[00:03:25] *** smellyfish has joined #opensolaris
[00:06:38] *** Tramsie has quit IRC
[00:08:07] *** yarihm has quit IRC
[00:08:52] *** smellyfish has quit IRC
[00:14:35] <stevel> gman: it's got nothing to do with kernel development
[00:14:44] <stevel> well, it doesn't require kernel development expertise rather
[00:15:57] <Gman> stevel, yeah, more familiar with kernel development process perhaps
[00:16:10] <Gman> [just reading through the wiki tasks list]
[00:16:13] <alanc> ooh - Gman's name was just used on p-team call
[00:16:23] <Gman> hah
[00:16:27] <richlowe> what'd he do now? :)
[00:16:34] <Gman> was it used with the words 'fricken' and 'idiot'?
[00:16:49] <alanc> volunteered to help get e-mail out for OGB elections apparently
[00:16:58] <Gman> yeah, i did and i will
[00:17:22] <Gman> though i do wonder if my level of english is up to that task
[00:18:38] <Gman> i'm still disgruntled that no one commented on the voting list thing
[00:18:59] <Gman> since it currently assumes, unless i'm wrong, that only core contributors can vote
[00:19:44] <richlowe> hah.
[00:19:53] <richlowe> the voting list includes 'contributors' as well.
[00:19:59] <richlowe> assuming I looked at the right list.
[00:20:00] <Gman> it does
[00:20:13] <Gman> but the constitution says that only core contributors are actually members
[00:20:14] <Gman> iirc
[00:20:16] <richlowe> Hm.
[00:20:34] <Gman> [again, unless i've read it wrongly]
[00:20:50] *** xushi has quit IRC
[00:21:04] *** lloy0076 has joined #opensolaris
[00:23:36] <Gman> stevel, any plan to get that wiki task list linked on scm-migration os.o page?
[00:24:29] <timeless> hrm
[00:24:40] <timeless> is there fuse for opensolaris that would handle jffs2?
[00:25:32] <timeless> all i want is a way to treat jffs2 as a read only archive (ala zip) and extract files :)
[00:25:55] *** slowhog has joined #opensolaris
[00:27:32] *** Arnald has quit IRC
[00:29:34] <stevel> gman: yeah - i'll do it now
[00:29:38] <stevel> thanks for reminding me
[00:29:40] <stevel> forgot to do that earlier
[00:29:53] <richlowe> well bugger.
[00:29:58] <richlowe> if I'd known it was just that I'd have done it.
[00:30:06] <Gman> awesome, thanks!
[00:30:08] <richlowe> I thought you were doing actual content on there.
[00:30:26] <stevel> me? write content?
[00:30:27] <stevel> :-P
[00:31:13] * Gman reading through hg manual and thinking about putting together a simple user guide presentation
[00:31:16] *** Odin- has quit IRC
[00:31:19] <richlowe> do it!
[00:31:27] <richlowe> sorry, was that a little too eager?
[00:31:33] <richlowe> I mean, uh, "that'd be great". :)
[00:31:38] <elektronkind> http://rafb.net/p/8L66jH88.html
[00:31:44] <stevel> :)
[00:31:50] <elektronkind> I wonder how I can make this more handy
[00:31:55] <Gman> richlowe, it's much for my own benefit of figuring out how things work
[00:32:03] <elektronkind> I have disk controller stats done too
[00:33:14] <timeless> hrm
[00:33:23] <timeless> are documents like http://opensolaris.org/os/project/fuse/fuse-kernel_function_specs_v1.pdf supposed to be accurate?
[00:34:16] <timeless> it seems that about half of the descriptions of opcodes in that document are wrong
[00:34:28] <timeless> FUSE Operation Name utime
[00:34:29] <timeless> Description The function changes the size of file specified.
[00:35:20] *** Odin- has joined #opensolaris
[00:37:09] * timeless should go find the flagday page and see if fuse is planning on turning up in an snv anytime soon
[00:38:21] <Gman> i don't think so, though there is a fuse project being started
[00:38:23] <salmandr> the bug i filed, 6517914, has been marked as Incomplete:Need More Info
[00:38:38] <salmandr> but i didn't get an email, nor do I see anywhere in the bug what other info is needed
[00:38:47] <_william_> gn
[00:38:47] *** _william_ has left #opensolaris
[00:39:09] <salmandr> is there anything else i can do at this point?
[00:39:12] <Gman> timeless, http://www.opensolaris.org/os/project/fuse/
[00:39:34] *** hile_ has quit IRC
[00:40:37] <Tpenta> From the trace it looks like a dup of 6509271, but need access to the core file
[00:40:38] <Tpenta> to confirm, is it available someplace?
[00:40:45] <Tpenta> that's what they are after in that bug
[00:40:48] <Gman> timeless, http://www.opensolaris.org/jive/thread.jspa?threadID=18625&tstart=0
[00:40:58] <richlowe> Tpenta: could you yell at them for not emailing the reporter then?
[00:40:59] <Gman> though someone needs to prod them into publishing that info on their project page
[00:41:12] <salmandr> i can make it available, sure
[00:41:13] <richlowe> Tpenta: it's not like folks outside have *any* way to know that without being told.
[00:41:26] <Tpenta> I'll drop something in there; the request *is* in the bug description
[00:42:09] <salmandr> well i can see the request has been added to the description
[00:42:15] <salmandr> but i don't have any way to say "yes, here it is"
[00:42:18] <salmandr> other than this channel :)
[00:42:19] <timeless> gman: yeah, i found the url for the project, hence my comment  about bugs in their pdf :)
[00:42:24] <Tpenta> hmmm
[00:42:41] <Gman> timeless, join the list and contribute! :)
[00:43:13] <timeless> gman: so you think they'd accept a comment about bugs in the pdf? :)
[00:43:19] <Gman> sure
[00:43:28] <timeless> http://www.opensolaris.org/os/community/on/flag-days/pages/2006121401/ <- yum
[00:43:45] <timeless> but i think my next update will probably be when i get fuse
[00:44:26] *** sartek has quit IRC
[00:44:54] <Tpenta> I've updated teh bug pointing out that the person who made that request won't have their email address seen, so teh submiteer has no idea who to talk to; and that they should email you direct.
[00:45:15] <salmandr> thanks Tpenta
[00:45:33] <Tpenta> the request *is* in the last 2 lines of the descriptin at http://bugs.opensolaris.org/view_bug.do?bug_id=6517914
[00:45:34] <timeless> http://www.opensolaris.org/os/community/on/flag-days/all/
[00:45:38] <timeless> Dec-27 #opensolaris Brand-specific handlers for zoneadm(1M) commands - PSARC/2006/651
[00:45:42] <timeless> http://www.opensolaris.org/os/community/arc/caselog/2006/651/ 404
[00:46:02] <timeless> what's the plan wrt having arcs actually be readable?
[00:46:54] <alanc> the plan is to taunt plocher until he gets it right
[00:47:10] <alanc> well, that's my plan at least 8-)
[00:47:17] <Tpenta> :-D
[00:48:43] <richlowe> works for me.
[00:52:07] <Gman> and you get to have fun at the same time, bonus!
[00:52:56] <richlowe> christ on a crutch, they managed to kick off the debian argument again too.
[00:53:01] <stevel> gman: there's someone you can hassle when you get here ;-)
[00:53:05] *** peteh has quit IRC
[00:53:46] *** sartek has joined #opensolaris
[00:53:49] <Gman> stevel, yeah, though you already got the IAM.xml published ;)
[00:54:16] *** peteh has joined #opensolaris
[00:54:27] *** nachox has joined #opensolaris
[00:54:33] <stevel> yeah
[00:54:42] <Gman> oh boy, plocher forwarded on private mail to the psarc-ext alias
[00:54:43] <Gman> ooops
[00:54:57] * nachox gets flooded by the gplv3 thread
[00:55:56] *** uebayasi has joined #opensolaris
[00:56:28] <Gman> [ahh, got permission to do so before doing it..]
[01:01:01] <richlowe> nachox: the vast majority of it is seriously not worth reading.
[01:01:07] <richlowe> bits of it, sadly, are, however.
[01:01:12] <richlowe> so you end up having to go through it anyway.
[01:02:09] *** broadcast has quit IRC
[01:02:53] <nachox> unfortunately, it should have been handled with more care, i've seen some very valid points from darren and casper though
[01:04:08] <nachox> stephen saw the mailing list was quiet and wanted to add some spice :P
[01:05:32] <timeless> not that i have any value, but i'm not a fan of gpl :|
[01:09:44] *** gm152 has joined #opensolaris
[01:09:56] <agliodbs> I can't say that I'm thrilled by the constitution
[01:12:12] <Gman> agliodbs, feel free to comment :)
[01:13:21] <nachox> the constitution is being openly drafted for that very same reason
[01:15:25] *** raph_ael1 has joined #opensolaris
[01:15:29] *** raph_ael has quit IRC
[01:27:29] *** estibi has quit IRC
[01:28:48] <lloy0076> I installed two zpools today, one on /opt and the other on /export ... and my memory counter suddenly went almost all green during a big rsync.
[01:28:59] <lloy0076> It was rather...unnerving.
[01:29:10] <stevel> i went down to the 2nd floor to play a game of table tennis and to get some water because i was feeling dehydrated
[01:29:26] <stevel> i came back with no water, no table tennis played, and a bottle of beer instead
[01:29:33] <Gman> score!
[01:29:40] <alanc> heh - stumbled into the MPK17 beer bash?
[01:29:44] <stevel> yeah
[01:29:51] <stevel> gman: sure, but it's not fixing my dehydration problem :-P
[01:30:00] <Gman> drink enough and it will :)
[01:30:05] <alanc> don't know why they moved from Friday to wednesday...
[01:30:21] <stevel> alanc: maybe too many people wfh on friday?
[01:30:22] <alanc> just need to finish one little putback before I head over
[01:30:22] <stevel> i dunno
[01:30:48] <lloy0076> wfh?
[01:30:54] <Gman> work from home
[01:30:57] <lloy0076> ah
[01:31:26] <lloy0076> Why don't they have Internal Sun Employee Internet Speak documented in "Solaris Internals"?
[01:31:28] <lloy0076> :(
[01:31:52] <alanc> isn't that in the OpenSolaris Developer's Guide?
[01:32:05] <alanc> the glossary with terms like "brickify"...
[01:32:14] <lloy0076> Ah, yes. So it is...
[01:32:49] *** alanc is now known as Xbot
[01:32:59] <Xbot> putback by alanc: 6489660 Xorg server 7.2 [PSARC 2007/051]
[01:33:02] *** Xbot is now known as alanc-beer
[01:33:07] *** bubbva has quit IRC
[01:33:10] <richlowe> nice. :)
[01:33:27] <nachox> how many bots do we have here? :P
[01:33:32] <timeless> but when do i get gnome2.18? :)
[01:33:35] *** Somethingelse has joined #opensolaris
[01:33:40] <elektronkind> ok, here's some sun humor
[01:33:41] <elektronkind> http://marc2.theaimsgroup.com/?a=117003618700002&r=1&w=1
[01:33:46] <alanc-beer> X is cheap, we can't afford a real bot, so I just put on a bot mask 8-)
[01:34:28] <nachox> damn, share the beer, you didnt want it in the first place :)
[01:35:17] <elektronkind> xorg 7.2... excellent alan
[01:35:25] <elektronkind> b57-ish you think?
[01:36:21] <Gman> timeless, when the community releases it for a start
[01:36:35] <Gman> timeless, you can get 2.17.x packages if you want - beta quality though
[01:38:47] *** jsubl2 has joined #opensolaris
[01:42:11] *** gisburn has joined #opensolaris
[01:48:00] *** yongsun_ has joined #opensolaris
[01:51:50] *** Yamazaki-kun has quit IRC
[01:52:31] *** hile_ has joined #opensolaris
[01:56:11] *** hile_ has quit IRC
[02:00:02] *** peteh has quit IRC
[02:00:50] *** yydzero has joined #opensolaris
[02:01:11] *** peteh has joined #opensolaris
[02:02:23] *** uebayasi has quit IRC
[02:02:58] *** simford has joined #OpenSolaris
[02:03:24] <alanc-beer> elektronkind: b58
[02:03:40] <alanc-beer> b57 images are already cut and out for internal test
[02:04:48] *** alanc-beer is now known as alanc
[02:07:21] *** sahafeez has quit IRC
[02:10:20] *** uebayasi has joined #opensolaris
[02:14:20] *** Odin-LAP has quit IRC
[02:17:32] *** dlynes_laptop has quit IRC
[02:19:24] *** hspaans has quit IRC
[02:19:28] *** lacaAFK is now known as laca
[02:19:30] *** Vratha has joined #opensolaris
[02:24:27] *** dalibor has joined #opensolaris
[02:26:45] *** steleman_work has joined #opensolaris
[02:28:52] *** steleman_work has left #opensolaris
[02:33:44] *** schily__ has joined #opensolaris
[02:40:44] *** mbirkis has joined #opensolaris
[02:41:55] *** schily_ has quit IRC
[02:42:36] *** dunc_ has joined #opensolaris
[02:45:37] <mbirkis> is it possible to keep the grub from my previous linux install and just add entrys for my fresh solaris express install?
[02:45:55] <alanc> hmm, wonder how long the onnv internal putback -> Hg external gate -> CIA -> IRC bot takes
[02:46:25] <alanc> since I see a new driver moved from usr/closed to usr/src putback from about 3 minutes ago in e-mail
[02:46:29] <CIA-22> xy150489: 6482530 Need Ethernet driver for Intel G965 chipset; 6491179 link aggregation with e1000g does not work unless snoop is running; 6494743 e1000g 5.1.4 driver does not work with s11_52; 6502458 e1000g is open source, move the source from usr/closed to use/src
[02:46:38] <alanc> and there's the answer 8-)
[02:46:53] <alanc> approx. 4-5 minutes
[02:47:10] *** stevel has quit IRC
[02:47:11] <elektronkind> CIA outputs right as I receive mail from onnv-notify
[02:47:36] <_syphilis_> elek: CIA uses push notification
[02:47:41] <_syphilis_> (usually mail)
[02:47:42] <elektronkind> it's probably some email address that's subscribed to onnv-notify and a procmail rule feeds it info when mail comes in
[02:47:50] <alanc> that should make some people happy
[02:47:53] <jbk> mbirkis: i seem to vaguely recall some modifications needed to grub for solaris, so if my memory is right, it probably depends on if your version also incorporates those changes
[02:48:02] <_syphilis_> i think there's an HTTP interface as well
[02:49:26] <mbirkis> jbk, ok thanks... do you know where i can find out more on the subject? i find it kind of hard to find info on solaris related stuff
[02:49:31] <movement> yes, I'm subscribed to onnv-notify and procmail it right into cia's email address
[02:49:58] <movement> and it's awesome that e1000g made usr/src, it was always beyond me why it wasn't!
[02:50:07] <elektronkind> whoa
[02:50:16] <elektronkind> e1000g is nolonger closed-bin?
[02:50:30] <elektronkind> wow
[02:50:37] <jbk> best bet is to post a message on opensolaris.org
[02:50:44] <mbirkis> ok
[02:51:09] <elektronkind> the fruits of "sun+intel"
[02:51:32] <richlowe> movement: there's all sorts of odd randomness.
[02:52:06] <richlowe> I'm more puzzled why that wasn't keyworded.
[02:55:09] <sommerfeld> elektronkind: actually, it was in the works well before the sun+intel announcement.
[03:09:58] *** SquareGuy has joined #opensolaris
[03:11:37] <timeless> so, i'm trying to understand how df and du work :|
[03:11:42] <richlowe> sommerfeld: does that explain why the license header is screwy, and explicitly states what is (iirc) the wrong version?
[03:12:28] <nachox> hmm, what license is the e1000g using? is it like the linux one?
[03:12:52] *** mrdeviant has joined #opensolaris
[03:13:00] <timeless> does zfs do delete operations in the background?
[03:13:11] <timeless>                        362G   466M   226G     1%    /export/home/timeless/nexenta
[03:13:16] <timeless>                        362G   321M   226G     1%    /export/home/timeless/nexenta
[03:13:25] <sommerfeld> richlowe: i'll bring this to the attention of people who tend to care.
[03:13:28] <timeless> between those two lines, i was not performing any deletes..
[03:13:56] <nachox> i ask only because that's the nic some iss appliances use and they did some nice stuff with them
[03:14:45] <sommerfeld> timeless: there was a recent putback which nuked a background delete thread.  so the bits you're using probably do that.
[03:14:58] <timeless> interesting
[03:15:14] <timeless> by nuke, does that mean the app that calls delete pays for dlete?
[03:16:26] <sommerfeld> If I read the bug report correctly, with the recent change, the thread which drops the last reference pays to free the blocks.
[03:18:36] <sommerfeld> richlowe: I looked at the copyright block.  i surmise that intel:CDDL::linus:GPL -- "current license version is ok.  if you have another one we should talk"
[03:19:18] <richlowe> sommerfeld: I believe that 1.1 is current though...
[03:19:26] <sommerfeld> Hmmm.
[03:19:41] <richlowe> sommerfeld: wasn't the wording altered slightly as clarification at some point?
[03:19:44] <richlowe> or am I thinking of something else.
[03:20:16] <richlowe> (I thought that happened about the same time as the header format changed)
[03:20:43] <timeless> sommerfeld: ok
[03:20:43] <sommerfeld> latest version at http://www.opensolaris.org/os/licensing is 1.0
[03:20:50] <timeless> i'm actually not sure which way i prefer
[03:21:01] <timeless> there are obvious benefits to background delete
[03:21:08] <timeless> if you ignore stupid users like me asking "what's up"
[03:21:23] <sommerfeld> delayed free makes rm -rf appear to go fast, and then slow everyone else down for the next N minutes...
[03:21:34] * timeless nods
[03:21:53] <sommerfeld> and there's actually a posix conformance issue with having delayed free visible in statvfs numbers.
[03:21:54] <timeless> normally i'd do zfs destroy
[03:22:09] <timeless> hrm
[03:22:17] <timeless> so the df output i got is "buggy"?
[03:22:24] <sommerfeld> strictly speaking, yes
[03:22:29] <timeless> (and hence the bug fix)
[03:23:47] <sommerfeld> someone posted to comp.unix.solaris a couple years back: delayed free broke some code which did a "delete old logfiles until we hit desired level of free space"
[03:24:02] <timeless> heh
[03:24:11] <richlowe> "oops"
[03:25:58] *** jamesd has quit IRC
[03:26:50] <alanc> I wonder how many people will be frightened that the Xorg build now requires lynx...
[03:26:53] <g4lt-U60> was it simon travaglia
[03:27:00] <timeless> heh
[03:27:09] <timeless> personally i'm more annoyed that git requires curl
[03:27:22] <timeless> (and yes, i know it doesn't really require curl)
[03:27:29] <mrdeviant> what does xorg need lynx for ?
[03:27:32] <timeless> (svn is worse, it requires some https thing)
[03:27:48] <timeless> the configurator? :)
[03:28:13] <timeless> haven't you always wanted to reconfigure X using a web browser? :)
[03:28:23] <richlowe> something twisted and documentation related maybe?
[03:28:24] *** jamesd has joined #opensolaris
[03:28:24] *** ChanServ sets mode: +o jamesd
[03:29:21] *** sahafeez has joined #opensolaris
[03:30:06] <alanc> xorg docs are in docbook in the source tree, translated to text for the packages
[03:30:18] <alanc> docbook2txt is basically "docbook2html | lynx -dump"
[03:31:33] <Auralis> ugh ....
[03:33:19] <alanc> I'm sure there's something much more clever a docbook guru could work out, but that's what the Xorg automakefiles use, so I just stuck with it
[03:35:28] *** hile_ has joined #opensolaris
[03:36:33] <CIA-22> ms148562: 6487584 X will not start - Unable to open /dev/agpgart; 6494695 agpgart: memory binding problem under i915GM chipset
[03:36:59] *** hile_ has quit IRC
[03:38:53] <e^ipi> where does CIA-22 pull these commit logs off of anyways?
[03:39:27] <e^ipi> are they public?
[03:39:54] *** dalibor has quit IRC
[03:40:22] * timeless sighs
[03:40:27] <timeless> mandb is apparently important
[03:41:07] <alanc> e^ipi: http://mail.opensolaris.org/pipermail/onnv-notify/
[03:41:36] <sommerfeld> and there's a mercurial mirror updated in real time.
[03:42:13] <alanc> onnv-notify is the mailing list notified when the mercurial mirror is updated
[03:49:37] <mbirkis> what does "bad PBR sig" meen? i get it when i try to boot, and it hangs
[03:50:15] <CIA-22> gc161489: 6516835 usb warlock is not clean due to CR6463853
[03:50:28] <richlowe> warlock still annoys me.
[03:50:32] <sommerfeld> generally means seriously messed up boot blocks.
[03:50:49] <timeless> can anyone here suggest a man page that suggests a way to fix a problem
[03:50:59] <nachox> PBR means Partition Boot Record, ie the first sector of a partition. so i'm guessing something wrong with the booting secuense
[03:51:14] <timeless> i'm tryhing to write a patch for a man page (debian: man 7 undocumented) suggesting how to recover from an error
[03:51:31] <timeless> but to do  that, i need a sample of a man page that does such a thing
[03:51:54] * timeless looks in man7 man
[03:53:09] <SquareGuy> Well I got OpenSolaris installed.  My disks on the promise controller (Ultra133 TX2 IDE) are not seen.  Reccomend any controllers?
[03:53:44] <timeless> that controller or something like it sounds very familiar, i think my old pc has it or an old relative :)
[03:54:04] *** sahafeez has quit IRC
[03:54:26] <timeless> would DIAGNOSTICS be an ok section?
[03:59:59] *** sartek_ has joined #opensolaris
[04:00:00] *** LeftWing has quit IRC
[04:00:08] *** LeftWing has joined #OpenSolaris
[04:00:09] *** sommerfeld has quit IRC
[04:00:10] *** sartek has quit IRC
[04:03:02] *** slowhog has left #opensolaris
[04:09:28] *** peteh has quit IRC
[04:09:47] *** peteh has joined #opensolaris
[04:13:56] *** whaq_ has joined #opensolaris
[04:14:15] *** whaq has quit IRC
[04:16:32] <Tempt> 23
[04:16:46] <Tempt> oops
[04:16:48] <Tempt> grr.
[04:17:49] <e^ipi> ?
[04:19:25] *** DMJC-L has joined #opensolaris
[04:19:32] <DMJC-L> is there a reiserfs module for solaris?
[04:19:41] <DMJC-L> I need to migrate ~200gb of data
[04:19:44] <Tempt> Nope.
[04:19:52] <e^ipi> no man
[04:19:57] <DMJC-L> it does have ntfs support?
[04:20:00] <dlg> hahaha
[04:20:02] <DMJC-L> read only?
[04:20:03] <dlg> hilarious
[04:20:07] <DMJC-L> I don't need write support
[04:20:17] <DMJC-L> freebsd has read only reiserfs support
[04:20:20] <e^ipi> isn't there a FUSE module kicking around?
[04:20:29] <DMJC-L> dunno, if there is that'd be fine too
[04:20:37] <Tempt> FUSE support on Solaris?
[04:20:50] <e^ipi> *shrug* i thought there was
[04:20:53] <e^ipi> could be wrong
[04:21:07] <e^ipi> i've never bothered to look, all my storage is network attached
[04:21:44] <Tempt> All the storage I care about these days uses ZFS.
[04:22:05] <DMJC-L> hehehe now there's an idea
[04:22:30] <axisys> how does sun java webserver compares to apache ?
[04:22:37] *** SquareGuy has quit IRC
[04:22:40] <DMJC-L> I assume you can emulate linux inside solaris through sometyhing like vmware?
[04:22:47] <e^ipi> the two are not exclusive... you can have network attached ZFS
[04:23:00] <e^ipi> DMJC-L: there's brandZ for running linux binaries
[04:23:10] <e^ipi> but i'm not sure why you'd want to
[04:23:10] <DMJC-L> that's not what I meant
[04:23:17] <_syphilis_> brandz won't support linux filesystems
[04:23:18] <Tempt> e^ipi: True enough. Physical storage is fibre attached, NFS served out to everything else.
[04:23:23] <Tempt> e^ipi: yay for NFSv4
[04:23:28] <DMJC-L> I'll explain..
[04:23:36] <DMJC-L> in windows I was able to boot an ubuntu livecd
[04:23:39] <DMJC-L> in vmware
[04:23:50] <DMJC-L> and access my linux partitions directly, then network share them
[04:23:56] *** nachox has quit IRC
[04:24:01] <DMJC-L> windows was then able to read the drives as network shares
[04:24:19] <DMJC-L> all I need is read support so i mght be able to do the same in solaris
[04:24:56] *** linma has quit IRC
[04:25:08] <e^ipi> can you not find another machine to put this harddrive in?
[04:25:33] <lloy0076> DMJC-L: Take a step back - are you trying to access your  Linux zones' files from within Solaris?
[04:25:34] <DMJC-L> not easily
[04:25:46] <lloy0076> Oh, it's a separate hard drive.
[04:26:10] <DMJC-L> I'm trying to access about 200gb of movies I want to migrate to a raidz array
[04:26:31] <lloy0076> DMJC-L: Get a commodity POS hardware running Linux, mount it using NFS or SMB and copy them over...
[04:26:46] <e^ipi> yeah, that's what i was getting at
[04:27:03] <lloy0076> OR, get GNU Parted to shunt the partition on the 200gb to be smaller.
[04:27:04] <e^ipi> it's by far the best way of getting around it
[04:27:13] <lloy0076> Then make a UFS partition on the 200gb drive.
[04:27:23] <dlg> whats cifs/smb mount support in solaris like?
[04:27:24] <dlg> usable?
[04:27:27] <lloy0076> And then...*hmmm*....hope that Linux can write Solaris UFS?
[04:27:39] <Tempt> dlg: non-existant.
[04:27:42] <Tempt> dlg: use sharity.
[04:27:47] <jbk> right now
[04:27:55] <DMJC-L> does solaris support boot from zfs?
[04:27:59] <e^ipi> Tempt: what are you talking about? samba works fine
[04:28:00] <jbk> there's supposed to be a project to implement that
[04:28:05] <DMJC-L> I was reading somewhere that it was complicated to setup
[04:28:16] <jbk> e^ipi: i think to mount, not share
[04:28:19] <Tempt> e^ipi: samba shares out nicely, but there isn't a mount facility.
[04:28:24] <Tempt> e^ipi: ala smbmount
[04:28:33] <e^ipi> meh, NFS is industry standard anyways
[04:28:43] <e^ipi> even windows has an implementation
[04:28:46] <DMJC-L> heh
[04:29:51] <mrdeviant> nfs and fat, how perfect for one another....
[04:30:00] <lloy0076> DMJC-L: If you really, really wanted to get complicated you could put a 20Gb separate partition on your Solaris machine, format it using PCFS (which I think Solaris can read and write to), boot into Knoppix, copy 20Gb over...boot back into Solaris, copy it to RAIDz...do that about 10 times (10x20 == 200) and then when you're finished add the 20Gb pcfs partition back into a zfs pool.
[04:30:10] <lloy0076> DMJC-L: That sounds horridly complex but it could work.
[04:30:21] <DMJC-L> hehehe
[04:31:25] <e^ipi> or burn a bunch of DVDr's
[04:31:34] <lloy0076> DMJC-L: Or, if you can get Solaris running in a VMWARE image on an Ubuntu/Linux/Something Not Solaris, then you could do a NFS or SMB copy from the host system to the Solaris in the vmware image.
[04:31:51] * lloy0076 thinks that it might be easier to burn DVD's
[04:31:57] <DMJC-L> heh i wonder if samba still works on this box
[04:32:12] <DMJC-L> it'll be down to 28gb of data soon
[04:32:17] *** jsubl2 has quit IRC
[04:32:28] <lloy0076> 28Gb = about 8 DVDs.
[04:32:32] <lloy0076> It's not *that* much.
[04:32:33] <DMJC-L> experimenting with a few things
[04:32:58] <Tempt> Just drop it on a tape.
[04:33:23] <e^ipi> Tempt: if he hasn't got another harddrive kicking around, why do you assume he's got tape?
[04:33:50] <Tempt> e^ipi: Aah, just used to having tape drives kicking around the place.
[04:33:51] <DMJC-L> this isn't datacenter style stuff
[04:33:58] <DMJC-L> this is just poor student heh
[04:34:03] <Tempt> e^ipi: They're always handy when you need to shovel data around.
[04:34:07] <e^ipi> obviously, it's 200g worth of pirated movies
[04:34:09] <DMJC-L> how redundant is raidz?
[04:34:18] <e^ipi> DMJC-L: think raid5
[04:34:21] <DMJC-L> I had raid5 fuck up on linux recently
[04:34:26] <DMJC-L> and it didn't recover properly
[04:34:33] <Tempt> That's linux.
[04:34:38] <DMJC-L> heh
[04:34:41] <e^ipi> yeah, blame linux, not RAID-5
[04:34:53] <DMJC-L> ok
[04:35:00] <DMJC-L> that's what a few people have said
[04:35:09] <DMJC-L> hence why I'm looking into a real unix now
[04:35:09] <delewis> no, software RAID-5 is inherently bad.
[04:35:16] <DMJC-L> how so?
[04:35:22] <delewis> of course, the fact it was the mdtools implementation didn't help things ;-)
[04:35:25] <Tempt> back when I actually had a linux box kicking around the volume manager used to shit me continually.
[04:35:32] <DMJC-L> wait. is this the write hole thing raidz talks about?
[04:35:33] *** crib- has quit IRC
[04:35:42] <delewis> but if you're going to bash Linux, at least provide some amount of justification
[04:35:51] <delewis> DMJC-L, yes.
[04:35:55] <Tempt> kept dropping devices and requiring partition fixes and stuff. ended up cronning a job to snap the first 10mb of the partition
[04:35:56] <e^ipi> DMJC-L: if you lose power in the middle of a write, which copy is correct?
[04:36:01] <e^ipi> the checksum, or the data?
[04:36:09] <DMJC-L> ah heh
[04:36:12] *** crib- has joined #opensolaris
[04:36:23] <delewis> the data; however, the parity has not been updated yet.
[04:36:24] *** Somethingelse has quit IRC
[04:36:25] <e^ipi> zfs fixes that through magic
[04:36:32] <DMJC-L> I'm not sure what happenned
[04:36:33] <delewis> thus when you "restore" your RAID-5 volume, your data is now corrupted
[04:36:41] <DMJC-L> I just was using kde, and the entire thing froze
[04:36:44] <delewis> because of the incorrect parity (it was never updated)
[04:36:51] <DMJC-L> rebooted and it had booted a drive out of the raid
[04:37:00] <DMJC-L> rebuilt the arrya and the FS was stuffed
[04:37:09] <DMJC-L> ran FS rebuild and it was totally boned
[04:37:25] <Tempt> Aah, the joy of it.
[04:37:39] <Tempt> Got a pile of spindles here waiting to become a raid5 set.
[04:37:44] <delewis> this is why you should use RAID-5 with hardware implementations that have battery-backed cache so the parity update is stored in cache if the system loses power.
[04:37:56] <delewis> *never*, *ever* trust your data to software RAID-5.
[04:38:07] <delewis> unless the volume manager is implementing various "hacks"
[04:38:43] <e^ipi> like ZFS does
[04:38:44] <delewis> actually, there aren't any hacks that I know of that software RAID-5 implementations could implement to avoid the write-hole.
[04:38:51] <delewis> e^ipi, that is not a hack. :-)
[04:39:02] <delewis> I'm referring to RAID-5 in the traditional sense.
[04:39:03] *** linma has joined #opensolaris
[04:39:14] <Tempt> Hardware RAID hardware is getting pretty cheap. An A1000 probably cost you $20 on ebay these days.
[04:39:26] <delewis> Tempt, A1000s are no longer supported in Solaris
[04:39:35] <Tempt> delewis: They work Just fine.
[04:39:40] <DMJC-L> I've moved to sata hdds
[04:39:56] <DMJC-L> I assume that makes it more expensive :)
[04:40:45] *** kimc has joined #opensolaris
[04:42:12] <DMJC-L> anyway now I have a decision to make
[04:42:17] <DMJC-L> solaris 10 or solaris 11
[04:42:40] <e^ipi> nevada's more fun, it gets all the features first
[04:43:01] <Tempt> Stability, stability. Unless you need the features just grab 11/06.
[04:43:07] <boyd> Tempt: an A1000 is still overprised at $20 :)
[04:43:16] <DMJC-L> 11/06 hmm
[04:43:22] *** karrotx has joined #opensolaris
[04:43:31] <kimc> i just built a pretty big project on Solaris 11 -its working great
[04:43:38] <DMJC-L> depends on the features I guess
[04:43:54] <kimc> yes
[04:44:01] <DMJC-L> they using gcc 4 at all?
[04:44:07] <Tempt> boyd: Bollocks.
[04:44:10] <Tempt> boyd: Good value.
[04:44:22] <delewis> DMJC-L, no.
[04:44:29] <delewis> and that's (probably) a good thing.
[04:44:30] *** alanc is now known as alanc-away
[04:44:33] <boyd> pfft... I'd prefer a D1000
[04:44:38] <delewis> I've seen instances on SPARC where gcc-4.x generates far worse code.
[04:44:40] <DMJC-L> k heh
[04:44:47] <Tempt> boyd: The difference is one board ;)
[04:44:50] <delewis> than gcc-3.x, that is.
[04:44:51] <DMJC-L> also amd64
[04:44:52] <Tempt> boyd: Swap 'em 'round
[04:44:57] <DMJC-L> 32bit vs 64-bit
[04:45:04] <delewis> DMJC-L, Sun Studio generates superior code, nowadays.
[04:45:06] <richlowe> delewis: I suspect that's why GCCfss is around.
[04:45:09] <delewis> for AMD64, that is.
[04:45:09] <boyd> Tempt: Bu I also need to get the stech of A1000 off me :)
[04:45:12] <richlowe> despite being flamewar bait.
[04:45:15] <DMJC-L> does 64-bit solaris do a thing like gentoo does?
[04:45:22] <DMJC-L> where it can load 32 bit libs??
[04:45:29] <Tempt> boyd: What, did you have an A1000 eat your data for brekkie?
[04:45:32] <delewis> richlowe, yes, I don't think gcc SPARC code generation quality will ever become remotely "decent"
[04:45:45] <delewis> gcc does not simple loop optimizations, even -- for *any* platform.
[04:45:48] <DMJC-L> multilib, that was it
[04:46:07] <boyd> Tempt: I'm being rhetorical, obviously
[04:46:11] <delewis> DMJC-L, Solaris was doing that before Linux supported a 64-bit platform.
[04:46:20] <DMJC-L> nice
[04:46:23] <richlowe> I don't think that's technically true.
[04:46:30] <richlowe> given how early linux's port to alpha was done.
[04:46:30] <CIA-22> Rod Evans: 6519951 bfu is just another word for exit today (RPATH -> RUNPATH conversion bites us)
[04:46:32] *** Theoden-Nexenta has joined #opensolaris
[04:46:35] <richlowe> (though alpha was 64bit only, I'll accept)
[04:46:51] <Tempt> boyd: I've got more spindles turning up tomorrow arvo, so I need to work out an array for them
[04:46:52] <delewis> richlowe, yes, I was dwelling on a few technical points in that statement :-)
[04:47:25] <Tempt> boyd: I'll have a total of 17 146Gb seagates to house.
[04:47:37] <boyd> Tempt: mmmmm noise & power
[04:47:46] <Tempt> That's the one.
[04:48:03] <delewis> Photons are louder
[04:48:16] <Tempt> Photons also eat about 500W before you drop the first spindle in.
[04:48:23] <Tempt> A big price to pay for the little display.
[04:48:26] <Tempt> Hmmm.
[04:48:29] <richlowe> photons are louder than many things.
[04:48:40] <Tempt> Got a few old photon boxes due for the dumpster here, I wonder how easy it would be to reuse that display.
[04:48:41] <delewis> Tempt, I pay about $20/month more if that for mine :-)
[04:48:52] * boyd has actually has worse luck with Photos
[04:48:56] <richlowe> delewis: curse you and your utility company.
[04:49:09] <Tempt> delewis: I've got these HP SureStore-E FCAL JBODs which are a lot leaner on power.
[04:49:19] <delewis> richlowe, yeah, I think my bill has gone up around $60 total since I've gotten the E4500 and A5200.
[04:49:42] <Tempt> Mmm, dee-licious 4500. That box will still be running in 20 years time given ample spares.
[04:50:01] <delewis> yes, it handles a load nicely. :-)
[04:50:14] <delewis> and an A5200 makes a nice storage toy.
[04:50:17] <Tempt> Just don't forget to stock up on parts.
[04:50:28] <Tempt> The photon is a nice storage toy because it comes with a VxVM license.
[04:50:43] *** kimc has quit IRC
[04:50:48] <delewis> Tempt, no, that and the number of disks.
[04:51:05] <delewis> I've got a full VxVM license -- no need for the Photon one that doesn't have VxFS
[04:51:06] <Tempt> Depends on what your stash of FCAL spindles is like.
[04:51:19] <delewis> Tempt, 22x18GB disks (more than I'll ever need)
[04:51:31] <Tempt> Fair cop.
[04:52:08] <Tempt> I don't really need to run a model for storage at home anymore. I just keep a few trays of 9Gbs at work for playing.
[04:52:45] <Tempt> Use my storage for ummn... "valuable data". That's it.
[04:55:36] <DMJC-L> got a url for nevada?
[04:59:41] <delewis> http://www.flickr.com/photos/85894987@N00/sets/72157594511620201/
[05:02:42] *** Theoden-Nexenta has quit IRC
[05:03:16] *** Theoden-Nexenta has joined #opensolaris
[05:06:32] *** laca has quit IRC
[05:09:06] *** gm152 has quit IRC
[05:10:04] <DMJC-L> hmm
[05:10:16] <DMJC-L> so what do i actually download to get a solaris install on my system?
[05:10:29] <DMJC-L> opensolaris website has a lot of things listed
[05:12:51] *** DMJC-L is now known as DMJC
[05:17:48] *** mrdeviant has left #opensolaris
[05:18:30] <elektronkind> pwew
[05:19:35] <elektronkind> where's L1-A on that sparcbook eh
[05:19:58] <delewis> elektronkind, Pause+A
[05:20:09] <DMJC> hmm
[05:20:17] <DMJC> is the amd64 version of solaris out yet?
[05:20:32] <delewis> DMJC, uh, Sun has been shipping AMD64 hardware for 2 years now.
[05:20:33] <DMJC> the solaris express dvd only seems to come in x86 and sparc versions
[05:20:48] <delewis> DMJC, why do you need a separate media for AMD64?
[05:21:00] <DMJC> 64-bit binaries?
[05:21:01] <delewis> the x86 media contains the packages that benefit from being compiled for AMD64.
[05:21:04] <DMJC> ok
[05:21:52] <delewis> very few binaries in Solaris are actually 64-bit.
[05:22:05] <delewis> kernel stuff (drivers, etc.), libc, libm, etc.
[05:22:15] <richlowe> I wouldn't call it "very few"
[05:22:24] <dlg> i wouldnt call those binaries per se either
[05:22:34] <delewis> richlowe, not even when you compare it to the number of binaries compiled for 32-bit?
[05:22:49] <delewis> dlg, I don't see why not. libs are binaries, as well.
[05:23:22] <dlg> unless a 64bit program is linked against them, theyre not used
[05:23:28] <dlg> kernel modules are fun though
[05:23:30] <DMJC> I don't mind if most of the system is 32-bit
[05:23:35] <DMJC> as long as I get the option
[05:23:50] <DMJC> stuff like mplayer is practically useless in 64-bit
[05:24:14] <delewis> DMJC, on AMD64, I'm not so sure about that.
[05:24:23] *** yydzero has quit IRC
[05:24:24] <delewis> 64-bit gives you twice as many general-purpose registers.
[05:24:33] <DMJC> mplayer uses win32 codecs
[05:24:42] <DMJC> it can use 64-bit ones but noone is making them
[05:24:42] <delewis> DMJC, those are thinning.
[05:24:51] <DMJC> there's REAL, and one other
[05:25:08] <delewis> there's an FFmpeg implementation of wmv, nowadays.
[05:25:09] <DMJC> afaik windows media player doesn't have 64-bit codecs yet
[05:25:13] <delewis> Solaris ships Real.
[05:25:18] <delewis> so that's fairly irrelevant.
[05:25:51] <delewis> DMJC, modern MPlayer does not use Windows Media Player codecs even if you have win32codecs installed.
[05:25:52] <_syphilis_> solaris ships java media player!!!*(£$%(*£
[05:26:00] * delewis throws stones at _syphilis_
[05:26:02] <richlowe> hah.
[05:26:15] <DMJC> delewis, I still have media files that need win32codecs
[05:26:22] <DMJC> they have a LOT of media player implemented
[05:26:25] <DMJC> but not all of it yet
[05:26:38] <DMJC> my last mplayer build was from about 2 weeks ago
[05:26:41] <delewis> DMJC, most of the things that aren't implemented, won't be implemented.
[05:26:50] <delewis> for various reasons
[05:26:55] <delewis> DRM should be obvious.
[05:26:58] <DMJC> yeah heh
[05:27:02] <DMJC> it wasn't drm though
[05:27:09] <DMJC> I had a file where audio played and video didn't
[05:27:12] <DMJC> and vice versa
[05:27:16] <elektronkind> DMJC: I think I see where you're coming from there... Solaris boots a 32bit kernel if it's not on EM64T/AMD64 hardware. It will boot 64 bit if it is. But unlike with most (every?) Linux distro, a comeplete 32bit environment can run under a 64bit kernel... and most binaries remain 32bit... but most every library in /usr/lib and so on has both 32 and 64 versions... so 32bit and 64bit apps run side-by-side.
[05:27:19] <DMJC> in 64-bit and 32-bit mplayers hehehe
[05:27:44] <DMJC> yeah elektronkind that's how gentoo does it
[05:27:57] <delewis> I need to get around to cleaning up a type conflict in lavc when compiling MPlayer in 64-bit mode.
[05:27:58] <DMJC> it's great, gives you lots of options for compatability vs performance
[05:28:01] <delewis> but alas, I haven't cared much lately.
[05:28:02] <elektronkind> yah
[05:28:54] <delewis> MPlayer does silly things with the large file interfaces, too (mostly because hardly anyone, including glibc has them properly implemented)
[05:29:00] <elektronkind> Solaris has been tha way since 64bit capabilties first appeared in it 8 or 9 years ago.
[05:29:13] <elektronkind> with Solaris 7
[05:29:25] <elektronkind> and what a stepchild that was
[05:29:35] <_syphilis_> 7 was the first 64-bit versions?  i thought ultrasparc was older than that
[05:29:36] <jbk> heh
[05:29:41] <delewis> _syphilis_, it is.
[05:29:44] <jbk> 7 i think first had 64-bit userland
[05:29:49] <delewis> but keep it mind the UltraSPARC-I was never run in 64-bit mode.
[05:29:51] <jbk> 2.6 i believe had 64-bit kernel
[05:30:14] <DMJC> heh these servers are so SLOW
[05:30:17] <elektronkind> _syphilis_: UltraSPARC is older than that. I got my firs US system in 1996.... but a fully 64bit Solaris didn't happen until 1999 with Solaris 7
[05:30:30] <elektronkind> 2.6 had 64bit *interfaces*
[05:30:39] <elektronkind> such as large file support
[05:30:46] <Stric> DMJC: they were blindingly fast when they came
[05:30:49] <delewis> elektronkind, and even so, a large chunk of the product line (UltraSPARC-I hardware) couldn't run in 64-bit mode (well, not safely, anyway)
[05:31:01] <elektronkind> yeah, the US-1 bug
[05:31:02] <DMJC> not the old stuff I mean the download of solaris i'm doing
[05:31:16] <_syphilis_> DMJC: sdlc is variable, i've gotten fairly fast speeds from it at times
[05:31:21] <elektronkind> bummer that was, but I gave away my Netra 150 ages ago so I don't care
[05:31:38] * delewis setup a Netra 150 on a real VT100 several weeks ago.
[05:31:44] <delewis> interesting experience that was :-)
[05:32:02] <elektronkind> you're kidding. There's actually a Netra 150 that is still in working order?
[05:32:15] <elektronkind> that was a styrofoam-filled POS imo
[05:32:26] <delewis> elektronkind, yes, my university has been using it as a mail server since the Netra 150s were released.
[05:32:33] <delewis> it was taken out of production about two years ago
[05:32:42] <delewis> and one of the faculty kidnapped it for a MySQL/PHP server for a class.
[05:32:55] <elektronkind> wow, well they certainly got their money's worth
[05:32:58] <delewis> which I had the honor "bestowed" upon me for setting it up.
[05:33:42] <elektronkind> I seem to recall that the CPU fans on those US-1 "Proton" motherboards had a penchent for seizing up
[05:33:53] <elektronkind> I ran one as a nntp server for a while
[05:34:04] <delewis> the only problem I had with it was the PROM battery ;-)
[05:34:23] <elektronkind> heh. deadbeef! badcafe!
[05:35:07] <delewis> that and getting accustomed to the DEC keyboard layout
[05:35:20] <_syphilis_> did the vt100 have an esc key?
[05:35:36] <delewis> _syphilis_, yes, but nowhere near where it should be :-)
[05:35:55] <delewis> made working with vi a PITA.
[05:36:08] * _syphilis_ has had the opposite problem, translating DEC keys into PC keyboard
[05:36:15] *** qbit has joined #opensolaris
[05:36:18] <_syphilis_> (not too hard until you come to "Press Gold...")
[05:36:19] <delewis> _syphilis_, ah.
[05:36:49] <elektronkind> I had a VT320 or something like that.. amber/orange CRT is all I can remember... but the keyboard was big enough to double as a ping pong ball table
[05:37:22] <delewis> yes, and on a VT100 you modify settings by setting various bits on the keyboard.
[05:37:53] <delewis> fortunately, someone had some old VT100 documentation laying around in a closet.
[05:37:59] <_syphilis_> someone wrote an Xresources to remap xterm to send VT escapes, which was quite useful
[05:38:16] <delewis> _syphilis_, I bet.
[05:42:55] *** adlpaf has joined #opensolaris
[05:43:32] *** adlpaf has quit IRC
[05:48:27] *** SunTzuTech has joined #opensolaris
[05:49:44] *** SquareGuy has joined #opensolaris
[05:50:28] <SquareGuy> how do i list all physical drives in opensolaris
[05:50:35] <_syphilis_> format
[05:51:14] <dlg> ls
[05:51:59] <SquareGuy> hmmm this is weird, i see the two drives on the onboard controller.  but by guesssing i have added 2 of the five drives on external controllers.
[05:52:19] <elektronkind> iostat -E is also a way
[05:52:21] <SquareGuy> i mean adding them to a pool
[05:52:36] <elektronkind> iostat -E doesnt really give pretty output tho
[05:53:26] <SquareGuy> looks like my pci controllers are not setup yet
[05:54:02] <SquareGuy> strange that i was able to add two drives to a pool though that are on it
[05:54:18] <lloy0076> Isn't it ironic that ZFS is actually quite easy to use, but its "User Guide / Book" (as produced by Sun) is 180 pages long?
[05:54:45] <delewis> lloy0076, the Veritas docs are much worse, but that's not really surprising.
[05:54:58] <jbk> heh
[05:55:05] <SquareGuy> Well I will get the controller up and going then give it another go.  Yeah ZFS is easier than anything I have seen.
[05:55:06] <jbk> don't get me started on veritas docs :)
[05:55:07] <delewis> I'd go as far to say that given the simplicity of ZFS, its docs are quite proportional to the Veritas ones.
[05:55:35] <delewis> installation guide, VxVM guide, VxFS guide, etc. etc.
[05:55:48] <elektronkind> I showed one of my devs how to do snapshots and clones. His response was "This is mighty."
[05:56:16] <Stric> iostat -E  doesn't show up-to-date info if you've been swapping around drives while up (scsi)..
[05:56:27] <Stric> while format re-probes it seems
[05:56:34] <SquareGuy> I plan on doing another install, just playing with it now.  The next install will be GUI-less.
[05:57:44] *** DMJC has quit IRC
[05:58:23] <SquareGuy> I will say that this reminds me of years past about Linux hardware and passing parameters.  I was hoping for plug and play with my IDE controller but believe me I am not one to complain.  This is sweet.
[05:59:13] *** deather_ has joined #opensolaris
[06:00:41] *** DMJC has joined #opensolaris
[06:00:49] <SquareGuy> I do have a question about hardware that I have not been able to find.  The onboard USB is inop.  Complains during boot (forget error) and no USB kb/mouse.  I plug into PCI USB controller and all is well.  Via chipset dual 933 P3... ASUS board, forget model#
[06:02:10] *** Theoden-Nexenta has left #opensolaris
[06:02:35] <SquareGuy> CUV4X-D is the model
[06:02:45] <richlowe> did the error relate to "No SOF interrupt"?
[06:02:53] <richlowe> (for the USB that is)
[06:03:18] <SquareGuy> yes that sounds right
[06:03:36] <richlowe> I've seen that on older via USB chips, but I've yet to get to the bottom of why.  I found that putting the USB in legacy mode caused it to work (though I think that's generally not a great idea)
[06:04:16] <SquareGuy> well it is only for a kb/mouse.  the pci USB controller is for external storage.
[06:05:10] <SquareGuy> Once setup the box will run headless.
[06:05:16] *** karrotx is now known as __
[06:05:50] *** __ is now known as karrotx
[06:06:07] *** yongsun_ has quit IRC
[06:06:16] <SquareGuy> richlowe: Thank you for the info though.
[06:07:56] <SquareGuy> I must admit I feel quite lost trying to manage Solaris coming from a Linux world.  Almost everything is different it seems.  Like where is my lspci and ls usb hehe.
[06:08:33] *** Gman is now known as GmanAFK
[06:08:36] <elektronkind> lspci... try /usr/X11/bin/scanpci
[06:08:57] <Stric> cfgadm -lav
[06:09:16] <SquareGuy> Sweet, thank you!
[06:09:30] <elektronkind> cfgadm is only for removeable devices...
[06:09:44] <Stric> which usb is ;)
[06:10:23] <SquareGuy> ahhh now I have my string to pass to fire up the IDE controller.
[06:10:26] <elektronkind> oh, right... but in terms of seeing PCI devices... scanpci or prtconf -v (for the masochistic)
[06:11:07] <SquareGuy> I am not very masochistic I prefer cfdisk over fdisk and nano over vi
[06:11:37] <Stric> I'm not masochistic either, so I prefer vim over nano ;)
[06:11:51] <SquareGuy> hehe never used that
[06:11:59] <jbk> could use ex
[06:12:02] <jbk> ;0
[06:12:04] <jbk> :)
[06:12:15] <elektronkind> I've, er, graduated to TextMate now
[06:12:31] <dlg> im not a masochist, thats why i learnt vi
[06:12:38] <dlg> its on every unix
[06:13:06] <SquareGuy> uhhh try edlin ;)  ok not *nix but hey.
[06:14:25] <lloy0076> SquareGuy: The problem is the hole is round.
[06:14:32] * lloy0076 ducks for cover
[06:15:06] <SquareGuy> I use large hammers to gently fit square pegs into round holes.
[06:16:14] <SquareGuy> Thanks for the help all, greatly appreciated.  Maybe my transition won't be too painfull.  I gotta get some sleep.
[06:16:20] *** deather has quit IRC
[06:19:08] *** SquareGuy has quit IRC
[06:24:44] *** postwait_ has joined #opensolaris
[06:25:08] <postwait_> So, I'm using the latest SX and I'm trying to use SVN
[06:25:21] <richlowe> ok...
[06:25:31] <postwait_> I have a metadb slice on two drives... and it works fine. (they are 10k 72GB Seagates)
[06:25:38] <richlowe> Oh, SVM not SVN
[06:25:41] <richlowe> then I'm less useful :)
[06:26:03] <postwait_> I have another identical machines (except it has 15k 72GB)
[06:26:09] <postwait_> (yes SVM sorry, typo)
[06:26:24] <postwait_> And I get errors about SVM metadb overlap.
[06:26:43] <postwait_> I have 8 of these 15k drives and the issue happens with all of them. (any two)
[06:26:53] <postwait_> I'm at a loss.
[06:27:13] <postwait_> I put linux on and can initialize md raid 1 and it "says" it works.
[06:27:38] <postwait_> I can't track any reasonably similar bugs on bugs.o.org
[06:28:13] <e^ipi> big surprise
[06:28:33] *** zaurus has quit IRC
[06:28:37] <delewis> postwait_, sounds like the slice layout is not identical.
[06:28:48] <delewis> if the drives are identical as you claim, use prtvtoc and fmthard.
[06:28:51] <e^ipi> boo-urns for b.o.o.
[06:29:04] <richlowe> the easiest way to search b.o.o is via google.
[06:29:11] <e^ipi> heh
[06:29:12] <richlowe> site:bugs.opensolaris.org <what you're looking for>
[06:29:13] <postwait_> I checked the VTOCs
[06:29:19] <richlowe> I'm quite serious, I just wish I wasn't.
[06:29:25] <postwait_> and they match cylinder for cylinder
[06:29:30] <e^ipi> richlowe: i know you're serious
[06:29:31] <delewis> postwait_, then use prtvtoc and fmthard
[06:29:37] <delewis> you should be able to metadb then
[06:29:42] <postwait_> prtvtoc showed the same
[06:29:45] <e^ipi> doesn't google sell little 1u google appliances?
[06:29:46] <postwait_> I'll look into fmthard
[06:29:58] <e^ipi> maybe sun should invest in a couple of them
[06:30:10] <delewis> postwait_, prtvtoc /dev/rdsk/c0t0d0s2 | fmthard -s - /dev/rdskc0t1d0s2
[06:30:28] <delewis> this copies the VTOC from c0t0d0 to c0t1d0.
[06:30:35] <Tempt> Hmmm
[06:30:49] <Tempt> Missed the fault (just came back) but it sounds interesting. What's the problemo?
[06:30:57] <postwait_> We did that from  working machine and it didn't work
[06:31:09] <delewis> postwait_, define "didn't work"
[06:31:10] <postwait_> too the prtvtoc and then moved it over and applied it
[06:31:22] <delewis> postwait_, then the drives are not identical.
[06:31:27] <postwait_> note that the prtvtoc on the "non working" machine was the same before and after.
[06:31:29] <richlowe> which kind of overlap are we talking about?
[06:31:40] <postwait_> not work meaning the SVM says it has overlaps
[06:31:49] <richlowe> (that the slices overlap, or that SVM says there's overlaps?)
[06:31:52] <richlowe> I believe they have different meanings.
[06:32:07] <postwait_> that the metadb overlaps with another slice
[06:32:18] <delewis> richlowe, both of which shouldn't happen if the drives are identical and he's using prtvtoc and fmthard.
[06:32:39] <richlowe> doesn't SVM complain similarly if you try to create a device composed of metadevices on the same disk, or put metadbs on the same device as metadevices?
[06:32:40] <postwait_> I have a rack of 12 machines that work fine.
[06:33:13] <karrotx> racks are for sissies and girls
[06:33:23] <postwait_> the boxes I'm using as a model were jumpstarted.
[06:33:30] <postwait_> they all have their metadb on s3
[06:33:39] <delewis> ugh.
[06:34:01] <postwait_> which doesn't appear to be "managed" by SVM -- its the metadb..
[06:34:02] <delewis> I guess s3 isn't too bad, but generally, you want metadb on s7.
[06:34:21] <postwait_> jumpstart put it on s3 automatically.
[06:34:30] <postwait_> maybe an issue with our jumpstart config there ?
[06:34:37] <postwait_> but that shouldn't be causing this problem.
[06:34:38] <Tempt> delewis: you can push jumpstart around more than that; tell it what you want.
[06:35:11] <delewis> postwait_, what slice the metadb is on is irrelevant. Its just fairly standard to use s7.
[06:37:20] *** MaGre has quit IRC
[06:41:56] *** spike723_ has quit IRC
[06:42:45] *** spike723_ has joined #opensolaris
[06:42:53] *** SunTzuTech has left #opensolaris
[06:43:38] <postwait_> I guess the code's out there now.
[06:43:51] <postwait_> So I can see what code path leads to my error message.
[06:44:18] <postwait_> figure out what bit of disk geometry is making this stuff go wonky,
[06:45:08] <Tempt> hmm, overlaps. Hmm. Are you using EFI labels?
[06:48:36] *** tsoome has quit IRC
[06:52:26] <postwait_> EFI labels?
[06:54:25] <postwait_> I'm on opteron. and they are SCSI.. so possibly??? I have no idea how to tell.
[06:54:58] <e^ipi> if you use ZFS on whole disks, you use EFI
[06:55:14] <postwait_> I am not on these guys
[06:56:42] <postwait_> perhaps they were once ZFS'd?
[06:56:56] <postwait_> and now they have EFI labels and they are staying as EFI...
[06:57:19] <postwait_> I see how I can _label_ it as SMI or EFI, but I don't see how I can print the current label type.
[06:57:42] <richlowe> if my memory is right, an EFI labelled disk will have 'one too many' slices.
[06:59:09] <postwait_> looks like I have an SMI label.
[06:59:36] <postwait_> prtvtoc will not show cyliners for EFI... mine look slike the SMI output from docs.sun.com
[07:01:20] *** bank__ has joined #opensolaris
[07:01:34] <postwait_> I don't the machine in question on right now .. it's a work.
[07:01:43] <postwait_> I'll take all this information and dig some more tomorrow.
[07:01:48] <postwait_> Thanks for shooting around ideas.
[07:01:50] <bank__> hi
[07:02:05] *** postwait_ has quit IRC
[07:02:31] *** bank__ is now known as bank
[07:02:36] <lloy0076> bank: Howdy :)
[07:04:21] <bank> :) at least I got em now http://61.19.242.221:1158/em
[07:07:33] <bank> :) how are you lloy0076
[07:08:00] <lloy0076> Looks good.
[07:08:19] <lloy0076> I'm fine; I rejigged my home system and put /opt and /export onto ZFS pools.
[07:17:07] *** lloy0076 has quit IRC
[07:17:16] <e^ipi> any muslims kicking around?
[07:17:25] <e^ipi> is eggs & cheese halal?
[07:20:57] *** anthony79 has joined #opensolaris
[07:23:14] <whaq_> eggs should be halal
[07:23:40] <e^ipi> no generally accepted prohibition on them in the same dish?
[07:23:47] <e^ipi> ( just curiosity )
[07:24:00] <whaq_> cheese also. they just can't eat those w/ pig and dog meats i think..
[07:25:32] <e^ipi> as i understand it certain combinations of food ( eggs & cheese is one IIRC) are not kosher
[07:25:42] <e^ipi> not sure if the same holds true of muslim dietary restriction
[07:26:12] *** SunTzuTech has joined #opensolaris
[07:26:25] <whaq_> it's probably there to maintain good taste (ha)
[07:26:33] <e^ipi> you don't like quiche?
[07:26:46] <whaq_> it's okay
[07:27:24] <whaq_> the recipe has potential, maybe i just haven't had a good implementation of it yet
[07:27:53] <e^ipi> *shrug*
[07:29:30] <elektronkind> yaar
[07:33:09] <whaq_> 's revenge.. a pretty good game..
[07:33:44] <elektronkind> can you be  pirate in it?
[07:35:54] *** kAv_ has joined #opensolaris
[07:36:40] *** kavie has quit IRC
[07:36:52] <boyd> I thought the kosher separation was because of food hygene... Meat and Dairy separation is the main one I can think of
[07:37:11] <elektronkind> no shellfish
[07:37:23] <elektronkind> no cloven hoove
[07:37:38] <elektronkind> no mixing of mother's milk and meat
[07:37:41] <whaq_> nah, you're in some sort of space-bug--ship..
[07:37:56] <e^ipi> most of the kosher rules make sense if you imagine yourself in a time pre-sanitation
[07:38:05] <boyd> I'm sure those originated from the likelyhood of food poisoning...
[07:38:07] <e^ipi> milk's got a lot of bacteria, meat spoils quickly
[07:38:10] <elektronkind> in fact if you'er really observant, you have separate plates for meats
[07:38:21] <boyd> Pork is subject to many parasites
[07:38:22] <elektronkind> like, only for meats. and preparation surfaces, pots, and pans
[07:38:26] <e^ipi> parasites in pork can live through cooking
[07:38:29] <whaq_> e^ipi: have you played a game called 'The Shivah'? It's about a rabbi in a crisis of faith (I'm not shittin)
[07:38:36] <e^ipi> shellfish gets sketchy *really* quick
[07:38:55] * boyd thinks shellfish is sketchy at the best of times :)
[07:39:11] <elektronkind> basically, if it's from the sea it must have scales
[07:39:18] <elektronkind> no scales, not kosher
[07:39:19] * whaq_ is with boyd re: shellfish.
[07:39:35] <e^ipi> i'm a vegetarian, so it's pretty much moot anyways
[07:39:40] <e^ipi> also, an atheist
[07:39:53] <whaq_> then the world is moot for u?
[07:39:58] <boyd> So a genetically modified abalone with scales would be ok? :)
[07:40:31] * boyd is an athiest too... and he wouldn't touch an abalone with scales :)
[07:41:51] <e^ipi> whaq_: who said anything about that? i have a strong sense of purpose quite aside from the existance or not of a deity... that said, if you do believe in a god, good for you *shrug*
[07:42:10] <boyd> +1
[07:42:25] *** Symmetria has joined #opensolaris
[07:42:30] <whaq_> haha take it easy
[07:43:06] <quasi> http://en.wikipedia.org/wiki/Richard_Dawkins for one who doesn't agree ;)
[07:43:52] <boyd> Actually, I'm more +1 for the sense of purpose and Richard Dawkins on the latter
[07:44:15] <e^ipi> heh, dawkins
[07:45:43] *** dunc_ has quit IRC
[07:45:44] <e^ipi> quite aside from whether i agree with him on all points, i will give him that he's highlighting the plight of atheists as of late
[07:48:21] <boyd> Indeed he is.
[07:48:30] * boyd is off to catch a train. Evening all
[07:48:32] *** djgregor has quit IRC
[07:57:45] <whaq_> take it easy, don't sit on the roof of it
[07:59:06] <e^ipi> ?
[07:59:48] <whaq_> some dozen of ppl got their head chopped off in India when they sit on top of the roof of their train
[08:00:28] <whaq_> evidently it's a common practice in india (and some of its neighboring countries).. but on that special day, they're traveling for a religious event, and some ppl brought flagpoles and the likes..
[08:00:37] <whaq_> ..which got the attention of the nearest powerline.
[08:00:46] <e^ipi> i've always wanted to visit india
[08:01:19] <whaq_> i just wanna go see Taj Mahal..
[08:03:54] <elektronkind> I'm heading there next january
[08:04:16] <e^ipi> taj mahal'd be neat
[08:04:20] <elektronkind> bombay, jaipur, delhi, calcutta, and dakar
[08:04:25] <e^ipi> but all over india would be cool
[08:04:35] <elektronkind> goa++
[08:04:48] <elektronkind> we need to have a conference or something in goa
[08:04:50] <g4lt-U60> I believe they have a special name for people riding on the tops of trains "hoboes"
[08:05:12] <e^ipi> why goa?
[08:05:37] <elektronkind> goa is the best beach. very international scene there, too. tons of beach parties
[08:05:40] *** gisburn has quit IRC
[08:05:54] <elektronkind> it was a portugues colony up until the early 70's
[08:05:56] <g4lt-U60> besides, there's zero distance to goa ;P
[08:06:04] <elektronkind> oh man
[08:06:10] <elektronkind> you know distance to goa
[08:06:25] <elektronkind> g4lt you goahead
[08:06:32] <whaq_> g4lt is in goa?
[08:06:33] <e^ipi> meh, touristy beach destinations never attracted me
[08:06:40] <whaq_> Bali is nice
[08:06:51] <e^ipi> i've got 0 interest in seeing coastal mexico, for example
[08:07:32] <elektronkind> well touristy places are where the best beaches are, otherwise you're literally in a trash heap
[08:07:35] <g4lt-U60> elektronkind, with the exception of a few friends, I'm probably the only one that does within a 150 mile radius :(
[08:07:59] <e^ipi>   i don't need to spend thousands of dollars to fall asleep on a beach
[08:08:10] <e^ipi> i can do that here for the $3 in gas it'd cost to get to a beach
[08:08:25] <elektronkind> oontz oontz. I missed simon posford in philly on nye :/
[08:09:28] <elektronkind> e^ipi: you're in CA?
[08:09:35] <e^ipi> at the moment i
[08:09:40] <e^ipi> 'm in the middle of nowhere
[08:09:44] <e^ipi> usually i'm in vancouver
[08:10:18] <elektronkind> ah vancouver. my wife is entertaining the a job offer there at UBC
[08:10:23] <elektronkind> s/the//
[08:10:24] <e^ipi> it's a good school
[08:10:35] <elektronkind> she's from Fernie, out in eastern BC
[08:11:04] <e^ipi> she should take it... we can get a VanOSUG going on
[08:11:10] <elektronkind> thing is, I know of jack squat jobs for me in vanc
[08:11:23] <e^ipi> do you like video games?
[08:11:44] <e^ipi> because there's TONS of electronic entertainment positions in van
[08:11:53] <elektronkind> not really. I mean, I like a few, but don't own a PS3... or a 2... or a Xbox for that matter
[08:12:00] <elektronkind> I'm mainly a storage guy
[08:12:50] <e^ipi> i'm sure you could find something
[08:13:03] <e^ipi> vancouver's a big city with a big internet link straight in to japan
[08:13:35] <elektronkind> yeah, and CANNET. I bet I can find something in academia, but I'm just about getting my fill of that here in the US
[08:14:14] <elektronkind> if UBC has a HPC center, I'd check it out in a heartbeat
[08:14:32] <e^ipi> i think they're a member of HPCVL
[08:14:57] <e^ipi> so all that sort of stuff is taken care of in toronot
[08:14:58] <bank> dclarke
[08:15:00] <e^ipi> toronto
[08:15:21] <e^ipi> UBC is a robotics school, the rest of compsci seems to be SFU's domain
[08:15:33] <elektronkind> gotcha
[08:15:48] <elektronkind> and there's always the goverment over in victoria
[08:16:53] <e^ipi> don't underestimate how much of a pain in the ass the ferry is
[08:17:20] <e^ipi> my brother commuted to victoria for 3 months... he couldn't handle it
[08:17:46] <e^ipi> waking up at 4:30 to catch the 5:30 ferry to get to work on time, leaving, and coming home at 9pm
[08:20:39] *** Dar has quit IRC
[08:21:43] <elektronkind> yeah, good point. I guess it wuld be too much to ask for a helicopter stipend ;)
[08:22:22] <e^ipi> heh
[08:22:26] <elektronkind> plus those ferrys seem to like bumping into docks alot
[08:22:27] <e^ipi> usually
[08:22:45] <e^ipi> that only happened this one time
[08:28:42] <elektronkind> hey g4lt-U60, since you're into that stuff, I have a few dj mixes you might dig
[08:28:46] <elektronkind> http://elektronkind.org/mixes/
[08:30:56] *** alfism_ has joined #opensolaris
[08:31:10] *** alfism_ has left #opensolaris
[08:34:01] *** anthony79 has quit IRC
[08:34:50] *** bbtm has quit IRC
[08:35:54] *** bbtm has joined #opensolaris
[08:38:51] *** mazon is now known as Mazon
[08:43:55] *** qdk has quit IRC
[08:49:34] <bank> see ya
[08:49:35] *** bank has left #opensolaris
[08:52:47] *** Odin-LAP has joined #opensolaris
[08:55:12] *** qbit has left #opensolaris
[09:01:41] *** tsoome has joined #opensolaris
[09:02:31] *** setuid has joined #opensolaris
[09:04:11] *** Cyrille has joined #opensolaris
[09:06:28] *** setuid has left #opensolaris
[09:08:36] *** linma has quit IRC
[09:09:01] *** mustang has quit IRC
[09:09:53] <Vratha> solaris have anything like "fast IPC" (i.e. shared memory-mapped IPC message passing?)
[09:10:14] <e^ipi> intimate shared memory ?
[09:10:25] <Vratha> and hopefully with some security built in for client/server-like apps running on a single system?
[09:10:37] <Vratha> e^ipi: no, not like threads
[09:10:55] *** setuid has joined #opensolaris
[09:11:19] <setuid> Does anybody have the nx6125 laptop?
[09:11:48] <e^ipi> isn't ISM between processes?
[09:12:02] <Vratha> as in IPC structures in memory shared between two processes... but when a "server" is working on the structure, the client is blocked when it tries to write to the memory
[09:12:10] <Vratha> ISM?
[09:12:23] <e^ipi> intimate shared memory
[09:12:39] * Vratha looks this up
[09:12:51] <e^ipi> Oracle apparantly uses it a lot
[09:13:45] <Vratha> does it have security mechanisms though?
[09:13:54] <Vratha> because i'm expecting to have untrusted clients access local servers
[09:14:17] <Vratha> and by security, i mean blocking clients when they access the memory while the server is working on it
[09:14:23] <e^ipi> i don't know, look it up
[09:14:26] <Vratha> so that the client can screw around with the server
[09:14:27] <Vratha> i did
[09:14:30] <Vratha> it makes no mention of it
[09:15:40] <Vratha> oh, bleh.. sun themselves say ISM sucks in terms of security
[09:15:49] <e^ipi> well, there you go then
[09:15:55] <Vratha> i guess i'll have to implement my own fast IPC unless there is a lib i find to do it for me
[09:16:04] <Vratha> though since i need it fast at the kernel level i might have to do it
[09:16:05] <e^ipi> there probably is
[09:16:15] <e^ipi> i just don't know of it
[09:16:18] <Vratha> yeah, i'm not sure about that.. it'd require kernel patches
[09:17:09] *** Mazon is now known as mazon
[09:17:13] <e^ipi> i dunno, judging by the market sun's going for, i'd be surprised if you couldn't find some facility or another to do what you want
[09:17:22] <Vratha> perhaps
[09:17:28] <Vratha> what market are they going after?
[09:18:18] <e^ipi> big servers
[09:18:29] <e^ipi> that theoretically may need big IPC
[09:18:40] <e^ipi> 1 sec, lemme find my copy of S:I
[09:18:47] <Vratha> S:I?
[09:18:52] <e^ipi> solaris internals
[09:18:55] <Vratha> oh ok
[09:19:12] <e^ipi> oh, shit
[09:19:15] <e^ipi> i'm retarded...
[09:19:16] <e^ipi> doors
[09:19:20] <e^ipi> you're looking for doors
[09:19:27] * Vratha looks up doors
[09:19:52] <e^ipi> i think anyways, unless i misunderstand what you're looking for
[09:20:07] <Vratha> that isn't a java thing is it
[09:20:12] <Vratha> i'm looking for something at the system level
[09:20:27] <e^ipi> no, it's not a java thing
[09:21:24] <Vratha> ok, good.. yeah, i just found something on sun.com talkinga bout it
[09:21:35] <e^ipi> some combination of doors & sysv shared memory ought to do what you're looking for
[09:21:50] <Vratha> does that S:I (book?) cover the kernel and overall structure of the OS?  if so, i'll probably buy it if it's some book
[09:21:57] <e^ipi> oh yeah
[09:22:00] <e^ipi> it's a fantastic book
[09:22:08] <Vratha> is it very up-to-date?
[09:22:13] <e^ipi> yep
[09:22:16] <Vratha> sweet
[09:22:24] <Vratha> i'll buy it on amazon in a month or so
[09:22:41] <e^ipi> it was just published something like 8 months ago
[09:23:09] <Vratha> great :)
[09:23:27] <e^ipi> http://www.amazon.com/Solaris-TM-Internals-OpenSolaris-Architecture/dp/0131482092/sr=8-1/qid=1170318183/ref=pd_bbs_sr_1/102-3013879-7097753?ie=UTF8&s=books
[09:23:30] <Vratha> it's just that i need that in order to bring over my userspace driver framework and to increase its performance
[09:23:54] <Vratha> damn; those are some good reviews
[09:24:01] *** linma has joined #opensolaris
[09:24:06] <e^ipi> it's a good book
[09:24:07] <Vratha> ooooo, and it's 1072 pages of OS goodness!
[09:24:25] <Vratha> e^ipi: thanks for the hookup on that title and name of the solaris tech
[09:24:31] <e^ipi> *nod*
[09:24:57] <Vratha> yeah, i would've thought they'd have something for what i want... seems like such an obvious thing to want
[09:25:05] *** qdk has joined #opensolaris
[09:25:08] <Vratha> minix has no fast IPC mechanisms though
[09:25:30] <e^ipi> minix isn't designed for massive databases
[09:25:34] <Vratha> e^ipi: however, i'm still having trouble finding info on it.. can you just do one small check for me and see if it implements any kind of security or protection from multiple access?
[09:25:47] <Vratha> yeah, well, minix still needs fast IPC to make the whole system faster
[09:26:08] <Vratha> they do a lot of copies rather than sharing memory that has a shared IPC struct in it
[09:26:29] <Vratha> though when you do that you need to make it so a server has that page locked and can expect no other process can modify that memory while it's working on it
[09:26:56] <e^ipi> can't you just call a mutex on the memory?
[09:27:17] <Vratha> yeah, but then that means your clients have to cooperate
[09:27:27] <Vratha> for the type of system i'm setting up, i can't trust the clients to do anything correct
[09:28:21] <Vratha> when they do something like a local send() to my server to send an IPC message, i need the client to receive it and use the mapped IPC struct like it was its own memory, without worrying that the client is going to get int here and fuck with it
[09:28:26] <e^ipi> well, use doors for RPC, the door server can do the checking & such
[09:29:22] <e^ipi> *shrug*
[09:30:31] *** linma has quit IRC
[09:30:32] <Vratha> blah.. i really need to get solaris installed on my new mini-itx server that i finally got built today just so i can read the man pages on this
[09:30:37] <Vratha> because ic an't find jack crap about doors on the net
[09:30:56] <e^ipi> man door_create & go from there
[09:31:05] <Vratha> yeah; that's what i'll do :)
[09:31:11] <Vratha> i just need to get it installed now
[09:31:20] <Vratha> i guess i'll do that tomorrow when i'm not busy
[09:32:11] *** linma has joined #opensolaris
[09:36:06] *** Dar has joined #opensolaris
[09:36:32] <Vratha> e^ipi: ah, so basically a client connects to a door, and when it does that and does a send(), it's forced to block until the server is finished?
[09:37:03] <Vratha> the man page made it look similar to tcp/ip programming, but just without all that network crap
[09:38:03] <e^ipi> that's what i understand it to be
[09:38:04] <Vratha> ah, that's what it looks like it does using door_call() and door_return()
[09:38:17] <Vratha> sweet!  and it looks like memory is shared between both in a fast-IPC manner?
[09:38:20] <Vratha> GREATNESS
[09:38:24] * Vratha kisses e^ipi
[09:38:31] <Vratha> hope you're a girl
[09:38:31] <e^ipi> not my thing man
[09:38:39] <e^ipi> not last i checked
[09:38:47] <Vratha> :-(
[09:41:19] *** simford|wfh has joined #OpenSolaris
[09:42:59] *** pikapika has quit IRC
[09:45:14] *** DMJC has quit IRC
[09:45:53] *** DMJC has joined #opensolaris
[09:47:41] *** simford has quit IRC
[09:48:04] *** SymmWork has joined #opensolaris
[09:48:18] *** merlin_ has joined #opensolaris
[09:48:55] *** Symmetria has quit IRC
[09:49:00] *** SymmWork is now known as Symmetria
[09:49:58] *** mazon is now known as Mazon
[09:50:10] *** triplah_ has quit IRC
[10:06:19] *** deedaw has joined #opensolaris
[10:07:59] *** Mazon is now known as mazon
[10:09:10] *** mazon is now known as Mazon
[10:10:04] *** alanc-away has quit IRC
[10:10:07] *** alanc_away has quit IRC
[10:10:28] *** alanc_away has joined #opensolaris
[10:10:35] *** alanc-away has joined #opensolaris
[10:13:25] *** setuid has left #opensolaris
[10:16:16] *** karrotx has quit IRC
[10:16:21] <DMJC> can someone walk me through a solaris install?
[10:16:52] <trygvis> just ask if you have any questions
[10:17:07] <DMJC> it won't try to auto partition will it?
[10:17:14] <DMJC> it'll ask me first how I want my drives?
[10:17:21] <tsoome> only thing you have to know - it does not change any disks till last ok
[10:17:26] <DMJC> cool
[10:17:41] <DMJC> it went through networking setup first which I assume is normal :)
[10:17:43] <tsoome> so you can play it trough
[10:20:30] *** Vratha has quit IRC
[10:25:33] <DMJC> hmm
[10:25:40] <DMJC> just a moment seems to hang the system
[10:25:46] <DMJC> after ipv6 setup
[10:29:53] *** raph_ael1 is now known as raph_ael
[10:30:42] *** karrotx has joined #opensolaris
[10:34:43] <raph_ael> hello
[10:34:50] *** DebolazX has joined #opensolaris
[10:35:34] *** Risky_ has joined #opensolaris
[10:36:16] <DMJC> excellent
[10:36:21] <DMJC> installing
[10:36:24] <DMJC> found my hdds
[10:37:46] *** mbirkis has quit IRC
[10:37:47] *** Risky has quit IRC
[10:37:48] *** svoboda_ has quit IRC
[10:37:48] *** gnubee has quit IRC
[10:37:48] *** DebolazY has quit IRC
[10:37:49] *** Kernel86_ has quit IRC
[10:37:49] *** jbalint has quit IRC
[10:37:49] *** noyb has quit IRC
[10:37:49] *** nightswim has quit IRC
[10:37:49] *** Berny_ has quit IRC
[10:37:49] *** lasseoe has quit IRC
[10:37:49] *** Drone has quit IRC
[10:37:49] *** Fetch has quit IRC
[10:37:50] *** master_baiter has quit IRC
[10:37:50] *** postwait has quit IRC
[10:37:50] *** DerJoern has quit IRC
[10:37:50] *** kirma has quit IRC
[10:37:50] *** hali has quit IRC
[10:37:50] *** cap_ has quit IRC
[10:37:50] *** TBCOOL has quit IRC
[10:38:47] *** lasseoe has joined #opensolaris
[10:38:48] *** hali has joined #opensolaris
[10:38:48] *** DerJoern has joined #opensolaris
[10:38:49] *** master_baiter has joined #opensolaris
[10:38:50] *** TBCOOL has joined #opensolaris
[10:38:56] *** Fetch has joined #opensolaris
[10:38:57] *** postwait has joined #opensolaris
[10:38:58] *** gnubee has joined #opensolaris
[10:39:00] *** nightswim has joined #opensolaris
[10:39:02] *** jbalint has joined #opensolaris
[10:39:04] *** mbirkis has joined #opensolaris
[10:39:04] *** cap_ has joined #opensolaris
[10:39:08] *** svoboda_ has joined #opensolaris
[10:41:27] *** Berny has joined #opensolaris
[10:41:32] *** Drone has joined #opensolaris
[10:42:03] *** Kernel86_ has joined #OpenSolaris
[10:43:55] *** noyb has joined #opensolaris
[10:49:48] *** mustang has joined #opensolaris
[10:55:24] *** uebayasi has quit IRC
[11:00:15] *** hachi has joined #opensolaris
[11:06:02] *** damienc has joined #opensolaris
[11:08:43] *** lplatypus has joined #opensolaris
[11:18:54] *** dunc_ has joined #opensolaris
[11:20:14] *** MattMan_ has joined #opensolaris
[11:30:54] *** Mazon is now known as mazon
[11:34:18] *** simford|wfh has quit IRC
[11:40:40] *** dunc__ has joined #opensolaris
[11:43:07] *** dunc_ has quit IRC
[11:48:30] *** whaq_ has quit IRC
[11:50:01] *** whaq has joined #opensolaris
[12:05:45] *** lplatypus has quit IRC
[12:05:59] *** Symmetria has quit IRC
[12:12:59] *** McBofhnyah has joined #opensolaris
[12:13:16] *** mazon is now known as Mazon
[12:14:03] *** pseudoXh4 has joined #opensolaris
[12:14:17] <pseudoXh4> Is it normal for /usr/X*/bin and /usr/sfw/bin not to be in PATH by default? :\
[12:14:38] <Berny> yes
[12:14:51] *** calumb has joined #opensolaris
[12:14:52] <Berny> .oO(who uses X anyway ;-P)
[12:15:01] <Berny> cheerie calum
[12:15:10] <pseudoXh4> How do you put stuff in your PATH so that it's there by default in csh for all users at bootup?
[12:15:33] <cmihai> Who uses csh anyway :)
[12:15:42] <pseudoXh4> Me. :p
[12:15:49] <cmihai> Bloody BSD people
[12:16:01] <andersmo> pseudoXh4: put it in one of the files csh executes on login.
[12:16:17] <cmihai> Drop it in your .profile, .cshrc or whatever
[12:16:32] <Berny> or .login
[12:17:01] <Berny> csh doesn't read .profile iirc, only .cshrc and .login (if login shell)
[12:19:15] *** jcea has left #opensolaris
[12:22:13] *** nyahMcBofh has quit IRC
[12:26:49] *** aliquis has joined #opensolaris
[12:27:14] *** Odin-LAP has quit IRC
[12:28:03] *** nyahMcBofh has joined #opensolaris
[12:36:41] *** McBofhnyah has quit IRC
[12:37:59] *** kirma has joined #opensolaris
[12:38:08] <aliquis> Anyone tried qnext in Solaris? I don't know if it's all or only partly in Java.
[12:39:32] *** mega has joined #opensolaris
[12:40:01] *** Vratha has joined #opensolaris
[12:40:46] <aliquis> I guess the linux version might run
[12:40:47] <aliquis> http://farm1.static.flickr.com/138/376404951_df114b5bc7_o.png  =P
[12:42:00] <aliquis> it's their own im (encrypted), icq, msn, yahoo, aim, irc, file sharing, photo sharing, music sharing, group chat and games in one.
[12:42:36] *** mega_ has joined #opensolaris
[12:42:41] *** mega has quit IRC
[12:45:29] *** jcea has joined #opensolaris
[12:46:44] <quasi> can't even be downloaded
[12:48:06] *** mega_ is now known as mega
[12:48:13] <DMJC> I assume oss is the way to go for sound?
[12:48:30] <DMJC> I also found: http://www.tools.de/opensource/solaris/audio/
[12:48:37] <DMJC> couldn't access the mixer in gnome though
[12:48:37] *** bunker_work has joined #opensolaris
[12:49:34] <aliquis> quasi: http://tucows.com/preview/425428 got a link to the old version
[12:50:05] <aliquis> it autoupdates when you run it anyway so that is ok
[12:50:29] <quasi> aliquis: well, then why haven't you tried it yet? ;)
[12:51:31] <aliquis> DMJC: I guess that depends, I like oss much better and it's mixer is way superior and my soundcard (audigy) is supported completely, but all/many apps is compiled without oss support so the combination kind of suck.
[12:51:53] <aliquis> quasi: Because I never boot Solaris because I'm lame / it's no good desktop os
[12:52:12] <aliquis> also I don't know how to setup the linux zones =P
[12:53:40] *** sniffy has quit IRC
[12:53:52] <DMJC> aliquis, so what do you use?
[12:54:02] <quasi> aliquis: something like http://blogs.sun.com/markp/entry/netbeans_mobility_pack_on_solaris
[12:54:09] *** McBofhnyah has joined #opensolaris
[12:54:12] <timeless> solaris zones aren't hard to setup
[12:54:24] <DMJC> looking at solaris as a drop in replacement for linux
[12:54:26] * timeless has centos and nexenta in zones :)
[12:55:01] <quasi> timeless: brandz zones is slightly harder
[12:55:02] <DMJC> something went bang heh
[12:55:19] <DMJC> told OSS to load drivers kernel freaked out and rebooted
[12:55:52] <timeless> ?
[12:56:08] <timeless> sorry
[12:56:11] <timeless> centos = linux
[12:56:32] <DMJC> are you saying you're booting a linux kernel under solaris?
[12:56:33] <timeless> getting nexenta working was actually more interesting than getting centos working
[12:56:36] <DMJC> or you're just running all the apps
[12:56:37] <timeless> no
[12:56:40] <DMJC> in a centos linux userland
[12:56:42] <timeless> brandz lx
[12:56:57] <timeless> solaris emulates linux kernel
[12:57:18] <timeless> (not an entirely modern kernel, but..)
[12:57:51] <DMJC> how modern?
[12:57:54] <DMJC> 2.6?
[12:57:55] <DMJC> or 2.4
[12:57:57] <timeless> read about brandz lx
[12:58:23] <timeless> [timeless@centos timeless]$ uname -a
[12:58:24] <timeless> Linux centos 2.4.21 BrandZ fake linux i686 i686 i386 GNU/Linux
[12:59:16] <cmihai> Kind of like FreeBSD/OpenBSD/SCO etc. do with Linux compat
[13:00:19] *** _357181_ has quit IRC
[13:00:32] *** gm152 has joined #opensolaris
[13:00:51] *** gm152 has quit IRC
[13:00:52] <cmihai> timeless: how did you get nexenta zones :)
[13:01:11] *** sniffy has joined #opensolaris
[13:01:16] <DMJC> hmm audio is definately causing it to bomb out
[13:01:23] *** gm152 has joined #opensolaris
[13:01:25] <timeless> i think i wrote it up actually
[13:01:27] <DMJC> failed to attach
[13:01:37] <DMJC> definately read but failed to attach
[13:01:50] <cmihai> Really would like running older Solaris versions / other Solaris stuff in zones.. especially stuff like Sol 8/9/10/nexenta etc
[13:01:54] <cmihai> timeless: got a link?
[13:03:03] <timeless> http://www.webwizardry.net/~timeless/osol/
[13:03:30] <cmihai> This should be fun :). Thanks.
[13:04:05] <timeless> get_nexenta.sh is probably the more important piece
[13:04:06] <timeless> there's some doc somewhere abour running snv0 or something
[13:04:38] <cmihai> Oh, I see.
[13:05:04] *** jcsmith_ is now known as jcsmith
[13:07:03] *** kloczek has joined #opensolaris
[13:07:11] *** nyahMcBofh has quit IRC
[13:08:53] <timeless> :(, i can't remember the magic word for solaris10 as a zone in snv
[13:09:54] <cmihai> "please?" :P
[13:11:07] <aliquis> DMJC: Now? Broken XP installation because it's so much work to move away all my data, Solaris doesn't wanna see my sata-disk so I have to move it from ata to other ata to sata from whatever os can use it ;)
[13:11:25] <cmihai> Bah, VIA SATA, what do you expect ;P
[13:11:30] <aliquis> or well, it sees it, but it doesn't give any nice values for cylinders and such
[13:11:30] <aliquis> ;D
[13:11:42] <DMJC> heh
[13:11:45] <aliquis> I will install OS X but I've borrowed my dvd to a friend
[13:11:46] <DMJC> I'm running XP atm
[13:11:55] <DMJC> lol osx
[13:12:07] <aliquis> DMJC: I would prefer to use os x or freebsd
[13:12:07] <cmihai> Vista FTW!
[13:12:08] <DMJC> OSX wouldbe fucking SWEET
[13:12:15] <DMJC> problem is no audio support
[13:12:18] <DMJC> :(
[13:12:23] <cmihai> Even better, SCO OpenServer 6 with KDE! :D
[13:12:28] <DMJC> I would pay $100 for audigy 2 drivers
[13:12:29] <DMJC> for os
[13:12:32] <DMJC> *osx
[13:12:35] <aliquis> timeless: I know solaris zones aren't hard to setup but the linux ones wheren't released when I did it so I don't know how to setup linux ones
[13:12:51] <DMJC> and I'd pay $300 for tv card support
[13:12:59] <DMJC> hell $500 combined for both
[13:13:08] <DMJC> that'd get me onto mac permanently
[13:14:13] <aliquis> DMJC: os x supports my ac97
[13:14:16] <aliquis> but not my audigy of course
[13:14:25] <DMJC> yeah sucks don't it
[13:14:29] <DMJC> onboard audio is crap
[13:14:34] <aliquis> also i have gf4 ti-4200
[13:14:37] <DMJC> no 7.1 output supported in osx
[13:14:41] <aliquis> so i don't know if i shall buy fx5200 for core image
[13:14:44] <DMJC> my 7600gt is supported perfectly
[13:14:47] <DMJC> I ran osx a while ago
[13:14:50] <aliquis> or nv6600 for decent gfx
[13:14:56] <aliquis> or better card and socket 939 + pci-express mobo
[13:15:01] <DMJC> everything except hdtv and audio worked
[13:15:02] <aliquis> or core 2 duo machine with better gfx
[13:15:13] <aliquis> I'd rather not spend so much because I want to get a real mac when they release one which doesn't suck
[13:15:14] <DMJC> hell give it audigy support and it's about equal with windows and linux
[13:15:21] <DMJC> macs suck ass
[13:15:36] <aliquis> DMJC: lol, if you can afford so much to make it work why not just buy a real mac?
[13:15:38] <DMJC> eg none of them have pci-e dual hdtv tuner cards
[13:15:48] <DMJC> unlike my pc
[13:15:56] <DMJC> which is F**CKINg cool
[13:16:13] <aliquis> the mac pro got 4 pci-express x16 i think
[13:16:18] <aliquis> but i don't know if those cards would work
[13:16:18] <aliquis> ;D
[13:16:22] <DMJC> but there's no drivers for saa7162
[13:16:34] <DMJC> there is for windows, and in the work for linux
[13:16:51] <aliquis> and for solaris?
[13:16:52] <aliquis> see
[13:16:57] <aliquis> you can just aswell use os x
[13:16:58] <aliquis> ;D
[13:18:13] <quasi> aliquis: socket 939? it seems a bit late for that now
[13:18:26] <aliquis> plans for solaris 11 with gnu userland and a decent package manager seems nice
[13:18:40] <aliquis> quasi: well 939 mobo + a64 3200+ venice is very cheap
[13:18:53] <aliquis> and atleast I could get a pci-express gfx card and lower power consumtion that way
[13:19:08] <aliquis> socket 764 or whatever it's called and a conroe and ddr2 memory is much more expensive
[13:19:19] <aliquis> err, 478 or whatever
[13:19:25] <quasi> aliquis: and a deadend - with the socket being out of commision, making for only secondhand
[13:19:26] <aliquis> 764 is the one i have for my old athlon64
[13:19:26] <aliquis> ;D
[13:19:34] <quasi> 754
[13:19:41] <aliquis> deadend what?
[13:19:45] <aliquis> like i will ever upgrade the cpu anyway
[13:19:46] <aliquis> =P
[13:19:55] <aliquis> on the same mobo that is
[13:19:57] <quasi> socket 939 has been replaced by AM2
[13:19:58] <aliquis> how many people does that?
[13:20:01] <aliquis> so?
[13:20:10] <aliquis> conroe owns am2
[13:20:14] <aliquis> and am2 doesn't perform better than 939
[13:20:17] <aliquis> and 939 is cheap
[13:20:20] <aliquis> so it's 939 or conroe
[13:20:43] <aliquis> or just fx5200/nv6600 agp
[13:21:09] <quasi> agp gfx tends to be expensive nowadays
[13:21:44] <aliquis> i don't think the difference is so big but it sucks because I don't wanna get an expensive one in case i will upgrade the rest of the machine
[13:21:57] <ofu_> i like my fast and quiet conroe
[13:21:59] <aliquis> but it also suck to upgrade the rest of the machine just to get pci-express gfx ;D
[13:22:08] <aliquis> i would have liked a non-sucking apple
[13:22:11] <aliquis> and cheap aswell
[13:22:12] <aliquis> ;/
[13:22:29] <ofu_> i liked dual-socket conroe, but i wont buy fb-dimm
[13:22:46] <quasi> ofu_: for solaris?
[13:23:11] <ofu_> sorry, no solaris on my desktop
[13:23:18] *** nyahMcBofh has joined #opensolaris
[13:23:30] <quasi> ofu_: so not really relevant ;)
[13:26:13] <aliquis> I would have wanted solaris to be a good enough desktop os but I would have wanted oss support, latest kde, better package management and more responsive ui for that (I still wonder if the ui issue is because I have via chipset or not =P)
[13:26:14] <ofu_> i heard rumors about vmware on solaris/x86, does anybody know anything about it?
[13:26:49] *** ofu_ is now known as ofu
[13:27:15] <quasi> aliquis: solaris != linux ;)
[13:28:39] <aliquis> quasi: which is a good thing ;), but it could be even better! ;D
[13:29:03] <dojtoll> aliquis: If you want the latest KDE. Why dont you just complie it?
[13:29:18] <aliquis> I want to learn it and I want to use zfs and zones and I want to try all the development tools which are free now but it doesn't work that well in combination =P
[13:29:27] *** jcsmith has quit IRC
[13:29:32] <aliquis> dojtoll: I don't know how to / expect issues
[13:29:48] <quasi> aliquis: stick solaris in a vmware
[13:29:58] <aliquis> Also hello, you know who I am and I know who you are but you might not be aware of it yet ;D
[13:30:44] <aliquis> dojtoll: If it was so easy to compile I guess there would already be a port ;)
[13:31:14] *** dlg has left #opensolaris
[13:33:38] *** gm152 has quit IRC
[13:34:04] *** DMJC has quit IRC
[13:34:45] *** DMJC has joined #opensolaris
[13:35:09] <aliquis> quasi: I don't wanna run stuff in multiple virtual machines, I would rather use OS X as desktop and Solaris on server and run the apps thru X11 forwarding instead in that case, but I only got one decent machine atm and electricity is expensive so I'm cheap ;D
[13:36:24] *** McBofhnyah has quit IRC
[13:36:29] <quasi> aliquis: and a pony too
[13:38:17] <aliquis> ponys are expensive? true, but a man must have his pleasure doesn't he?
[13:38:32] *** nostoi has joined #opensolaris
[13:39:07] *** pseudoXh4 has quit IRC
[13:40:14] <timeless> aliquis: i tried getting debian into a zone and had problems
[13:40:23] <timeless> but then it turned out that all i needed was apt, and not linux
[13:40:28] <timeless> so i switched to nexenta
[13:40:44] <quasi> aliquis: at the rate you were asking for impossible things, throwing in a pony shouldn't make much difference ;)
[13:40:49] <timeless> the other requirement i had was for a linux, and centos comes as some sort of shrinkwrapped zone
[13:40:50] <cmihai> Hm... anyone got any experience with HITACHI thunder 9500V Fiber Channel SAN?
[13:40:55] <cmihai> Trying to connect to it serial...
[13:41:01] <dojtoll> aliquis: I cannot send a priv.msg to you but. No I'm running Solaris on a SPARC.
[13:41:04] <cmihai> 96008N1 fails... so I'm out of ideas :P
[13:41:13] <cmihai> Can't find specs heh
[13:41:35] <dojtoll> aliquis: Still got some x86-boxes.
[13:41:48] <timeless> sn-1!
[13:42:28] <timeless> yeah, someday i need to try sn-1
[13:44:34] <cmihai> Well, docs are worthless.
[13:44:45] <cmihai> 9600 8 N 1, and I'll be damned if it works :(
[13:45:04] <timeless> anyway, the sn-1 brandz isn't really that useful
[13:45:09] <timeless> it's just amusing for end users
[13:47:35] <cmihai> Ugh, I think I need to install the "sol" tar to connect to this thing via serials. That's dumb...
[13:49:43] <timeless> so, um, how exactly do i post to fuse-discuss?
[13:49:54] <timeless> i tried subscribing, once by email, and once using the web interface
[13:50:31] <ofu> cmihai: wtf? We will get HDS-Boxes next week...
[13:50:50] <cmihai> Heh
[13:50:57] <ofu> is there a binary for linux/ppc, so that I can install it on a cyclades?
[13:51:18] * cmihai inserts the disk labeled "Propriotary and Confidential" in the cdrom...
[13:51:48] <cmihai> ofu: I have HP-UX, Sol, AIX, WIN and LIN blobs here
[13:52:04] <cmihai> it's Jaba
[13:52:09] <cmihai> it's Jaba
[13:52:12] <cmihai> Erm, Java.
[13:52:26] <ofu> hooray
[13:53:10] <cmihai> Actually, the client is just java 1.4.2 for the platform + some jar files.
[13:53:21] <cmihai> "Array Manager CLI / GUI" bla bla
[13:54:41] <tsoome> not a clean java, some dll/.so as well....
[13:54:45] <tsoome> unfortunately...
[13:55:36] <cmihai> Yeah... some strange binaries there too
[13:56:05] <cmihai> tsoome: so anyway, do I really need to use their client to connect via RS232? Manual says 9600 8N1, but tip/cu won't do the trick.
[13:56:17] <tsoome> no!
[13:56:24] <axisys> my named process is taking 10% cpu.. pretty high load since two days ago.. the dtrace shows this http://rafb.net/p/GDwHJA80.html
[13:56:28] <cmihai> No what?
[13:56:33] <axisys> anyidea where to look next..
[13:56:44] <tsoome> ethernet is the only way to manage it
[13:56:53] <cmihai> Oh... really?
[13:56:56] <tsoome> yes
[13:57:01] <cmihai> Then what are the 2 RS232 ports for?!
[13:57:20] <tsoome> basically you can see vxworks booting;)
[13:57:27] <tsoome> if you are lucky;)
[13:57:29] <cmihai> lol
[13:57:30] *** gallium has joined #opensolaris
[13:57:35] <tsoome> or not so lucky
[13:57:47] <tsoome> it's not for customers
[13:58:08] <ofu> hmmm, do i really want to have a look at those boxes?
[13:58:12] <cmihai> VXWorks is pretty nice and all... but you're saying you don't get any access?
[13:58:15] *** nyahMcBofh has quit IRC
[13:58:28] *** nyahMcBofh has joined #opensolaris
[13:58:33] <cmihai> Fick... I need to setup the IP and stuff so I can manage via LAN first
[13:58:54] <tsoome> there are preconfigured IP addresses
[13:58:57] <ofu> arp -s?
[13:59:06] <tsoome> and in fact there are 2 sets od ethernet ports
[13:59:12] <cmihai> Yep.
[13:59:25] <cmihai> Any idea what the preconfigured IP is? I'd really hate to scan for it now
[13:59:26] <tsoome> closed ones from 10.0.0
[13:59:59] <cmihai> Crap...
[14:00:02] <tsoome> 10.0.0.7 and 8 if my memory serves
[14:00:08] <cmihai> Guess I'd better go get a switch and laptop
[14:00:20] <tsoome> and "customer" ports 192.168.0.7 and 8
[14:00:44] <tsoome> I may be wrong as well:)
[14:01:31] *** whaq_ has joined #opensolaris
[14:01:32] <tsoome> you did just ordered the array and not the installation?
[14:01:39] *** whaq has quit IRC
[14:02:13] <tsoome> usually they sell it only with initial installation.....
[14:04:15] *** Vratha has quit IRC
[14:04:31] <cmihai> tsoome: yupo
[14:05:05] *** yarihm has joined #OpenSolaris
[14:06:01] *** Vratha has joined #opensolaris
[14:11:02] *** mega has quit IRC
[14:11:47] *** mega has joined #opensolaris
[14:13:56] *** Odin-LAP has joined #opensolaris
[14:16:04] *** gallium has quit IRC
[14:16:46] *** bank__ has joined #opensolaris
[14:17:44] <bank__> hi
[14:18:58] <cmihai> tsoome: well, it's not 10.0.0.7,8 or 192.168.0.7,8.. so what now? Scan all IP's? try to arping?
[14:26:22] *** uebayasi has joined #opensolaris
[14:29:32] <jteo> hello *.
[14:32:26] <PerterB> shouldn't that be ".*" ;)
[14:32:53] <bank__> hello -R
[14:33:05] *** Vratha has quit IRC
[14:33:28] <tsoome> umm wait a bit
[14:34:46] *** McBofhnyah has joined #opensolaris
[14:34:49] *** Vratha has joined #opensolaris
[14:34:54] *** nyahMcBofh has quit IRC
[14:35:25] <tsoome> 192.168.0.16 and 192.168.0.17 for open ports and 10.0.0.16 and 10.0.0.17 for closed
[14:35:30] *** calumb is now known as calLNCH
[14:48:16] *** uebayasi has quit IRC
[14:50:28] *** Vratha has quit IRC
[14:56:12] <aliquis> "You can buy an inexpensive USB flash drive and use up to 4GB of cheap RAM to boost your system performance."
[14:56:16] <aliquis> I never understand that vista argument
[14:56:21] <aliquis> since then are usb memory drives fast?
[14:57:32] <tsoome> ;)
[14:57:57] <tsoome> since you install vista, after that, every other thing seems vast;)
[14:58:00] <tsoome> fast
[14:58:20] <bank__> andersmo
[14:58:49] <PerterB> I think the idea is to kind of use it as an L2 cache for the disk, which isn't so dumb: http://en.wikipedia.org/wiki/Readyboost
[14:59:15] <jteo> wouldn't it be cheaper to buy more RAM?
[14:59:24] <PerterB> I would imagine so, yes
[14:59:28] *** mbirkis has quit IRC
[14:59:31] <aliquis> PerterB: But are they really faster than a hdd?
[14:59:36] <trygvis> ram is volatile, that's the whole point
[14:59:40] <andersmo> aliquis: access time.
[14:59:49] <PerterB> aliquis: for random reads only - read the web page
[14:59:56] <aliquis> yeah, I've figured their access time is a little lower
[14:59:58] <andersmo> (which is what hard drives suck at)
[15:00:05] <aliquis> but I don't know how fast that means they are
[15:00:05] <aliquis> =P
[15:00:09] <bank__> hi andersmo
[15:00:13] <andersmo> bank__: ?
[15:00:18] <andersmo> bank__: hi? =)
[15:00:20] <bank__> Are you still working on postgre?
[15:00:41] <bank__> I remember you explain me on postgreSQL development last year.
[15:00:43] <axisys> how do I burn a flash archive in a cdrom?
[15:00:48] <axisys> dd should do it?
[15:00:56] <andersmo> Nope, I'm only a postgresql user "by proxy" (using applications that use postgresql =)
[15:01:11] <andersmo> bank__: just ask the question anyway, maybe someone else here knows. =)
[15:01:22] <bank__> no, I have no specific question.
[15:01:32] * andersmo does, however, hack on http://db.apache.org/derby =)
[15:01:42] <bank__> oh yes. you are derby
[15:01:49] <bank__> sorry, I remember wrong.
[15:02:02] <aliquis> The device should have an access time of 1ms or less.
[15:02:03] <aliquis> The device must be capable of 2.5 MB/s read speeds for 4 kB random reads spread uniformly across the entire device and 1.75 MB/s write speeds for 512 kB random writes spread uniformly across the device.
[15:02:18] <aliquis> ok, decent access time compared to hdd but read and write speeds? =P
[15:03:34] <LeftWing> The future alleged use of other "networked Windows Vista PC"'s RAM seems somewhat dubious...
[15:03:55] <LeftWing> Imagine the realm of new remote vulnerabilities that would create in Windows.
[15:04:35] <aliquis> LeftWing: I guess over gigabit/10 gigabit ethernet that remote ram is faster than disks aswell
[15:04:56] <LeftWing> I can't see them pulling it off without fumbling it security-wise though.
[15:05:15] <bank__> andersmo: are there any solution for data warehousing with derby?
[15:05:25] <andersmo> aliquis: could be useful for paging-like work, yes. Just get the first few blocks from the device and get the application going while the disk drive's head is still busy repositioning itself. =)
[15:06:20] *** nachox has joined #opensolaris
[15:06:33] <andersmo> bank__: I'm not aware of anything using Derby and targeting data warehousing and olap workloads. Kinda isn't Derby's niche. =)
[15:06:58] <bank__> ok :-)
[15:07:58] <trygvis> you're also forgetting that this is for booting the machine and the disks take quite a while to start up
[15:09:02] <razrX> hmm, just luupgraded snv_55b to 56 and was nicely presented with following msg during boot: elfexec: [ID 991571 kern.warning] WARNING: elfcore: core dump failed for process 280; address space is changing
[15:09:26] <razrX> everything appears to be running though
[15:09:51] <razrX> references to this warning can be found in elf.c
[15:10:08] <razrX> http://cvs.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/uts/common/exec/elf/elf.c (lines 1573 and upwards)
[15:10:21] <razrX> don't know what it means
[15:30:09] *** mrdeviant has joined #opensolaris
[15:42:06] *** nostoi has quit IRC
[15:44:13] *** bzcrib has joined #opensolaris
[15:45:35] *** laca has joined #opensolaris
[15:50:30] *** bzcrib has quit IRC
[15:50:53] *** bzcrib has joined #opensolaris
[15:51:29] *** calLNCH has quit IRC
[15:53:52] *** bzcrib has quit IRC
[15:54:06] *** bzcrib has joined #opensolaris
[15:58:17] <timeless> libc.so.1`_waitid+0x15(428, 80444e4, 0)
[15:58:18] <timeless> libc.so.1`waitpid+0x97(428, 80444e4, 0)
[15:58:49] <timeless> does that mean my process is waiting for pid 428?
[15:59:35] <quasi> pid_t waitpid(pid_t pid, int *stat_loc, int options);
[16:03:07] *** RaD|Tz has joined #opensolaris
[16:08:36] *** Mazon is now known as mazon
[16:11:55] <jbk> so yes
[16:14:54] *** jsubl2 has joined #opensolaris
[16:16:02] *** bzcrib has quit IRC
[16:18:20] *** hell has quit IRC
[16:18:42] *** hell` has joined #opensolaris
[16:20:14] *** gongoputch has quit IRC
[16:20:17] *** jsubl2 has quit IRC
[16:20:23] *** McBofhnyah has quit IRC
[16:23:03] *** bunker_work has quit IRC
[16:23:50] *** jsubl2 has joined #opensolaris
[16:24:07] *** bank__ has left #opensolaris
[16:31:30] <timeless> quasi: i ask because there was no pid 428 :)
[16:34:03] *** kAv_ has quit IRC
[16:35:30] <PerterB> might be a long wait then :)
[16:36:20] <jteo> bah.
[16:43:01] *** mrdeviant has quit IRC
[16:44:19] *** Dar is now known as Dar_HOME
[16:46:37] *** Cyrille has quit IRC
[16:46:49] <axisys> i seem to cannot install jumpstart image of sol 8 cdrom on zfs.
[16:46:54] <axisys> i get this error http://rafb.net/p/HfjpoE33.html
[16:47:05] <axisys> anyone know of a work around?
[16:47:43] <quasi> don't put it on zfs ;)
[16:47:47] *** boyd has quit IRC
[16:48:07] <tsoome> export it then start script
[16:48:18] <axisys> tsoome: how?
[16:48:29] <tsoome> man zfs
[16:48:36] <delewis> no, just comment out the 4 or 5 lines in the script that do that test.
[16:48:39] <axisys> quasi: i have not much space.. i put most of the disk as zfs
[16:48:50] <quasi> zfs share
[16:48:54] <delewis> tsoome, that's kill a bird with a boulder.
[16:49:01] <delewis> quasi, too.
[16:49:03] <tsoome> :)
[16:49:42] <axisys> delewis: i cant.. it is cdrom iso.. i guess i have to copy the whole cd first
[16:49:51] <quasi> delewis: why?
[16:49:53] <delewis> yes, copy the CD
[16:49:57] <delewis> then edit the script.
[16:50:10] <axisys> quasi: zfs share ?
[16:50:21] <quasi> axisys: 16:48 < tsoome> man zfs
[16:50:50] <axisys> quasi: i know that .. but how would zfs share help to setup jumpstart?
[16:51:12] <axisys> delewis: let me do the copy
[16:51:24] <tsoome> if fs is already shared, script will not try to share it
[16:51:37] <quasi> axisys: you'll find out eventually
[16:51:48] <tsoome> but true, looking inside the script is probably a good idea
[16:51:57] <delewis> "probably"? it *is*.
[16:52:10] <delewis> there's just 4 or 5 lines that need to be commented out.
[16:52:27] <delewis> I've done Solaris 8 and 9 Jumpstarts this way.
[16:52:32] *** bzcrib has joined #opensolaris
[16:52:55] <quasi> delewis: remember that it is axisys - you'll be needing to give him a patch and the step by step guide to patch ;)
[16:53:03] *** boyd has joined #opensolaris
[16:53:12] *** mrdeviant has joined #opensolaris
[16:53:19] <jteo> help here goes end to end, all the way!
[16:53:23] <axisys> quasi: ha ha.. :P
[16:53:35] <delewis> grep for the string, and comment out the conditional that encapsulates it
[16:53:38] <delewis> not too hard :-)
[16:54:13] <tsoome> expectation is a mother of all fuckups.....
[16:54:29] <delewis> tsoome, not my fuckup :-)
[16:54:36] <tsoome> ;)
[16:54:49] <tsoome> so they say all.....
[16:56:18] <jbk> hmm.. i don't supposed anyone knows offhand which kjp for solaris 10 provides large pages for the kernel?
[16:56:35] <tsoome> kjp?
[16:56:42] <jbk> kernel jumbo patch
[16:56:53] <jbk> the patch that shows up in uname -a
[16:56:57] <delewis> the kernel cannot use large page sizes, AFAIK.
[16:57:10] <delewis> standard 8KB-size pages on SPARC.
[16:57:11] <jbk> as solaris solaris 10 update 2, supposedly it can
[16:57:18] <delewis> interesting.
[16:57:27] <jbk> that's why i was curious which specific patch actually did it
[16:57:45] <tsoome> then check what kjp is provided with u2:)
[16:58:01] *** Cyrille has joined #opensolaris
[16:58:23] <jbk> well there's a decent possibility it was integrated sometime between u1 & u2, so that the kjp w/ u2 might not be the minimum revision needed
[16:58:32] <delewis> it wouldn't make sense to use large page sizes for the kernel, though, given a somewhat significant majority of the hardware out there has limited large page size functionality (AMD64 and UltraSPARC-III)
[16:59:03] <delewis> I should hope they're selecting which hardware to use large page sizes on with care :-)
[16:59:08] <jbk> Support for mapping the kernel heap with large pages was added in S10 update 2, but by default is only enabled for systems with 1GB of memory or more.
[16:59:24] <jbk> (been browsing http://www.solarisinternals.com/wiki/index.php/Multiple_Page_Size_Support)
[17:00:17] *** mazon is now known as Mazon
[17:00:31] *** tsoome has quit IRC
[17:01:45] <delewis> hmm, they don't say anything about AMD64 or UltraSPARC-III, which have very limited large page size functionality.
[17:02:03] <delewis> the do point out that the UltraSPARC-II has a fully associate TLB, which makes it a lot more functional for large pages.
[17:02:11] <delewis> associative, rather.
[17:02:32] <mrdeviant> i believe it will use large pages on us3 and higher
[17:03:03] <delewis> mrdeviant, US2 is better for large pages than <= 900MHz (non-CU) UltraSPARC-IIIs.
[17:03:29] <delewis> you'll get very, very little benefit from large pages on those UltraSPARC-IIIs
[17:04:16] <mrdeviant> i wasn't debating the benefits of using it
[17:05:41] <jbk> what about IV or IV+
[17:06:10] <delewis> jbk, IV or IV+ have a least a two-way associative 512-entry TLB as the latter UltraSPARC-IIIs do.
[17:08:37] <quasi> IV+ is pretty nice
[17:08:49] <delewis> and AMD64 has even more restrictive large page support than old UltraSPARC-III's do.
[17:08:51] <delewis> :-(
[17:08:56] *** boyd_ has joined #opensolaris
[17:09:05] <delewis> at least current AMD64s do.
[17:09:48] <jbk> i guess i can see if i can glean it from the bugid list in the 118833 readme
[17:11:30] <axisys> delewis: installing jumpstart server .. modified setup_install_server is now copying image.. thnx
[17:11:39] <delewis> axisys, sure :-)
[17:11:55] <delewis> the same trick works for the Solaris 9 Jumpstart script, too.
[17:12:24] <axisys> delewis: cool.. by the way dont pay attention to quasi.. he is out there to get me ;-)
[17:13:33] <quasi> axisys: your comment the other day about "help me, but don't tell me to read the man page" put you on the target list
[17:14:08] <axisys> quasi: i c.. lession learned.. :P
[17:14:57] <quasi> yeah, I learned you're a lazy bum who wants to be spoon fed ;)
[17:15:18] * delewis hands quasi a shovel
[17:15:44] <quasi> delewis: you don't think the hole is deep enough yet?
[17:15:45] <axisys> quasi: :P
[17:15:52] <delewis> quasi, :-)
[17:17:27] <richlowe> mornin' all.
[17:18:40] *** Theoden-Nexenta has joined #opensolaris
[17:20:40] *** Tpenta has quit IRC
[17:21:24] *** boyd has quit IRC
[17:23:07] *** hile_ has joined #opensolaris
[17:29:58] <PerterB> yay... finally finished what was meant to be a simple backup setup review for a client and turned into a big ugly detour through Oracle Recovery Manager
[17:32:07] *** Odin-LAP has joined #opensolaris
[17:32:18] <quasi> PerterB: I hope they pay you extra for that ;)
[17:33:21] *** MattMan_ has quit IRC
[17:34:30] <PerterB> I get paid a daily rate... my employers lose out because they quoted fixed price for this (and didn't tell me about all the RMAN stuff when they asked me how long it would take me)
[17:34:41] <quasi> backups should be enough, why bother about recovery? ;)
[17:34:58] * PerterB laughs
[17:35:15] <PerterB> sadly that attitude is way too common for real
[17:36:01] <quasi> $old jobs tsm people seemed to think that backups were the primary product and restores were merely a feature
[17:36:48] <quasi> (like starting a restore and getting an average speed that would hit close to 80 years for a single machine)
[17:36:59] <PerterB> I've never actually seen TSM... This was Legato, which has its quirks but can be configured to do a good job (we used ti quite extensively in a previous job)
[17:37:13] *** mega has quit IRC
[17:37:45] <quasi> tsm is easy to abuse - incremental forever sounds nice, till you have to recover a system from 400 different tapes
[17:37:49] <PerterB> but from a conversation I had with a DBA, TSM actually tracks expiry times on a per file basis not per tape, so I can guess it has a lot of overhead
[17:39:03] <quasi> another bad habit is building file lists in memory - that used to trigger the free memory alarm on a system with a couple of hundred thousand files in one dir
[17:41:17] *** b3stbuddy has joined #opensolaris
[17:41:46] <axisys> speaking og backup.. anyone used amanda or bacula?
[17:42:13] <axisys> we are planning to pick one to backup about 200 systems all cross the country
[17:42:33] <b3stbuddy> we were talking about bacula over here
[17:42:50] <b3stbuddy> but to me it seems kinda involved
[17:43:03] <PerterB> I've used amanda in the past for small-scale stuff... not sure how well it would scale to 200 systems
[17:44:35] <axisys> PerterB: how did u like it? some engineers here want to have a gui for the restore .. easy to manage for 200 systems
[17:44:36] <delewis> I guess TSM or NBU is out of the question?
[17:44:49] <jbk> heh we just ran into that annoying 'rman thinks it's smarter than the os' issue when using it w/ nfs
[17:44:57] <delewis> I've seen a single TSM server scale to several thousand clients.
[17:45:12] <delewis> its quite scalable.
[17:45:24] <PerterB> axisys: it was for half a dozen systems of my own at home, so it was fine for that (no gui)
[17:46:07] <PerterB> I always liked Legato's restore GUI and CLI better than NBU, but that's just me (and 200 systems is well within its capability but it wouldn't be cheap)
[17:46:25] <delewis> PerterB, the TSM quite is nice, too.
[17:47:03] <lasseoe> PerterB: looked at Arkeia?
[17:47:13] <PerterB> I always assumed anything with Tivoli in its name would be way too expensive ;)
[17:47:19] *** comay has joined #opensolaris
[17:47:26] <delewis> http://helpdesk.rpi.edu/update.do?artcenterkey=277
[17:47:34] <delewis> well, TSM scheduling is greatly superior to NBU's.
[17:47:38] <jbk> PerterB: or a bunch of perl scripts? :)
[17:47:39] <delewis> which makes it far more scalable.
[17:47:57] <PerterB> lasseoe: sorry, never heard of it
[17:48:22] <delewis> the TSM server I administered backed up every organizational server and workstation (approximately 1,500) and had backup windows of all sorts.
[17:48:42] <lasseoe> PerterB: it loosk decent enough for small setups, and it's affordable
[17:48:51] <lasseoe> looks even
[17:48:58] <PerterB> nice looking gui (tsm)... does it do RMAN integration and all those good things?
[17:49:03] <delewis> given this was a hospital, *everything* has to be backed up and duplicated, but TSM does a complete "disaster recovery" strategy for you.
[17:49:14] <delewis> PerterB, yeah.
[17:49:51] <delewis> meaning it tells you when you need to put and fetch tapes in the "offsite vault"
[17:50:07] <delewis> aside from the onsite one
[17:50:19] *** richlowe has quit IRC
[17:50:22] <PerterB> sounds good... getting operators to follow a decent tape policy with other backup solutions can be a pain
[17:50:24] *** jsubl2 has quit IRC
[17:50:44] <delewis> PerterB, a TSM operator (for the most part) is a tape monkey.
[17:51:07] <delewis> "Fetch tape foo from vault 0"
[17:51:13] <delewis> "5 tapes ready to be transferred to vault"
[17:51:16] <PerterB> gotcha
[17:51:17] <delewis> that sort of thing.
[17:51:44] <hile_> just hope that the TSM admins have enough sense to actually conduct a DR test
[17:51:52] <delewis> indeed.
[17:51:56] <hile_> *cough
[17:52:02] <delewis> fortunately, when IBM came and setup the TSM prior to my arrival they did all that.
[17:52:15] <jbk> sounds like an interesting project to write scalable open-source backup software :)
[17:52:28] <delewis> we had a nice closet at multiple locations that contained backup images of all the servers along with complete TSM DR plans, etc.
[17:52:29] <PerterB> jbk: that was the intention behind Amanda :)
[17:52:41] <jbk> it sounds like it's not there (I have not looked at it)
[17:52:50] <jbk> my problem is more ideas than time :)
[17:52:57] <delewis> can Amanda at least span backups across tapes, now?
[17:53:00] *** b3stbuddy has left #opensolaris
[17:53:28] <PerterB> not unless they changed their tape format... (it used to be a label then a bunch of ufsdumps)
[17:53:34] <delewis> that's what I thought.
[17:53:41] <delewis> for some reason I thought they were coming up with their own format or something.
[17:54:30] <PerterB> maybe... it's been 7 or more years since I looked at it
[17:54:47] <delewis> I looked at it a little over a year ago
[17:54:48] *** dunc__ has quit IRC
[17:54:51] <delewis> and it still could not span tapes.
[17:55:40] <delewis> with TSM, you have ways to span and localize backup sets to a fixed number of tapes.
[17:55:42] <delewis> its quite flexible :-)
[17:55:58] <delewis> "Do we want to waste tape space or do we want to minimize restore time?"
[17:56:03] <PerterB> hmmph, why can't vendors like Arkeia put actual prices on their website
[17:56:39] <hali> PerterB: i got some arkia backup quotes a few months back
[17:56:46] *** coffman has joined #opensolaris
[17:56:51] <hali> PerterB: from our standpoint it made more sense to buy veritas netbackup linux starter edition
[17:57:03] <PerterB> so not all that cheap then?
[17:57:05] <hali> (which works on all platforms)
[17:57:10] *** dclarke has quit IRC
[17:57:18] *** bzcrib has quit IRC
[17:57:19] <hali> PerterB: 5 clients, server and library was in the $7k range for us
[17:57:26] <delewis> oh, yeah.
[17:57:28] <lasseoe> damn, it's gotten expensive.
[17:57:31] <delewis> NBU would fit that nicely.
[17:57:33] <hile_> ugh netbackup
[17:57:36] <delewis> TSM would be overkill.
[17:57:50] <hali> veritas was $7k, but early maintenance was more expensive... but we still went with netbackup
[17:57:50] <delewis> NBU scheduling really, really sucks.
[17:57:53] <hali> im quite fond of netbackup
[17:57:57] <PerterB> line
[17:57:57] <delewis> it gets exponentially more complicated as you scale up :-)
[17:58:00] <PerterB> oops :)
[17:58:05] <delewis> hali, its only practical for small shops.
[17:58:17] <hali> delewis: we have several installs so i don't know
[17:58:17] <delewis> TSM scheduling is a gem.
[17:58:26] <hali> we've got 2-3 5-10 machine setups and one with perhaps 150 nodes
[17:58:32] <jbk> heh.. veritas sold it to mgmt here
[17:58:42] *** Theoden-Nexenta has left #opensolaris
[17:58:47] <jbk> might explain some stuff..
[17:59:00] <delewis> NBU also uses gtar as a backup format.
[17:59:05] <delewis> that should tell you something.
[17:59:12] <hali> well, thats a "feature"
[17:59:19] <delewis> right.
[17:59:22] <PerterB> my last place had NBU covering about 450 solaris and 1000+ wintel servers... luckily there was a dedicated team, I couldn't make head or tail of the scheduling
[17:59:36] <hali> you can encrypt and use their propriatery backup format as well
[17:59:37] <delewis> PerterB, oh, god.
[17:59:40] <delewis> I can't even imagine.
[18:00:04] <hali> but sure, TSM is quite nice
[18:00:11] <hali> as long as noone mentions networker :P
[18:00:21] <hali> or dataprotector
[18:00:21] <delewis> FedEx uses Networker.
[18:00:25] <hile_> my new gig uses networker
[18:00:29] <jbk> delewis: I know :(
[18:00:32] <_syphilis_> is there a way to see which processes are paging?
[18:00:33] <delewis> and they have very large deployments as you can imagine.
[18:00:43] <hile_> i don't deal with it though, so i don't know the ins and outs
[18:00:43] <delewis> so I tend to assume that Legato scales as well as TSM does
[18:00:46] <hali> i haven't used netbackup since emc bought it... but it used to suck donkeys through a garden hose
[18:00:58] <PerterB> plus, because they skimped on drives, if we wanted to book dedicated restore times (eg backout of last resort in a major software go-live), they could never guarentee when stuff would get backed up or restored
[18:01:02] <hile_> when did EMC buy netbackup?
[18:01:09] <jbk> you mean legato?
[18:01:12] <hile_> i thought they might have bought legato?
[18:01:23] <PerterB> they definitely bought Legato
[18:01:35] <PerterB> NBU is still veritas/symantec, no?
[18:01:43] <delewis> FedEx is in the process of moving to NBU, though :-(
[18:01:46] <delewis> PerterB, yes.
[18:01:57] *** richlowe has joined #opensolaris
[18:01:59] <hali> hm, must be 3 years ago they bought legato
[18:02:55] <hali> didn't CA used to have some backup software as well?
[18:03:20] <jbk> probably
[18:03:30] <jbk> they bought every other software in existance :)
[18:04:38] <hali> sorry, they don't have a backup software... they have "arcserve" :)
[18:05:07] * PerterB shudders
[18:05:34] <coffman> oh dear
[18:05:38] <hali> last time i saw arcserver it was loaded as an nlm :)
[18:05:43] <PerterB> ditto
[18:05:44] <coffman> u scare people of with this..
[18:08:08] <delewis> http://www.open-mag.com/features/Vol_83/arcserve/arcserve.htm
[18:08:35] <delewis> doesn't look *too* bad.
[18:08:58] <delewis> apparently, they're using Ingres for the database.
[18:09:49] <delewis> IBM is supposedly going to use DB2 eventually for TSM.
[18:10:38] <quasi> currently they have flat files (which by the way is not good for 300M sized files)
[18:10:54] <delewis> quasi, who? TSM?
[18:11:12] <Stric> delewis: TSM already uses DB2:ish as backend now
[18:11:19] <quasi> delewis: yeah
[18:11:24] <delewis> Stric, with 5.3, sort of.
[18:11:36] <delewis> quasi, yes, I know. I experienced database corruption on one instance.
[18:11:39] <hali> i thought they already switched to db2?
[18:11:56] <delewis> had to do a full dump to tape, restore, and then audit the database.
[18:12:02] <delewis> took the better part of a weekend.
[18:12:07] <quasi> delewis: max is supposed to be 60M, but some fools I know ran 200M or more
[18:12:09] <hali> i remember some ibm global services seminar like in 2002 i attended where they said "the next version will use db2 and you can run sql straight to the database to get reports and stuff"
[18:12:20] <delewis> of course that's what you get when you don't have rollback enabled in your logs.
[18:12:27] * delewis curses at the IBM GS tech that set it up
[18:12:45] <delewis> quasi, hah.
[18:12:52] *** stevel has joined #opensolaris
[18:12:52] *** ChanServ sets mode: +o stevel
[18:12:52] <delewis> we had *a lot* more than that.
[18:12:53] <delewis> ~ 25GB.
[18:12:54] *** stevel has left #opensolaris
[18:13:01] *** stevel has joined #opensolaris
[18:13:01] *** ChanServ sets mode: +o stevel
[18:14:01] <quasi> hali: maybe they were talking about next major version - they've been on 5.x for quite a while
[18:14:52] <Stric> SELECT (Perform an SQL Query of the IBM Tivoli Storage Manager Database)
[18:14:58] *** Chris_D has joined #opensolaris
[18:15:08] <delewis> Stric, that doesn't imply its DB2.
[18:15:15] <delewis> that implies it implements a subset of SQL.
[18:15:37] <Stric> I've heard our backup admins talk about it being DB2 in the back though
[18:15:48] <delewis> and I'm fairly sure 5.1, 5.1.5, 5.2, and 5.2.2 had SELECT, as well.
[18:15:57] <Chris_D> Good morning, all. I'm having horrible problems with NFS against Solaris 10, and I thought the OpenSolaris community might have some 'top-of-the-head' responses.
[18:16:06] <delewis> but alas I haven't touched TSM in well over a year now.
[18:16:11] <delewis> so my memory may be foggy :-)
[18:16:17] <Chris_D> On Linux clients, great write performance. Virtually non-existent read performance.
[18:16:50] <quasi> delewis: they had a select option on the command line
[18:22:33] *** qdk has quit IRC
[18:22:57] <elektronkind> Chris_D: solaris 10 is the server?
[18:23:33] <elektronkind> Chris_D: describe your setup with a little more detail if you could
[18:24:49] <Chris_D> Solaris 10 is the server. Clients are Linux (Ubuntu 6.2). NFS sharing a ZFS filesystem. On the Sun machine, aggregated 4 GigE ports. All machines are connected to the same switch.
[18:25:26] <elektronkind> and read performance is a problem for the clients you say
[18:25:33] <PerterB> usually Linux client people complain about bad write performance, not read I thought
[18:25:52] <Chris_D> PerterB: Yes, that's *usually* the case.
[18:25:53] <Chris_D> :)
[18:26:13] <Chris_D> Yes. Write performance, sustained over all 6 machines, is very good--iostat peaking at 200M.
[18:26:22] <Chris_D> Read performance, though, is horrible.
[18:26:27] *** tsoome has joined #opensolaris
[18:26:36] <Stric> Chris_D: ubuntu 6.2 doesn't exists.. 6.06 or 6.10 does..
[18:26:49] <Chris_D> Watching iostat, there's no activity for long periods (~30s), then a burst.
[18:26:50] <Stric> Chris_D: have you tried without link aggregation?
[18:27:05] <Chris_D> Stric: Sorry, 6.10.
[18:27:12] <Chris_D> Stric: No, not yet.
[18:27:28] <Stric> have you tried network transfers without nfs involved?
[18:27:51] <Chris_D> Stric: Yes. Without NFS, all appears to be good. I'm going to run some Samba tests this morning.
[18:28:26] <Stric> have you tried those linux clients against anything else but your s10 server?
[18:28:38] <elektronkind> Chris_D: how many files are being accessed on the NFS server? lots? hundreds? thousands?
[18:28:46] <Chris_D> elektronkind: 1, at the moment.
[18:28:55] <Chris_D> This is purely a testing environment.
[18:29:14] <Chris_D> Stric: Yes, we ran the same series of tests against a NetApp.
[18:30:43] <Stric> not sure if I have a 6.10 client anywhere close to a s10 server, but 6.06 clients against s10 (and snv40..55) server works just fine here..
[18:31:13] <Chris_D> Stric: Did you change any of the NFS defaults on client or server?
[18:31:36] <Stric> nope
[18:32:19] <Stric> Chris_D: tried tcp and udp?
[18:32:43] <Chris_D> OK. I just forced the NFS version to 3 in /etc/default/nfs and I'm getting much better performance (in that I'm getting *performance*).
[18:33:06] <Stric> what does 'snoop rpc nfs' say on the server when it "freezes" .. last packets before idle and first packets after traffic starts again
[18:33:51] <Stric> so it used nfsv4..
[18:33:56] <Stric> (which is tcp only)
[18:34:45] <Chris_D> After forcing to v3, I no longer get the dropouts.
[18:35:04] <Stric> so there's some v4 interoperability problem..
[18:36:14] <Chris_D> Yes. I went from 30 minutes to read a 20 Gb file to 108 seconds.
[18:36:17] <Chris_D> Oops.
[18:36:24] <Chris_D> 198 seconds.
[18:36:33] <Chris_D> 108 MB/s
[18:36:58] <_syphilis_> hmm, does the 'de' column in vmstat mean anything anymore?  it seems to be permanently 0
[18:38:44] <Chris_D> These past few minutes, the tests were q&d--I'm going to rerun my benchmarks now that I have saner numbers on 1 machine.
[18:41:28] <elektronkind> you can also force nfsv3 on the clients
[18:41:51] <Chris_D> Tried that before trying to force it on the server--didn't work.
[18:42:10] <elektronkind> nfsvers=3 would be the mount option
[18:43:57] *** damienc has quit IRC
[18:44:47] <Peanut> Does anyone know what an SDC number is and/or how to get one?
[18:46:25] <Chris_D> elektronkind: Yes, that's what I used.
[18:46:41] <Cyrille> As in a Sun Download Center number? Where are you asked for one?
[18:48:02] <Peanut> No, Sun Developer C??? I guess, I've asked for a quote for an U20 and they're asking me that number.
[18:48:34] <Cyrille> sun developer connection maybe?
[18:48:37] <Cyrille> no idea, really.
[18:49:24] *** RaD|Tz has quit IRC
[18:50:19] <gdamore> good morning *
[18:50:32] <Stric> good evening
[18:51:41] *** RaD|Tz has joined #opensolaris
[18:54:19] *** slowhog has joined #opensolaris
[18:54:34] *** qdk has joined #opensolaris
[18:55:21] *** peteh has quit IRC
[18:56:37] *** peteh has joined #opensolaris
[19:00:48] *** jsubl2 has joined #opensolaris
[19:07:30] <axisys> how come this works in bash but not in sh .. ls -al /opt/tools/agents/data/datastream/*/agent/
[19:08:14] *** axxl has joined #opensolaris
[19:08:20] <axisys> i am trying it on sol 8
[19:08:21] *** GmanAFK is now known as gman
[19:08:26] *** gman is now known as Gman
[19:08:37] <richlowe> hey Gman.
[19:08:46] <Gman> hey hey richlowe
[19:10:11] <axisys> hmm.. seems like it works on other sol 8 but not this one system
[19:10:27] *** Mazon is now known as mazon
[19:12:36] *** sommerfeld has joined #opensolaris
[19:12:36] *** ChanServ sets mode: +o sommerfeld
[19:13:08] *** coffman has quit IRC
[19:14:41] *** kAv_ has joined #opensolaris
[19:15:04] *** alanc-away is now known as alanc
[19:17:43] *** tsoome has quit IRC
[19:21:28] *** postwait_ has joined #opensolaris
[19:21:38] <postwait_> picking back up on my SVM troubles
[19:21:55] <postwait_> I'm trying to add a third slice into a mirror set.
[19:22:03] <postwait_> metainit -f d6 1 1 c1t2d0s0
[19:22:11] <postwait_> And I get: c1t2d0s0: overlaps with device in d3
[19:22:21] <postwait_> "d3" is c1t1d0s0
[19:22:29] <postwait_> (an entirely separate physical drive)
[19:22:35] <postwait_> how is this possible?
[19:23:46] <quasi> postwait_: are you absolutely sure your partitions don't overlap?
[19:24:22] <postwait_> How could they?  they on different physical media.
[19:24:28] <postwait_> one is t2 and the other t1
[19:24:44] <postwait_> So.. I have three drives with identical SMI labels and VTOCs.
[19:24:58] <postwait_> And if have the primary metadb on c1t0d0s4
[19:25:07] <postwait_> I can add c1t1d0s4
[19:25:18] <quasi> I was going to say slices rather than partitions
[19:25:18] <postwait_> (as a another metadb)
[19:26:12] *** wesolows has joined #opensolaris
[19:26:18] <postwait_> But then I try to add c1t2d0s4 (as a third metadb) and I get overlaps with device in metadevice state database
[19:26:21] *** ChanServ sets mode: +o wesolows
[19:26:40] <postwait_> If I add c1t2d0s4 first, it works.. then on adding c1t1d0s4 as the third, I get the same error.
[19:27:11] <postwait_> It's like it is confusing the last two drives as the same drive or something.
[19:27:12] *** dunc has quit IRC
[19:28:40] <postwait_> I've been looking at this for days and can't figure it out.
[19:29:40] *** Chris_D has left #opensolaris
[19:33:10] *** hile_ has quit IRC
[19:36:22] *** hile_ has joined #opensolaris
[19:39:32] *** karrotx has quit IRC
[19:39:47] <slowhog> uname -a on a T2000 returns: SunOS new-host 5.11 snv_56 sun4v sparc SUNW,Sun-Fire-T200.
[19:40:02] <slowhog> notice the T200. What's the story? ;-)
[19:40:10] <sommerfeld> oh, that's easy.
[19:40:30] <sommerfeld> last minute product rename
[19:41:00] <sommerfeld> same for V880 and V480
[19:41:32] <slowhog> oh, so this is obvious and apply to all the produce line or it's somehow updated
[19:42:08] <slowhog> I mean, later products has been updated to correct name or stay as T200
[19:42:23] <sommerfeld> left forever as a testament to the indecisiveness of marketing :-)
[19:43:10] <sommerfeld> (V480 was originally going to be 480R; V880 was originally going to be 880, then they got v-happy...)
[19:48:34] *** yarihm has quit IRC
[19:48:37] *** logic_ has quit IRC
[19:51:31] <richlowe> and started using z's for random purposes.
[19:52:32] <hile_> v880 z was rather cool
[19:52:38] <hile_> don't think i'd want one undermy desk, though
[19:53:07] <alanc> z's were Opterons, except when they were SPARCs with Zulu graphics
[19:54:09] <alanc> and of course you all know the version after Solaris 10 will be "Solaris Z" to highlight ZFS, Zones, and BrandZ
[19:54:17] <richlowe> that would be what I was referring to.
[19:54:26] <delewis> oh, no.
[19:54:31] <delewis> like Solaris 11Z?
[19:54:32] <richlowe> alanc: I think you mean DynFS, Containers and <foolish name yet to be thought up> ;)
[19:54:59] * alanc laughs maniacally
[19:55:25] <alanc> though I should probably be careful joking about such things, might give marketing ideas... 8-)
[19:55:26] *** RaD|Tz has quit IRC
[19:56:00] <delewis> pretty soon Solaris releases are going to sound like names of Japanese car imports.
[19:56:11] <delewis> Solaris 15MCXZ
[19:56:12] <alanc> richlowe: I thought they already chose Linux Application Environment or something close to that
[19:56:44] <alanc> all the cool kids add letters to their version names - Oracle 8i, HP-UX 11i, AIX 5L
[19:56:58] <Cyrille> and these are cool kids?
[19:57:09] <alanc> and of course Microsoft, who just dropped the numbers altogether and went pure letters
[19:57:11] <delewis> IRIX kept it simple :-)
[19:57:24] <delewis> and you see where it is now
[19:57:29] <alanc> but SGI has been killing IRIX for years, so it doesn't coutn
[19:57:51] <e^ipi> irix's only been dead for a month and a half
[19:57:53] <e^ipi> show some respect
[19:58:00] <delewis> well
[19:58:11] <delewis> as long as Sun doesn't take tips from IBM's DB2 marketing team, we'll be ok
[19:58:13] <Cyrille> following up on your idea, the next name could be sOlariZ (eXistenZ style ;-))
[19:58:22] <delewis> where DB2 8.2 == DB2 8.1+some maintenance pack.
[19:58:39] <delewis> regressive versioning :-)
[19:59:53] *** bengtf has joined #opensolaris
[20:00:00] *** bengtf_ has joined #opensolaris
[20:00:19] *** bengtf_ has left #opensolaris
[20:00:56] *** logic_ has joined #opensolaris
[20:01:14] <alanc> delewis: you mean how Sun Studio 10 update 1 morphed into Sun Studio 11 to make it sound bigger?
[20:01:42] <delewis> alanc: oh, no. I didn't know that. :-(
[20:01:45] <alanc> e^ipi: yes, but it was a long slow poisoning
[20:02:22] *** McBofhnyah has joined #opensolaris
[20:02:24] <delewis> though, Sun Studio 11 did constitute a new release, as the licensing changed.
[20:02:29] *** McBofhnyah is now known as McBofh
[20:02:29] <alanc> they did the same with Studio 10, but that made more sense - adding AMD64 was bigger than just a "Studio 9, Update 1" sort of thing
[20:02:39] <delewis> DB2 versioning is just pure annoying.
[20:02:48] <alanc> yeah, I think the freebieness of it was a large part of the version inflation
[20:02:49] <delewis> IBM has a nice long table in the DB2 documentation detailing what's what
[20:02:58] <delewis> DB2 7.1.1.1.1.1.1 == maintenance pack 99 reversion 3
[20:03:07] <delewis> (exagerating, but you see the trouble)
[20:03:21] <andersmo> oracle versioning is fun as well.
[20:03:36] <delewis> andersmo, they've at least made their entire product line consistent.
[20:04:06] <delewis> no more "Oracle Developer version Foo", "Oracle Database version Bar", "Oracle Application Server version Baz"
[20:04:18] <andersmo> Like database 10g being version 10.0.1.0 and application server 10g bering version 9.4.2.1 or something. =)
[20:04:29] <andersmo> s/bering/being/
[20:04:30] <delewis> andersmo, yeah, but ignore that :-)
[20:04:38] <delewis> focus on the 10g!
[20:05:17] * cap_ nominates RedHat kernel versioning to the discussion: 2.6.9-42.0.8-ELsmp
[20:05:20] *** nostoi has joined #opensolaris
[20:05:47] <andersmo> delewis: ignore it? But it's exactly what I'm grumbling about. =)
[20:05:58] <andersmo> again, marketing.
[20:06:35] <delewis> I guess marketing departments all over the world think that rapid versioning and version scheme modification signify "activity" in the product.
[20:07:40] <delewis> maybe inactivity in the brains of those marketing drones :-)
[20:08:23] <andersmo> And quite to the contrary: "Do you think we'll ever see a 2.7 or 3.0 Linux kernel? - It's not likely. Linus has essentially said, that a 2.7 kernel will not happen unless someone pries control of the kernel out of his hands. [...]" - http://kerneltrap.org/node/7637
[20:08:57] <andersmo> "Marketing? We don't need no steenkin' major bersion bumps, let alone marketing."
[20:10:40] <alanc> it's been amusing watching the LSB lists lately as the Linux world has discovered their success brings them very much the same demands for ongoing compatibility that Solaris has had for 15 years, and they're finally learning to accept it
[20:11:50] <Gman> lsb is so scary.
[20:12:02] <e^ipi> don't let it scare you
[20:12:05] <e^ipi> nobody pays attention to it
[20:13:00] <Gman> thing is, people do
[20:13:04] <Gman> they just think it's crack too :)
[20:16:21] *** Cyrille has quit IRC
[20:17:46] *** dunc has joined #opensolaris
[20:21:49] *** Kmays has joined #opensolaris
[20:24:23] *** mon has joined #opensolaris
[20:24:28] <mon> hiya
[20:24:46] <Kmays> !
[20:24:47] <mon> im afraid i was tricked with a nvidia (silicon image) sata controller
[20:24:59] <mon> i read that with the latest builds this should work, but it doesn't
[20:25:18] <mon> cant find to much info about controllers so does anyone here know something?
[20:28:43] *** RaD|Tz has joined #opensolaris
[20:30:04] *** deedaw has quit IRC
[20:30:24] *** estibi has joined #opensolaris
[20:34:05] *** Claudemir1074 has joined #opensolaris
[20:35:20] <Claudemir1074> hi guys, anybody already installed opensolaris in a laptop "Acer 4230"
[20:35:41] <e^ipi> that's really vague
[20:35:43] *** pikapika has joined #opensolaris
[20:36:17] <pikapika> hello
[20:36:18] <e^ipi> the collector of parts isn't important... what's the chipset? what ethernet chip? what wifi card? sound card?
[20:36:45] *** dunc_ has joined #opensolaris
[20:38:22] *** dunc has quit IRC
[20:38:51] <Claudemir1074> hmm..I go to look at!!!
[20:39:19] <e^ipi> acer just takes a bunch of stuff & puts it in a box
[20:41:39] <Claudemir1074> tks my friends
[20:43:03] <Kmays> mon <- http://www.sun.com/bigadmin/hcl/data/sol/
[20:43:47] <_syphilis_> the HCL does actually list complete systems, but not very many :)
[20:43:50] <mon> Kmays: right thanks. i forgot to mention i just tried
[20:43:50] <Claudemir1074> i am want to buy one notebook to install opensolaris
[20:44:03] <mon> cant find anything on sata, it's weird
[20:44:52] <Kmays> Its under technology field (Storage: RAID)
[20:44:57] <e^ipi> acer ferarri seems to be used a lot by the developers... pretty decent bet it works ok
[20:45:00] <_syphilis_> what allocates from kmam_va?  according to ::kmastat, 1GB is allocated to kmem_va_4096 .. filesystem cache?
[20:45:12] <Kmays> mon: What OS are you using?
[20:45:13] <Claudemir1074> but, i wanted to buy one powerfull notebook that it opensolaris
[20:45:22] <mon> Kmays: thats what i mean. search for all under that category. no results
[20:45:27] <mon> Kmays: now? linux
[20:45:33] <mon> ubuntu
[20:46:23] <mon> think ill try gnusolaris but that's not really what i want :(
[20:46:48] <gdamore> look at www.tadpolecomputer.com if you want a Solaris notebook. :-)
[20:46:58] <gdamore> (disclaimer: they pay my salary)
[20:47:34] <Kmays> mon <- Belenix 0.5.1 or SXCR b56 is good for testing.
[20:48:18] <Kmays> http://www.sun.com/bigadmin/hcl/data/sol/components/views/external_storage_all_results.page1.html
[20:48:51] <Kmays> A few companies support SATA:RAID under Solaris.
[20:48:52] <mon> it works for you? hmm
[20:48:58] *** whaq__ has joined #opensolaris
[20:49:01] <mon> yeah i was pretty sure of that too :)
[20:49:39] <Claudemir1074> but Acer Ferrari...expensive a lot...hehehe
[20:50:07] <mon> and not nice imo
[20:50:16] <mon> i can recommend a macbook ;)
[20:51:09] <e^ipi> how's the solaris wifi support on the macbook, btw?
[20:51:15] <whaq__> thinkpad x60 (s, tablet..)
[20:52:36] <mon> e^ipi: probably nonexistant. the broadcom thing was/is a pain in the *ss on linux too
[20:52:54] <e^ipi> oh, it uses ndiswrapper?
[20:54:27] *** RaD|Tz has quit IRC
[20:54:51] <Claudemir1074> thanks guys
[20:54:58] <mon> could be that's possible
[20:55:07] <mon> never tried that much, i like osx :)
[20:55:22] *** whaq_ has quit IRC
[20:56:49] <e^ipi> yeah, it's alright
[20:57:49] <postwait_> How well wourld the IP stack in Solaris deal iwht a LOT LOT LOT of /128 routes?
[20:57:55] <postwait_> (IPv6 of course)
[20:57:57] *** deather_ is now known as deather
[20:58:26] <Claudemir1074> the thinkpad x60,,,is good? install opensolaris?
[20:59:30] *** axxl has quit IRC
[20:59:34] <postwait_> will the algorithms/datastructures used in the route table continue to preform well with a million IPv6 host routes?
[21:01:16] * delewis seconds gdamore's suggestion :-)
[21:02:07] *** dunc_ has quit IRC
[21:02:24] *** dunc has joined #opensolaris
[21:02:58] <e^ipi> Claudemir1074: looks like everything's supported
[21:05:12] *** logic__ has joined #opensolaris
[21:05:32] <sommerfeld> postwait_: that's a good question for networking-discuss at opensolaris dot org
[21:05:36] *** Claudemir1074 has quit IRC
[21:05:54] <Kmays> Claudemir1074: Most people have Thinkpads and MacBook Pro laptops (x86-based).
[21:06:15] <delewis> postwait_, given Solaris has had IPv6 support (and I've seen a few shops using IPv6 on Solaris with great success) since Solaris 8, the algorithms and data structures are probably fairly mature.
[21:06:37] <sommerfeld> postwait_: so, there's fundamentally no reason it couldn't be made to work given sufficient memory and a 64-bit kernel
[21:07:04] <sommerfeld> postwait_: all the relevant code is open-source under CDDL
[21:07:46] <delewis> Kmays, "most" :-)
[21:07:56] <sommerfeld> postwait_: I have tried millions of IPsec policy rules and IPsec security associations; haven't tried millions of host routes.
[21:08:40] <mon> dont want to be smart but why not just buy an expensive router?..
[21:09:02] <richlowe> sommerfeld: I seem to recall you having fixed an "ipv6 route hashing doesn't" kind of bug, too.
[21:09:07] <richlowe> did that hit a 10 update?
[21:09:15] *** Darwin has quit IRC
[21:09:23] <richlowe> (or was that SAs, not v6 routes?)
[21:09:57] *** deedaw has joined #opensolaris
[21:10:11] <sommerfeld> it was SA's.  parens in the wrong place in the hash function leading to truly psychotic exponential hash distribution across buckets.
[21:10:21] <delewis> hah.
[21:10:59] <sommerfeld> mon: it's been a while since I followed it but I didn't think purpose-built routers were built for millions of v6 routes yet.
[21:11:26] <mon> sommerfeld: oh could be, wouldn't know :)
[21:11:43] *** Darwin has joined #opensolaris
[21:11:53] <sommerfeld> mon: they were more typically built for high PPS forwarding rates with modest-sized routing tables
[21:12:20] <sommerfeld> using fast and thus expensive memory
[21:12:26] <postwait_> I think this app is for very low PPS with enormous routes.
[21:13:15] <_syphilis_> it's not really decided how the ipv6 routing table is going to scale yet in any case
[21:13:49] <_syphilis_> a lot of people would like to see something better than the ipv4 solution, to prevent the problem with millions of routes
[21:13:49] <sommerfeld> postwait_: sounds like some voip control plane/mgmt apps i've heard about
[21:14:49] <postwait_> Here we're going to manage 50M routes.
[21:15:02] <postwait_> however, it doesn't need to be on one machine.. it could be on a handful.
[21:17:08] <mon> ive got this feeling since it's a stable bsd it'd do pretty good
[21:17:39] <postwait_> freebsd does harribly on this front as it uses particia trees to hold route tables.
[21:18:20] <mon> well it's more the sun thing :)
[21:18:29] <mon> i really like linux, but it's all a bit chaotic
[21:18:34] *** logic__ has quit IRC
[21:18:35] <e^ipi> a bit?
[21:18:36] <mon> choices are good but this
[21:18:47] <mon> distro's versions, pkg management systems, kernels
[21:19:01] <mon> it needs a bit of dictatorship sometimes i think :)
[21:19:15] <mon> that's why i chose sun, i expect it to be good like osx
[21:19:45] <e^ipi> the kernel's a lot better than OSX
[21:20:10] *** Darwin has quit IRC
[21:20:11] <mon> wouldnt know, dont touch it that much :)
[21:20:21] <mon> with osx it's more the software experience, like integration
[21:20:24] <mon> really just works
[21:20:25] <e^ipi> XNU is very chatty
[21:20:28] <jbk> anyone know offhand of anything that will play .asx in solaris?
[21:20:32] <e^ipi> it spends a lot of time in IPC
[21:20:43] <e^ipi> so it's slow ...
[21:20:53] <e^ipi> also the memory allocator is pretty crappy
[21:21:20] <mon> osx seriously feels fast to me
[21:21:21] *** logic_ has quit IRC
[21:21:26] *** Darwin has joined #opensolaris
[21:21:32] <mon> it's a matter of opinion and taste i guess
[21:21:43] <e^ipi> not really, it's a matter of benchmarks
[21:22:08] <mon> got some?
[21:22:21] <sommerfeld> mon: except when the psychadelic beachball starts spinning
[21:22:22] <mon> still ill take my feelings over i chart though :) but i do like charts
[21:22:41] <mon> that's only with vlc which happens to crash sometimes when i play with the lid ;)
[21:22:48] <mon> dont see him around much
[21:23:08] <delewis> mon, what you're seeing is the performance of the interface.
[21:23:16] <delewis> not the actual underlying operating system.
[21:23:42] <delewis> OS X does things like caching bitmaps of each window to make window movement and other things appear "fast"
[21:23:54] <mon> excuse my ignorance but i dont mind what happens there. i like it to just work (tm)
[21:24:38] <mon> i totally love Java. this guy at work also explained how the memory allocation was actually better then c++'s and all that
[21:24:50] * delewis holds head in shame
[21:24:59] <mon> that's nice and all, but still the gui-stuff is... slow
[21:25:15] <mon> maybe just my perception of things :)
[21:25:21] <McBofh> mon: java gui? use at least java5, preferably java6 ... it'll change your mind
[21:26:11] <mon> i do :)
[21:26:22] <mon> made an app in 5 and recently used it with 6
[21:26:50] <mon>  especially with this guy with "only" 256Mb ram. had to wait 20 secs (timed)
[21:26:59] <McBofh> ugh ...
[21:27:05] <McBofh> did you tune the startup args for the jvm?
[21:27:14] <mon> hm nope no tuning
[21:27:17] <mon> what'd i miss?
[21:27:29] <McBofh> there's stuff about how often gc runs, for example
[21:27:35] <McBofh> and how much ram to grab
[21:27:38] <postwait_> I'm still totally screwed on my metadb issue :-(
[21:27:57] <delewis> and you definitely want the Java VM go grab less memory than the amount of physical memory you have in order to prevent page faults.
[21:27:59] <postwait_> My install keeps things parts of disk 3 overlap with parts of disk 2... (which makes no sense!)
[21:28:13] <delewis> which was more than likely what was happening to your friend.
[21:28:15] <mon> i saw you could manually do a gb collect, but i dont know that much about it
[21:28:28] <mon> in this case it was XP too i guess. should mention that
[21:28:35] <mon> eh make that win2k
[21:29:04] * delewis wonders when the consumer market will move beyond 256MB or 512MB of memory.
[21:29:21] <richlowe> when they stop raking in 'upgrade' money with purchases?
[21:29:26] <mon> this was at work. it was my "coach" :)
[21:29:30] <mon> i agree though
[21:29:31] <delewis> richlowe, :-)
[21:29:39] <mon> he got 1G this week finally after a month
[21:29:45] *** Darwin has quit IRC
[21:29:50] <e^ipi> OSX is a nice windowing system bolted to a terrible OS
[21:30:04] <postwait_> the OS is pretty nice.
[21:30:04] <delewis> well, applications are already bloated, but I would rather have bloat and lots of memory than page faults. :-)
[21:30:07] <postwait_> Mach is cute.
[21:30:14] <delewis> postwait_, it is cute.
[21:30:21] <delewis> but still terrible practically-speaking.
[21:30:27] <delewis> paging is slow as fuck.
[21:30:29] <postwait_> no so much for desktop apps.
[21:30:47] <postwait_> the mach-o architecture for binaries, dylibs, bundles and frameworks
[21:30:50] <delewis> uh, its a micro-kernel, and therefore, will suffer from the disadvantages of micro-kernels.
[21:30:53] <mon> delewis: agreed. osx is huuuge too but it does make stuff work
[21:30:58] <postwait_> makes it really efficient to share application level resources.
[21:31:05] <delewis> outrageous overhead, difficult to debug, etc.
[21:31:09] <sommerfeld> part of the problem is that pages haven't grown as fast as memory has.
[21:31:10] <postwait_> applications cross-exposing services to other apps is easy
[21:31:16] <mon> have any experience with cocoa by any chance?
[21:31:18] <delewis> sommerfeld, true.
[21:31:19] <postwait_> And that's "mach's fault"... so I like Mach for that.
[21:31:26] <postwait_> mon: I do
[21:31:30] <mon> seems like a really good programming platform to me
[21:31:39] <mon> /easy
[21:31:44] <postwait_> The user-space paging facilities do kinda blow.
[21:31:46] <delewis> postwait_, Mach has nothing to do with anything on the application-level.
[21:31:58] <postwait_> delewis, it does a bit
[21:32:04] <postwait_> Much of that is accomplised using mach ports.
[21:32:11] <postwait_> which are facilitated by the mach kernel.
[21:32:20] <delewis> oh, god.
[21:32:32] <postwait_> It's very non unix kernel ish
[21:32:36] *** toploup33 has joined #opensolaris
[21:32:43] <toploup33> hello everyone
[21:32:44] <postwait_> And has many disadvantages...
[21:32:45] <e^ipi> possibly because it's not a unix kernel
[21:32:47] <delewis> and context switching galore.
[21:32:52] <mon> btw paging doesn't happen that often here
[21:32:53] *** Darwin has joined #opensolaris
[21:32:53] <postwait_> I can't imagine running it for a server OS.
[21:32:56] *** Darwin has quit IRC
[21:33:01] <e^ipi> apparantly apple can
[21:33:03] <postwait_> mon: happens for me enough :-)
[21:33:10] <mon> too bad
[21:33:11] <e^ipi> and they're seeking UNIX03 for leo for some reason
[21:33:19] <postwait_> e^ipi: microsoft does to...
[21:33:21] <toploup33> i am in the process of installing solaris 10 11/06 but i have a huge problem
[21:33:24] <mon> ah well you can never convert those true opensource guys i guess :)
[21:33:25] <delewis> e^ipi, doesn't surprise me.
[21:33:34] <delewis> they're trying to get a lot of the POSIX specifications implemented.
[21:33:39] <delewis> POSIX threads, POSIX AIO, etc.
[21:33:54] <delewis> POSIX threads were implemented properly in 10.2, IIRC, and AIO just got implemented in 10.4
[21:33:57] <e^ipi> mon: who said anything about open source? osx  being crap has nothing to do with it being not open-source
[21:34:06] <postwait_> implementing POSIX in Mach isn't too hard.
[21:34:19] <toploup33> it gives me this error message: Partition 4 extends beyond the end of the disk (c0d0), i googled on that but no success
[21:34:21] <postwait_> They would stand to gain tremendously by pulling the VM into the core kernel.
[21:34:30] <postwait_> It would "break" the mach concept abit...
[21:34:40] <postwait_> but welcome to the difference between theory and practice ;-)
[21:34:41] <toploup33> what can i do?
[21:34:42] <delewis> postwait_, unfortunately, that's one of the things that Mach is all about.
[21:34:45] <delewis> for God knows what reason.
[21:34:56] *** Gman has quit IRC
[21:35:05] <postwait_> delewis.. with 128 cores.
[21:35:06] *** Darwin has joined #opensolaris
[21:35:11] <postwait_> who cares that stuff runs in another process?
[21:35:20] <mon> e^ipi: i meant opensolaris
[21:35:51] <postwait_> this whole multi-core fiasco starts to legitimize Mach more and more.
[21:36:09] <postwait_> I am still not particularly fond of core OS services in userland.. but, if they can make it work -- more power to them.
[21:36:27] <postwait_> The Mach-O file format has tons of advantages over ELF.
[21:37:45] <postwait_> besides, I can create a software mirror on Mac OS X... and on Solaris I get "your metadb overlaps with a slice on another physical disk" error.
[21:37:47] <postwait_> :-D
[21:38:04] <delewis> postwait_, I suspect PEBKAC in that specific instance.
[21:38:13] <delewis> I have no trouble creating mirrors.
[21:38:19] <delewis> have you tried mirroring root with OS X?
[21:38:24] <postwait_> I don't one any of my other machines.
[21:38:32] <postwait_> (no.. I was just being snarky)
[21:38:35] <mrdeviant> i didn't think you could mirror the boot volume in OS X
[21:38:48] <delewis> mrdeviant, not without a lot of low-level hacking from what I've heard.
[21:38:49] <postwait_> the svm stuff works on all my other boxes... but with this one it is all wonky
[21:38:54] <mrdeviant> Apple's built-in utilitiy got rather annoyed with me when I tried to do it
[21:39:00] <postwait_> and I've replaced the drives twice.
[21:39:21] <PerterB> mrdeviant: I thought you could with Soft Raid (3rd party app), but it's not quite the same as having SVM bundled
[21:39:24] <postwait_> I'm thinking there is a problem with the device mapping or the hardware revision on the drives.
[21:39:37] <mrdeviant> postwait_, xnu has the equivalent of a BKL, I think multi-core is going to be a big problem for Apple
[21:40:55] <postwait_> mrdeviant.. good thing the kernel doesn't do much
[21:40:58] <postwait_> It's mach.
[21:41:03] <e^ipi> apple's sold multi-cpu machines for a while
[21:41:08] <mrdeviant> no, it's more than mach
[21:41:15] <mrdeviant> xnu is mach + parts of the FreeBSD kernel.
[21:41:25] <delewis> e^ipi, not very large ones, though.
[21:41:37] <e^ipi> well, how "multi" do you think it's going to get?
[21:41:44] <e^ipi> they've sold 4 core machines before
[21:42:06] <e^ipi> ( 2x dual core G5's , 2x dual core xeons)
[21:43:25] <Kmays> Once we get Xorg 7.2 on Solaris, it'll squash al that old Xsun odditites.. :O
[21:44:15] <e^ipi> we should harras alanc about that more
[21:44:33] *** coffman has joined #opensolaris
[21:44:36] <stevel> he put it back yesterday
[21:44:37] <alanc> harass away, I put back yesterday
[21:44:50] <e^ipi> pick up the pace man!
[21:45:31] <alanc> nv_58 unless QA finds it's really broken
[21:45:33] <e^ipi> ;)
[21:46:30] <Kmays> w00t!! Finally!
[21:46:37] <alanc> though it won't save you from Xsun on SPARC or Sun Ray, so we'll still bear that burden a while longer
[21:46:59] <e^ipi> i don't have a sparc or a sun ray
[21:47:03] <alanc> need to poke at the SPARC wsfb driver a bit more to get 24-bit working, and XVR-100 working
[21:47:04] <e^ipi> so that's not an issue
[21:47:07] <Kmays> darn... is there no escaping it...
[21:47:23] *** Snake007uk has joined #opensolaris
[21:47:36] <alanc> so nv_58 will give you 8-bit Xorg on m64, and that's it for SPARC
[21:47:58] <richlowe> not going to use the cgsix driver they have? :)
[21:47:58] <alanc> but it's a first step, and the code is all out there for people who want to help poke
[21:48:09] *** Snake007uk has quit IRC
[21:48:14] <Kmays> DRI/DRM status?
[21:48:14] <e^ipi> i assume: http://dlc.sun.com/osol/x/downloads/current/ <--that's the correct place?
[21:48:25] <alanc> richlowe: actually, I think cg6 will work with wsfb, but haven't plugged in the Ultra 2 to test it
[21:48:36] <gdamore> it should, in theory, be possible to hack on the Xorg radeon to make it work with XVR-100
[21:48:41] <alanc> e^ipi: I haven't posted a 7.2 tarball yet - but yes
[21:49:00] <gdamore> (although memory mapping registers might not work on SPARC.... ugh)
[21:49:15] <alanc> Kmays: umm, well, DRI won't work so well until the ON guys catch up
[21:49:45] <alanc> the i810 driver in Xorg 7.2 requires a newer version of the kernel i915 DRM module than they ship
[21:49:50] *** _william_ has joined #opensolaris
[21:50:13] <alanc> though someone building ON themselves could probably fix that
[21:50:19] <_william_> hi all
[21:50:26] <gdamore> heh.  i'm trying to hack on i945G hardware right now.  I _hate_ loonix.  the fbcon code is seriously drain bamaged.
[21:52:02] <Kmays> I thought Sun China had an updated DRI
[21:52:11] <gdamore> speaking of which, does the console code we have for XVR-100 in nevada work with any x86 graphics cards?
[21:52:36] <alanc> no idea
[21:52:39] <gdamore> (i.e. do we have a plain framebuffer available in Solaris?  or just X and text console?)
[21:53:33] <Kmays> (there is a note in PSARC a few months ago on Intel driver and ATI driver working)..let me look again
[21:53:41] <alanc> Kmays: they're supposed to be working on it - I don't know when they're putting back
[21:53:58] <alanc> they have an older version of i915 DRI now, ATI is supposed to be later this year
[21:54:41] <alanc> so if you use nv_55 or nv_56 you'll get Xorg 6.9 + DRI on i855 throught i915, but no i945 or i965
[21:55:07] <Kmays> so skip b57?
[21:55:18] <postwait_> So I have SVM working, but the problem is very weird.
[21:55:25] <alanc> nv_57 should be pretty much same as nv_56 for X/DRI
[21:55:28] <postwait_> I have three drives with identical VTOCs.
[21:55:33] <Kmays> ok
[21:55:44] <postwait_> slice4 is the the metadb on each.
[21:55:57] <postwait_> however, I can't add _both_ the 2nd and 3rd drive as metadb
[21:56:04] <postwait_> it says the third drive overlaps with the second.
[21:56:21] <postwait_> If I move s4 on the third drive to a different cylinder range (than drive 2), it works.
[21:56:36] <postwait_> How one drive has anything to do with the other is beyond me.
[21:56:43] <postwait_> It's like the kernel is confused or svm tools are.
[21:57:37] <mon> well t's been nice, byebye
[21:57:40] *** mon has left #opensolaris
[21:58:37] <postwait_> Ah ha!
[21:58:58] <postwait_> On my working box... the "Device ID" is different on all three drives.
[21:59:09] <postwait_> On the busted one.. the last two drives has the same Device ID
[21:59:13] <McBofh> postwait: that's a known issue
[21:59:15] <postwait_> (as reported by metastat
[21:59:30] <McBofh> yes .. check the output of "prtpicl -v" for the devid field for those devices
[21:59:31] <postwait_> That _must_ be confusing svm to thing they are the same
[21:59:32] <BadKarma> meh
[21:59:45] <McBofh> postwait: yup
[21:59:57] <postwait_> yeah
[22:00:03] <postwait_> id1,sd@n5000000000000000
[22:00:04] <postwait_> on both
[22:00:08] <postwait_> how do I change that?
[22:00:15] <McBofh> I don't think you can
[22:00:17] <McBofh> are they fc?
[22:00:22] <postwait_> No SCSI
[22:00:27] <McBofh> model?
[22:00:30] <postwait_> Seagate ST373455LC
[22:00:36] <McBofh> sun-branded?
[22:00:39] <postwait_> No
[22:00:50] <McBofh> well that rules out a sun fw upgrade :)
[22:01:08] <postwait_> Is that a firmware id?
[22:01:11] <McBofh> no
[22:01:35] <McBofh> the firmware isn't providing the correct info to the inquiry command that sd sent, so sd "fakes" a devid
[22:01:53] <postwait_> Ah.. so the firmware needs an upgrade.
[22:01:56] <McBofh> postwait: got sunsolve access?
[22:02:00] <postwait_> Can I pull the firmware off another drive.
[22:02:06] <McBofh> no
[22:02:08] <postwait_> I have _one_ that reports something else
[22:02:14] <postwait_> I have sunsolve access.
[22:02:28] <McBofh> ok, check out srdb document 48730
[22:02:39] <McBofh> it's a long time since I had to look for that ....
[22:02:54] <postwait_> I should be able to get a firmware update from my non-Sun vendor right?
[22:02:59] <postwait_> this is a seagate firmware upgrade.
[22:03:01] <McBofh> I would hope so
[22:03:02] <sickness> evening all
[22:03:03] <postwait_> not a sun one... right?
[22:03:07] <McBofh> postwait: correct
[22:03:08] <postwait_> Okay.. I'll try that first.
[22:03:11] <postwait_> Thanks alot!
[22:03:18] <McBofh> postwait: read the srdb - it'll get you around the problem for the moment
[22:04:42] <McBofh> a few years back I added a workaround entry to the bug that the srdb is based on -- http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=4347117
[22:05:13] <McBofh> postwait: running x86 or sparc?
[22:07:39] *** bank__ has joined #opensolaris
[22:08:06] <bank__> hi
[22:09:57] *** jmcp has joined #opensolaris
[22:10:45] <bank__> hi jmcp
[22:12:42] <tg> hey
[22:13:15] <jmcp> gday
[22:13:39] * jmcp works
[22:13:58] <bank__> bye
[22:14:01] *** bank__ has left #opensolaris
[22:14:26] <sickness> jmcp: you back at sun? =)
[22:15:37] <alanc> he better be - otherwise he's hacked into our internal IRC server too
[22:15:54] <jmcp> :)
[22:15:55] <tg> :)
[22:15:56] <jmcp> sickness: yes
[22:16:03] <jmcp> only on a contract though :()
[22:16:03] <sickness> jmcp: cool! congratulations :)
[22:16:06] <sickness> :/
[22:16:13] <sickness> alanc: lol
[22:17:36] <coffman> heh
[22:20:16] <alanc> it's starting to disturb me how many "NOTICE: core_log: prodreg core dumped" messages I'm getting while live upgrading my nv_53 machine to nv_57
[22:20:45] *** yarihm has joined #OpenSolaris
[22:20:47] *** whaq__ has quit IRC
[22:26:18] *** mazon is now known as Mazon
[22:26:34] <sickness> it's starting to disturb me how many http://rafb.net/p/rTjqha83.html I have in the dmesg and I don't know why :/
[22:29:41] *** dep has joined #opensolaris
[22:30:41] *** freakazoid0223 has quit IRC
[22:31:00] *** lloy0076 has joined #opensolaris
[22:33:21] *** freakazoid0223 has joined #opensolaris
[22:39:30] *** dep has quit IRC
[22:47:20] *** tsoome has joined #opensolaris
[22:48:03] *** lloy0076 has quit IRC
[22:59:47] *** tsoome1 has joined #opensolaris
[23:00:31] *** lloy0076 has joined #opensolaris
[23:04:54] *** lloy0076 has left #opensolaris
[23:05:50] *** Tpenta has joined #opensolaris
[23:05:55] <gdamore> sickness: that sounds bad.  are you using unsigned binaries?
[23:06:04] *** CosmicDJ has joined #OpenSolaris
[23:06:11] <sickness> gdamore: I don't know, I'm using snv55b
[23:06:22] <sickness> gdamore: but here I just found this: http://docs.sun.com/app/docs/doc/817-0552/6mgbi4fgi?a=view
[23:06:26] <gdamore> maybe RE didn't sign them properly
[23:06:28] *** ChanServ sets mode: +o Tpenta
[23:06:31] <sickness> Workaround: None. Ignore the error messages. <- ghgh =)
[23:06:42] <sickness> now I'm about to liveupgrade to snv56
[23:07:10] <sickness> so I'll see if they remain, IIRC they weren't here before snv55b
[23:07:19] <gdamore> lol.
[23:07:52] <sickness> (i'm liveupgrading this sistem since snv47)
[23:07:55] <sickness> eheh :)
[23:08:18] *** tsoome has quit IRC
[23:11:06] <e^ipi> sickness: i've bfu'ed my server since schillix was still current
[23:11:09] <sickness> anyway aside from this warning messages, system works like a charm :)
[23:11:20] <e^ipi> so that'd be what.. b31?
[23:11:22] <sickness> e^ipi: cool =) all works?
[23:11:29] <e^ipi> yeah, never had a problem
[23:11:32] <sickness> e^ipi: yeah something around that, maybe b36
[23:11:37] <e^ipi> just watch the flag days
[23:12:01] <coffman> i wished that liveupgrade would handle zfs proper
[23:12:06] *** bunker has joined #opensolaris
[23:12:46] *** mrdeviant has quit IRC
[23:13:00] <coffman> cause i have my /opt on zfs
[23:13:15] *** hspaans has joined #opensolaris
[23:13:38] <hspaans> g'day all
[23:19:52] *** nostoi has quit IRC
[23:20:04] <sickness> coffman: heh, and zones too =)
[23:24:21] *** _william_ has quit IRC
[23:25:17] *** McBofhnyah has joined #opensolaris
[23:26:23] <coffman> i dont do liveupgrade at servers :)
[23:26:35] <coffman> mainly my workstation/laptop
[23:26:48] <delewis> why? that's what Live Upgrade is suited for.
[23:26:53] <coffman> and zones would be so much away :/
[23:27:00] <delewis> to minimize downtime during an upgrade, and to provide you with something to fall-back on.
[23:27:19] *** McBofh has quit IRC
[23:27:26] <coffman> all my servers have zones :)
[23:31:46] <coffman> and they got 10 03/05, 10u2, 10u3
[23:34:10] *** bunker has quit IRC
[23:38:17] *** tsoome has joined #opensolaris
[23:38:36] * jmcp growls @ sdestika on osol-discuss
[23:38:47] <richlowe> hah.
[23:38:59] *** tsoome_ has joined #opensolaris
[23:39:19] *** tsoome1 has quit IRC
[23:39:50] * delewis suggests that sdestika evaluate Linux/SPARC.
[23:39:57] <jmcp> heh
[23:40:04] <richlowe> jmcp: you mean you guys actually help us?
[23:40:06] <richlowe> jmcp: I'm shocked ;)
[23:40:13] <jmcp> nah ...
[23:40:17] * jmcp looks @ Doc for a counter-example
[23:40:24] <jmcp> s/for/as/
[23:44:55] <tsoome_> anyone installed oracle rac on sun 6540 storage with oracle clusterware?
[23:44:56] *** lopa has joined #opensolaris
[23:48:24] *** gm152 has joined #opensolaris
[23:56:10] *** tsoome has quit IRC
[23:58:08] *** linma has quit IRC
[23:58:34] *** linma has joined #opensolaris
[23:58:39] *** jolts has joined #opensolaris
[23:59:08] *** Mazon is now known as mazon
[23:59:37] *** pikapika has quit IRC
[23:59:58] <lopa> I'm reading the stuff Vista can supposedly do, and it's like someone read my mind and put together a list of top things I absolutely do not want my OS to do, ever, for any reason





top