58:08] <gdamore> i started looking at fixing it, but quickly lost interest.
[22:58:11] <Auralis> blastwave has qemu or solaris
[22:58:27] <gdamore> An older build, and maybe even sparc only.
[22:58:40] *** nwf has quit IRC
[22:58:53] <gdamore> well, it has an x86 binary.  but its ancient 0.7.0.
[22:58:54] *** eboutilier has joined #opensolaris
[22:59:01] <Auralis> older is possible, lost of stuff on blastwave seems a bit old, but for x86 as well
[22:59:12] <alanc_work> gdamore: I don't even have a roadmap for sparc graphics beyond the XVR-300 that came out this month
[22:59:16] *** nwf has joined #opensolaris
[22:59:17] *** Poptartz has quit IRC
[22:59:32] <gdamore> i never heard of XVR-300.  what is it?
[22:59:41] <alanc_work> the low-end PCI-E board
[22:59:45] <gdamore> PCIe radeon?
[22:59:51] <alanc_work>  	 ATI FireMV 2200
[22:59:58] <gdamore> yep.
[23:00:12] <alanc_work> http://www.sun.com/products-n-solutions/hardware/docs/html/819-6651-10/
[23:01:20] <gdamore> Alan: what can I do to help resolve the problem of getting a graphics solution for SPARC going forward?  I _want_ to help.
[23:01:37] <Kush-> http://www.sun.com/emrkt/blackbox/images/zot_bb_scenario_2.jpg this one is funny
[23:01:39] <gdamore> eww... XVR300 has only a single DVI port.
[23:01:42] <alanc_work> not sure - I'm off to a meeting right now, so I can't talk at the moment
[23:01:46] *** alanc_work is now known as alanc-away
[23:02:05] <sickness> gdamore: correct me if I'm wrong, but I think that the qemu packages (0.8.2+patches) at opensolaris.org are built on an ancient solaris 8 :)
[23:02:20] *** sahafeez has quit IRC
[23:02:33] <gdamore> i've not looked at those, so I can't speak to them.
[23:02:35] <sickness> http://www.opensolaris.org/os/project/qemu/
[23:02:56] <Error_404> i must say, the "humanitarian relief" idea, i don't much get
[23:03:16] <gdamore> i never knew there was a qemu opensolaris project.
[23:03:38] <Error_404> shipping crate full of food is, IMO,more important than shipping crate full of compute power in case of tsunami or whatever
[23:04:01] <Error_404> "oh no, these people are suffering! what's the 10 quadrillionth digit of pi?!"
[23:04:09] <gdamore> maybe.  if you need a data center to reestablish local telecomm or whatever, it might be almost as important.
[23:04:26] <gdamore> (think rack full of NEBS equipment)
[23:04:34] <sickness> gdamore: yeah, and it's actually really cool IMHO =)
[23:04:42] <sickness> gdamore: I've started to try it from day one :)
[23:05:08] <Error_404> yeah, i hadn't thought of telecom
[23:05:18] <gdamore> cool, there is a kqemu accelerator too.  nice.
[23:05:29] <sickness> yeah
[23:06:23] <gdamore> does it work on S10, or do you need OpenSolaris to use it?
[23:06:28] <jengelh> N[Pi, 10^15]
[23:08:06] <sickness> I think it works on S10 it's built on S10...
[23:08:43] <gdamore> i'm gonna give it a whirl later.  when I can afford risking a crash. :-)
[23:08:52] <sickness> eheh, good :)
[23:10:18] <sickness> I love it, I also managed to run an opensolaris install with svm mirrored root as a guest :))) http://www.opensolaris.org/jive/thread.jspa?threadID=15124
[23:15:53] *** tsoome has joined #opensolaris
[23:17:23] *** esproul has left #opensolaris
[23:17:29] *** tsoome has quit IRC
[23:18:05] *** Kernel86_ has quit IRC
[23:25:20] *** esproul has joined #opensolaris
[23:29:07] *** mkhl has quit IRC
[23:32:33] <Gman> stevel, you involved in the svn stuff at all?
[23:32:33] <stevel> yeah
[23:32:33] <stevel> unless you're reporting something as broken
[23:32:33] <stevel> then no. i'm not involved
[23:32:33] <stevel> :-P
[23:32:33] <Gman> nope - just petty ugly behavior of the notification script to mail out
[23:32:33] <stevel> does it involve the raw svn command output?
[23:32:33] * Gman sees it attaches some svnlook crap at the end
[23:32:33] * Gman finds archive for example -
[23:32:33] <stevel> that would be it
[23:32:34] *** Ireul has quit IRC
[23:32:34] <stevel> no need
[23:32:34] <stevel> i know
[23:32:34] <stevel> i plan on removing it
[23:32:34] <richlowe> stevel: would this be a bad time to ask about the source drop? :)
[23:32:34] *** mrdeviant has quit IRC
[23:32:37] <stevel> richlowe: it's coming along. i forgot to update my sparc build machine per the tamarack flag day - so it borked
[23:32:44] <stevel> should be ready in a bit
[23:32:47] <Gman> http://mail.opensolaris.org/pipermail/jds-notify/2006-October/000008.html
[23:32:56] <Gman> stevel, sweet
[23:33:03] <Gman> just wasn't sure that dermot was really fighting for that change
[23:33:34] * Gman is an anal son of a gun.
[23:34:03] <sickness> http://www.gaelic.com/images/journal/GayInside.jpg ?
[23:34:24] *** alanc-away is now known as alanc_work
[23:34:25] <gdamore> hmm... i guess NetBSD will support sparc64 soon.  I guess that means that sparc64 is on its way to becoming a dead architecture. :-)
[23:34:55] <sickness> lol
[23:37:37] <stevel> richlowe: ON nightly is up now
[23:37:39] *** jerome__ has joined #opensolaris
[23:37:39] <gdamore> i meant us3, of course.
[23:37:44] *** stevel changes topic to "Latest releases: SXCR: 49, ON build: 50, ON nightly: 20061016"
[23:37:49] <jerome__> Hi
[23:37:55] <jerome__> quick question
[23:38:02] <jerome__> where do I set my default route
[23:38:11] <gdamore>  /etc/defaultrouter
[23:38:55] <jerome__> alright
[23:38:58] <jerome__> we'll see
[23:39:01] <jerome__> I'm rebooting now
[23:39:05] <jerome__> better be working :)
[23:39:20] <gdamore> heh.
[23:39:54] *** sahafeez has joined #opensolaris
[23:40:30] <jerome__> seems to works
[23:40:31] <jerome__> thanks man
[23:40:45] <AbeFroman>         NAME                                     STATE     READ WRITE CKSUM
[23:40:46] <AbeFroman>         raidinc                                  ONLINE       0     0   340
[23:40:47] <AbeFroman> wtf
[23:40:54] *** jerome__ has quit IRC
[23:41:41] <gdamore> jerome: np.
[23:42:58] <gdamore> wtf.  first time booting sparc s10 06/06.  it gives fatal errors during boot about being unable to open a zfs device.  I'm booting from _DVD_.
[23:49:35] *** gisburn has joined #opensolaris
[23:50:11] <gisburn> stevel: requesting permission to strangle the mailman server admin to death with fatal and permanent results...
[23:50:43] * stevel appoints gisburn to the mailman server admin position
[23:50:45] <stevel> permission granted
[23:50:52] <gisburn> stevel: "Message has implicit destination" ...
[23:51:06] * gisburn starts the execution...
[23:51:09] <gisburn> chhrrllllllll
[23:51:13] <gisburn> hhggllll
[23:51:19] <gisburn> chhrrrrrrrrrllllrrrrrrlll
[23:51:30] <Gman> which list?
[23:51:33] <gisburn>  **cchhrrll**
[23:51:58] 
[23:52:23] <Gman> well, at least you can admin that one yourself :)
[23:52:50] <gisburn> Gman: yes, I know... but I thought the problem was solved...
[23:52:59] <Gman> nup
[23:53:09] <Gman> well, sort of.
[23:53:44] <gisburn> Gman: the interesting bit is that right now all postings from *sun.com suffered the fate,
[23:54:07] <gisburn> Gman: since these were the only postings right now.
[23:54:23] <sahafeez> anyone think the snia certs are worth anything?
[23:54:25] <Gman> i've done my arguments of what the better mechanism will be
[23:54:26] <sahafeez> snia.org
[23:54:28] <Gman> it's out of my hands.
[23:54:54] <gisburn> chhrrrrrrrrrllllrrrrrrlll *chhrrrlll*
[23:54:55] <Gman> suspect it's out of stevel's too :)
[23:55:25] <gisburn> Ok, self-*chrrrll*-strangulation is tricky
[23:55:50] * gisburn would prefer to do it the good old "komodo dragon pit"-style
[23:56:07] <alanc_work> the problem is that Don put the ksh93 in the implicit cc for the new PSARC case instead of listing it on the cc line
[23:56:10] <gisburn> Gman: your proposal was ?
[23:56:56] <Gman> gisburn, just the way future arc works
[23:57:03] <gisburn> alanc_work: just curious: how do I specific an implicit CC: ?
[23:57:03] *** benr has joined #opensolaris
[23:57:03] *** ChanServ sets mode: +o benr
[23:57:11] <Gman> hey ben
[23:57:27] <alanc_work> in mailman?  dunno
[23:59:30] *** coffman has quit IRC





top