[00:02:57] <quants> this is wonderous, and frustrating all at the same time
[00:03:24] <quants> a system where I can kill off root
[00:04:16] <Stric> trusted solaris has been shipping without a root account
[00:08:37] *** richlowe has quit IRC
[00:09:42] *** comay_ has joined #opensolaris
[00:10:09] *** comay__ has quit IRC
[00:11:50] *** Gman has joined #opensolaris
[00:12:19] *** tsoome has quit IRC
[00:13:58] <quants> Stric: but it's _in_ the system for use any time
[00:14:19] <Stric> quants: I'm not so sure about that..
[00:14:51] <Qapf> anyone willing to help me with zfs performance? im getting the weirdest behaivor. i have 3 250gb drives im adding to a raidz pool, and they add and work just fine, but when writing to the pool they seem to never write constantly, they just burst, so its 1 second of writing at 50mb/s, and then 3 seconds of nothing, i can format the drives with ufs and mount them and they all work and behave just fine, but even if i create a zpool for a
[00:14:51] <Qapf>  single drive and write to it, i get this weird behaivor
[00:14:54] <Stric> hrm. seems like it exists, but doesn't have any privileges
[00:15:39] *** andrei has joined #opensolaris
[00:16:41] *** richlowe has joined #opensolaris
[00:17:21] <Qapf> how can i start to try and trace where this problem is coming from?
[00:20:49] *** comay_ has quit IRC
[00:21:01] *** comay_ has joined #opensolaris
[00:26:39] <sommerfeld> Qapf: write bursts every 5s is normal for zfs; it accumulates a bunch of stuff to write, then splats it out.
[00:27:01] <sommerfeld> (in the absence of fsync, etc.,)
[00:28:16] <richlowe> lockstat may help.
[00:28:28] <richlowe> it's also worth looking to see if you're on a build affected by the recent ZFS perf regression.
[00:28:34] <richlowe> (sadly, I forget the range, 51-55?)
[00:31:42] *** comay__ has joined #opensolaris
[00:32:06] *** comay_ has quit IRC
[00:32:17] <sommerfeld> richlowe: something close to that.  definitely fixed in 55a but not 55
[00:36:26] <alanc> ooh, SXCR 55 finally, so people can stop asking about it
[00:39:00] *** adlpaf has quit IRC
[00:39:21] <richlowe> alanc: you really think?
[00:39:24] <quants> alanc: bot report %% is out... ;)
[00:39:34] <quants> 55*
[00:39:41] <richlowe> alanc: and it isn't up yet.
[00:40:03] <alanc> Derek sent the mail to opensolaris-announce, so I figured he'd have it up
[00:40:10] <alanc> didn't actually check it
[00:40:12] <richlowe> unless they buggered the URL format and broke my bookmark.
[00:40:35] <richlowe> nope, sxcr_dvd/sxcr_cd takes me to 54, bookmark says 55 ain't there. :)
[00:40:48] <alanc> hmm, takes me to 55
[00:40:58] <quants> lol
[00:41:01] <alanc> http://www.opensolaris.org/os/downloads/sol_ex_dvd/
[00:41:24] <alanc> same for http://www.opensolaris.org/os/downloads/sol_ex_cd/
[00:41:27] <quants> so I gotta install all over again?
[00:41:41] <alanc> that's what live upgrade is for
[00:41:53] <richlowe> until the build where dboot breaks it, at least. ;)
[00:42:17] <richlowe> alanc: ah, they buggered my bookmark *and* steve's scripts.
[00:42:18] <richlowe> alanc: thanks.
[00:42:25] <richlowe> unless steve's just waiting for cron to update him.
[00:42:35] <alanc> I just followed the links from the ON page that Derek posted
[00:42:45] *** comay__ has quit IRC
[00:42:52] <alanc> it should be 55b
[00:42:57] *** comay__ has joined #opensolaris
[00:43:12] <quants> beta?
[00:43:33] <richlowe> second respin.
[00:44:44] <alanc> first we made 55, then they found european keyboards didn't work and ZFS was slow, so they made 55a, then they found more bugs, and made 55b
[00:45:12] <alanc> and just minutes ago, 56 finally came out internally, after waiting for last-minute ON stopper fixes
[00:46:05] <quants> oh frezzes
[00:47:01] <Qapf> sommerfeld, what i am seeing is with a service like samba, ill copy 5gb from another machine onto the share, and instead of writing at a constant 8-9mb a sec for a 100mbit lan, it bounces from 9mb sometimes as low as stalling, the server has a gbit nic so if i jump over to a gbit device and spit data at it, ill surge to about 20-30mb, and then it will stall down to 0 and then jump back up, is this behaivor normal?
[00:47:12] <alanc> of course in X, we've already finished 57, sent it to QA for testing before we integrate to the OS images, and are working on 58...   this train doesn't stop
[00:49:31] <Viking667> erm, what kernel's inside 6/06? Or don't we get to find that out?
[00:49:45] <richlowe> alance: the desktop groups are small, but never sleep?
[00:51:23] <alanc> sleep?   isn't that a libc call?
[00:51:45] <alanc> Viking667: the kernel inside Solaris 10 6/06 is the Solaris 10 6/06 kernel
[00:51:47] *** nprice has joined #opensolaris
[00:52:15] <alanc> so the question doesn't really make sense - what are you trying to find out?
[00:52:29] *** estibi has quit IRC
[00:52:45] <Peanut> Is there something like a "what's new" for SXCR55?
[00:52:54] <alanc> and the desktop groups span the world, so someone's always awake - whether they be in Ireland, Beijing, California, or New Zealand
[00:53:18] *** comay_ has joined #opensolaris
[00:53:23] <quants> wow we got lively w/ a 55 anounce
[00:53:25] <alanc> Peanut: they don't really do what's new for SXCR releases
[00:53:44] <sommerfeld> for OS/Net: http://www.opensolaris.org/os/community/on/flag-days/51-55/
[00:53:54] <nprice> What are some common causes of bad ZFS performance and what can I do to help diagnose it?  Individual drives perform fine but when added to a raidz1 they get horridly slow.
[00:53:56] *** comay__ has quit IRC
[00:53:59] <alanc> they do when they pick one for the main SX release train, but for SXCR, your best sources are the ON flag days, and the changelogs from ON, JDS & X
[00:55:06] <Peanut> sommerfeld: nice link, thanks.
[00:55:26] <Peanut> THough I'm wondering what's so exciting about 55 as compared to 54 is, then :)
[00:55:51] <alanc> for most consolidations, 55 was mostly a bugfix build, not new features
[00:56:04] <alanc> plus it was just before the holidays, so lots of people were out early
[00:56:26] <richlowe> Peanut: not much.
[00:56:31] <alanc> X didn't even deliver to 55, just reused the packages from 54
[00:57:26] <alanc> http://dlc.sun.com/osol/on/downloads/b55/on-changelog-b55.html has a bit more detail
[00:58:06] <richlowe> I think that should contain everything.
[00:58:26] <richlowe> ah, no, it'll be missing changes solely to the closed tree, but otherwise it would. :)
[00:58:30] <sickness> i'm back
[00:58:53] <Peanut> Ah, the Australia timezone fixes *grin*
[00:59:17] *** benh has left #opensolaris
[00:59:54] <sommerfeld> richlowe: Hmm.  The links to source files are broken.
[01:00:05] <richlowe> sommerfeld: src.opensolaris.org snafu
[01:00:08] <Viking667> hm. SO, what's this 55/56/58? Is that release numbers of Sol Express?
[01:00:24] <richlowe> the mercurial interaction goes boom, regularly, so it's not longer present, I suspect steve linked the bits as they used to be.
[01:00:31] <alanc> well, everything for ON
[01:00:38] <alanc> install had a bit of change in 55
[01:00:54] <sommerfeld> Viking667: roughly every 2 weeks a full Solaris development build is released.  that's the build number
[01:01:07] <sommerfeld> (sometimes there are 3 weeks between builds)
[01:01:07] <Viking667> oh.
[01:01:22] <sommerfeld> those appear via solaris express: community release
[01:01:33] * Viking667 has enough trouble just keeping up with ONE release, let alone a release every three-four weeks.
[01:02:17] <alanc> this install bug (fixed in 55!) has bitten me and a few other people I know: 6497684 Unable to login due to live upgrade setting incorrect permissions on /tmp and /var/run
[01:02:19] <sommerfeld> those are a preview of the next Minor release (i.e., the one after Solaris 10, which marketing  gets upset if we call "Solaris 11")
[01:02:59] <Viking667> lol.
[01:03:07] *** Fish- has quit IRC
[01:03:25] <Viking667> Right. Guess I'll stick with what I've got then.
[01:03:36] <richlowe> alanc: yeah, that going away would be awesome.
[01:03:39] <alanc> and this fix in JDS in nv_55 was from a conversation Gman & I had here a few weeks ago: 6505616 firefox should be default browser in Nevada, should not prompt for confirmation.
[01:03:42] *** comay__ has joined #opensolaris
[01:04:00] <richlowe> sommerfeld: you state that as if upsetting marketing would be a bad thing.
[01:04:09] <richlowe> sommerfeld: scaring them into keep consistent names sounds like a feature, to me...
[01:04:11] *** comay_ has quit IRC
[01:04:14] <richlowe> 'into keeping'
[01:04:48] <sommerfeld> I'm merely stating facts, not making value judgements...
[01:05:03] <alanc> whole bunch of StarOffice 8 packaging fixes in 55 too
[01:05:05] <sommerfeld> I'll let someone else pick that fight
[01:05:50] <sommerfeld> I'm fully engaged in a bunch of others (someone claims senior management expects them to integrate something which requires them to ignore Secure by Default rules...)
[01:06:14] <Viking667> ohhh. Where do I find what release of SO is inside 6/06? All I've got so far is Release: 1.00
[01:06:34] <alanc> StarOffice in Solaris 10 is StarOffice 7
[01:06:37] <Peanut> The i945 bugfixes, is that the same chipset as the Appel Macbooks?
[01:07:06] <alanc> I don't remember whether Apple uses i945 or the newer i965
[01:07:11] <sommerfeld> I *think* I've given them the clue required for them to push back..
[01:07:12] <Viking667> alanc: thank you.
[01:07:31] <richlowe> sommerfeld: the ARC handbook "it's a marketing requirement isn't a reason" is all you'd need...
[01:07:38] <Viking667> now all I have to figure out is what diffs there are between 7 and 8.
[01:08:05] <sommerfeld> richlowe: you might think this, but you'd be wrong.
[01:08:15] <richlowe> on a technical or social level?
[01:08:33] <alanc> marketing isn't bound by architectural rules
[01:09:11] <sommerfeld> I was arguing with someone who claimed to be a CISSP.
[01:09:27] <Peanut> alanc: actually, their site says GMA950, but I think the graphics part of that is i945, according to intel.
[01:09:39] <alanc> sounds right
[01:10:12] <alanc> though the Xorg driver changes for i945 haven't made it in yet, since they depend on a 7.1 or newer Xorg server
[01:10:29] *** LordKing has quit IRC
[01:10:30] <andrei> dboot putback today sounds like a really good cleanup
[01:10:35] <andrei> I always hated multiboot
[01:10:39] <alanc> which some lazy ass engineer hasn't gotten done yet since he spends all his time on IRC
[01:10:48] <alanc> 8-)
[01:10:51] <Peanut> *nod* wondering how that would perform, OpenGL-wise.
[01:11:34] <sommerfeld> andrei: yah, though i'm a little nervous about how many bricks will be created in practice by the required grub menu updates..
[01:11:37] *** edwardocallaghan has joined #opensolaris
[01:11:46] <edwardocallaghan> Hi all
[01:12:36] <Viking667> "bricks"?
[01:13:29] <Peanut> You know, a 'brick' - a lump of metal that might have been a computer once, but now, it won't even boot.
[01:14:42] *** comay__ has quit IRC
[01:14:46] <Viking667> mrm. Sounds bad.
[01:14:51] *** comay__ has joined #opensolaris
[01:15:07] * Viking667 wants grub to "grok" Linux partitions
[01:15:08] <sommerfeld> commmon side effect of an upgrade gone horribly wrong.
[01:15:19] <sommerfeld> very common if you're doing OS development
[01:15:43] <Viking667> The Linux grub works fine, but needs persuading to boot Solaris. "rootnoverify (hd2,0); makeactive; boot"...
[01:21:50] <sommerfeld> i'm now firmly convinced that a CISSP certification does not have positive value.  not sure whether it is of zero value or negative value, though.
[01:22:09] <richlowe> sommerfeld: I've bfu'd prior xen bits, and the only brickification was bug induced, not grub induced.
[01:22:27] <richlowe> sommerfeld: the larger problem is that it can't mess with entries it didn't create, so if you're using LU or similar *nothing* is fixed on your behalf.
[01:22:28] *** gm152 has joined #opensolaris
[01:22:31] <Viking667> What I was rather after was more how the Solaris grub works:  root (hd2,0,a); kernel ..../multiboot; module whatever/it/is/boot-archive
[01:22:32] *** comay has joined #opensolaris
[01:22:45] <Viking667> Any clues of what I need to do?
[01:24:19] <richlowe> actually, the larger issue with that was the console server transforming from a rather nice ssh-accessible device to a cloud of acrid smoke mid-way through unbricking.
[01:24:20] * Peanut is off to bed, good night all and have a fun weekend.
[01:24:45] <hspaans> slaapse Peanut
[01:24:52] <sommerfeld> richlowe: ick.
[01:24:55] <Peanut> Hey hspaans , dank en trusten :-)
[01:25:11] <sommerfeld> "who let the smoke out?"
[01:25:22] *** comay_ has joined #opensolaris
[01:25:44] <edwardocallaghan> Oh wants to have some fun
[01:25:47] *** comay__ has quit IRC
[01:25:49] <edwardocallaghan> perl -e 'for($i=400;$i<4000;$i+=0.01) { print chr((sin($i*$t++/8000) * 127) & 255);}' > /dev/dsp
[01:26:05] <alanc> once you let the smoke out, you can never get it back in, and it won't work the same without the smoke in it
[01:26:29] <Viking667> won't work at all, in fact.
[01:26:37] <Viking667> no smoke, no work.
[01:26:53] <Viking667> cos smoke is gaseous daemons, you see...
[01:28:17] <edwardocallaghan> tcpdump -i web  -nn -w /dev/dsp
[01:29:13] <Viking667> "web"?
[01:29:52] <edwardocallaghan> It was a joke
[01:30:46] *** richlowe changes topic to "Latest builds: SXCR: 55 | ON: 55"
[01:30:59] <edwardocallaghan> ar cool
[01:31:02] <Viking667> lol
[01:32:01] <Viking667> ummm.
[01:33:37] *** comay has quit IRC
[01:33:44] <Viking667> why two nn?
[01:33:47] *** comay has joined #opensolaris
[01:34:32] <tomww> n is numeric, nn ist much more numeric :-)
[01:34:46] <edwardocallaghan> cat /dev/urandom > /dev/null &
[01:35:09] <Viking667> ah.
[01:35:16] <Viking667> Well, I'm still not hearing anything...
[01:35:51] <edwardocallaghan> perl -e 'for($i=400;$i<4000;$i+=0.01) { print chr((sin($i*$t++/8000) * 127) & 255);}' > /dev/dsp
[01:36:12] <edwardocallaghan> cat /dev/urandom > /dev/dsp
[01:36:27] *** comay_ has quit IRC
[01:36:47] *** sommerfeld has quit IRC
[01:36:50] *** comay_ has joined #opensolaris
[01:38:13] <Viking667> heh. Guess things aren't working well. I'm getting quarter second chunks about every thirty seconds.
[01:39:11] *** Kmays has joined #opensolaris
[01:39:28] <Viking667> so much for having a dialup connection
[01:42:10] <Kmays> I'll be glad when FreeBSD dumps UFS2 for ZFS.
[01:42:47] <delewis> Kmays, doubt that'll happen if ZFS still has the same hardware requirements for adequate performance as it does on Solaris.
[01:43:03] <Kmays> :(
[01:43:18] <delewis> its worthwhile to point out that a large percentage of the FreeBSD userbase are users that using "constraining" hardware.
[01:43:24] <delewis> are*
[01:43:53] <alanc> I'd be surprised if they picked a non-BSD-licensed FS as their primary FS instead of just an option
[01:44:19] *** comay__ has joined #opensolaris
[01:44:35] <delewis> indeed, and XFS isn't an option for "everyone's filesystem"
[01:44:44] *** comay has quit IRC
[01:44:47] <delewis> its largely suited for large files and has some memory requirements, itself.
[01:47:45] *** comay has joined #opensolaris
[01:47:47] *** comay_ has quit IRC
[01:49:56] *** Viking667 has quit IRC
[01:51:01] * tsoome1 is away - bbl - [AcidJazz - Log ON]
[01:51:10] *** dunc has joined #opensolaris
[01:55:18] *** comay__ has quit IRC
[01:55:28] *** comay__ has joined #opensolaris
[01:58:35] *** comay has quit IRC
[01:58:59] *** comay has joined #opensolaris
[02:00:06] *** peteh has quit IRC
[02:00:30] *** lloy0076 has joined #opensolaris
[02:00:35] *** peteh has joined #opensolaris
[02:04:55] *** adp` has left #opensolaris
[02:05:59] *** comay_ has joined #opensolaris
[02:06:27] *** comay__ has quit IRC
[02:08:18] *** Mazon is now known as mazon
[02:09:23] *** comay__ has joined #opensolaris
[02:09:50] *** comay has quit IRC
[02:16:32] <lloy0076> heh
[02:16:54] *** comay_ has quit IRC
[02:16:56] <lloy0076> When I was downloading 4 parts of ON 55 at once, I got 200KB/sec for each part.
[02:17:09] <lloy0076> Now that it's download only 2 parts, I'm getting 70KB/sec for each part.
[02:17:12] <lloy0076> Go figure :P
[02:17:19] *** comay_ has joined #opensolaris
[02:19:06] <Kmays> delewis <- The great three-letter company told me GPFS seeks a challenger to the throne. ;)
[02:20:20] *** comay__ has quit IRC
[02:20:33] *** comay__ has joined #opensolaris
[02:21:07] <delewis> that three-letter company always says silly things :-)
[02:22:02] <alanc> good thing I work at a company with lots more than three letters
[02:22:10] <alanc> Microsystems is a long word...
[02:22:41] <delewis> alanc, watch out for those that claim the name is still "SUN" and an acronym.
[02:22:54] <Kmays> I told them to 'take that sword out of their stone'..but they just laughed...
[02:23:58] <alanc> well, of course a true Berkeley alum like myself wouldn't be working at a company named "Stanford University Network" - that's preposterous
[02:24:30] <delewis> IBM is just one of those companies with a double personality -- "Hey, here's Linux. Well, if you really, really want AIX because you know its better, we'll hook you up with a support contract!"
[02:24:36] <delewis> alanc, true :-)
[02:25:13] <alanc> but I've had enough work for today, so I'm outta here - see you all on Monday....
[02:25:18] *** alanc is now known as alanc-away
[02:25:26] <Kmays> alanc <- Imagine the cost to have that put on a T-shirt nowadays.
[02:26:40] *** alanc-away is now known as onbot
[02:27:00] <onbot> new ON flag day message: FLAG DAY: IP Instances: http://www.opensolaris.org/os/community/on/flag-days/pages/2007011902/
[02:27:05] *** onbot is now known as alanc-away
[02:27:27] * alanc-away is really leaving now 8-)
[02:27:58] <lloy0076> This might sound silly, but I thought that onbot was a BOT.
[02:28:20] *** comay_ has quit IRC
[02:28:31] *** comay_ has joined #opensolaris
[02:28:37] <delewis> lloy0076, we tried getting AJ_ZO in here to be onbot, but that fell through :-(
[02:29:36] <lloy0076> Is AJ_ZO a Hungarian doll which was abducted at age 3 and taken to the depths of Siberia, where a mummy in the museum cursed the doll to roam the world never to see its homeland again?
[02:29:37] *** coffman has joined #opensolaris
[02:30:04] <delewis> no, AJ_Z0 is the bot in #solaris that uses a very small flat textfile for quotes.
[02:30:17] <lloy0076> Ah
[02:31:31] *** comay__ has quit IRC
[02:31:55] *** comay__ has joined #opensolaris
[02:32:55] *** slowhog has joined #opensolaris
[02:33:09] *** slowhog has quit IRC
[02:33:28] <Kmays> xorg 7.2..out in b58 I hear.
[02:33:31] *** slowhog has joined #opensolaris
[02:34:42] *** broadcast has quit IRC
[02:34:43] *** slowhog has quit IRC
[02:34:43] *** Tpenta has quit IRC
[02:35:22] <edwardocallaghan> see you guys later
[02:35:25] *** edwardocallaghan has left #opensolaris
[02:36:59] *** slowhog has joined #opensolaris
[02:37:00] *** Tpenta has joined #opensolaris
[02:37:42] *** ChanServ sets mode: +o Tpenta
[02:39:27] *** comay_ has quit IRC
[02:39:54] *** comay_ has joined #opensolaris
[02:41:48] *** nachox has joined #opensolaris
[02:42:51] *** comay has joined #opensolaris
[02:42:56] *** comay__ has quit IRC
[02:49:12] *** abackos has joined #opensolaris
[02:49:49] <richlowe> alanc-away: heh, I'd kind like to see a bot deal with that, but getting them mailed out is of a more practical use.
[02:50:52] *** comay_ has quit IRC
[02:51:13] *** comay_ has joined #opensolaris
[02:52:37] * lloy0076 sigh
[02:53:11] <lloy0076> PHP5 uses GCC...nothing wrong with that...but WAIT!
[02:53:20] <lloy0076> MySQL uses C++...
[02:53:35] *** comay has quit IRC
[02:53:36] <lloy0076> And the packages on mysq.com are obviously built with Sun Studio...
[02:53:58] <lloy0076> ...because those libraries get the infamous ABI link errors.
[02:54:01] *** comay has joined #opensolaris
[02:54:06] <lloy0076> I think I may have to roll my own MySQL5 :(
[02:54:23] <richlowe> or build PHP with studio.
[02:54:49] <lloy0076> richlowe: I suspect I'm linking against more GNU *crud* than Sun Studio crud.
[02:55:12] <delewis> MySQL builds with Sun Studio
[02:55:18] <delewis> 5.0.27 does, anyway.
[02:55:19] <lloy0076> I'll have a go at it.
[02:55:28] <lloy0076> delewis: That's my whole point :P
[02:55:44] <lloy0076> delewis: I'm using the GNU toolchain which has "enhanced" the C++ ABI.
[02:55:57] <lloy0076> (so that it's incompatible with anything else)
[02:56:00] *** Jiko_ has joined #opensolaris
[02:56:10] <delewis> lloy0076, there is no standardized C++ ABI
[02:56:23] <delewis> and that's why you shouldn't write code in C++ and expect it to be portable :-)
[02:56:45] <lloy0076> Someone posted a link to a document which basically said that the GNU guys knew what they were doing would make it incompatible.
[02:57:02] <lloy0076> Apparently it's meant to be "better" but is "faster but incompatible" a good tradeoff?
[02:57:07] * steleman could point to the same link and explain that the GNU guys did not know what they were doing
[02:57:20] <delewis> yes, I saw that as well, but the article also claimed C had no standardized ABI, which is blatantly false.
[02:57:26] <delewis> so I'd question the validity of the article :-)
[02:57:39] <steleman> delewis: there are ABI problems in C as well between compilers
[02:58:10] <lloy0076> steleman: I don't think delewis might argue about that.
[02:58:24] <lloy0076> steleman: He is saying, though, that there IS a standard to throw at incompatible compiler writers.
[02:58:32] <steleman> therefore there is no standardized C ABI
[02:58:44] <steleman> if there were one, there would be no problems
[02:58:59] <lloy0076> This could degenerate into: "Please define what standard means" debate.
[02:59:28] <steleman> that has already been established: THAT which is codified in an ANSI/ISO/IEC document.
[02:59:59] <nachox> huh? "there are abi problems between C compilers therefore there is no standard C abi"? logic fails bad there
[03:00:15] <steleman> nachox: quote the ISO/IEC document
[03:00:54] <richlowe> the ABIs are codified, just not there.  Given that objects from different OSes or ISAs are unlikely to be linked...
[03:01:09] <steleman> im talking about same OS, same ISA
[03:01:44] *** comay__ has joined #opensolaris
[03:02:14] *** comay_ has quit IRC
[03:04:30] *** andrei has left #opensolaris
[03:04:58] *** comay_ has joined #opensolaris
[03:05:10] *** comay has quit IRC
[03:05:31] *** swa_wtf is now known as swa_mobil
[03:06:16] <nachox> sun studio and gcc?
[03:07:38] <steleman> nachox: yup. C with gcc and C with sunstudio
[03:09:24] <delewis> steleman, is a matter of not complying with SVID (the generic and processor-specific portions)?
[03:09:27] <delewis> it*
[03:10:06] <steleman> it's actually several things: structures packed differently and alignment of struct elements doens't match between compilers
[03:10:33] <delewis> that sounds like non-compliance with the spec.
[03:10:40] <delewis> the processor-specific portion covers all of that.
[03:10:59] <delewis> http://www.sparc.com/standards/psABI3rd.pdf
[03:11:09] *** Jiko has quit IRC
[03:11:50] <delewis> so a standard is there -- it just seems certain suspect compilers aren't complying with it :-)
[03:12:02] *** alfism has quit IRC
[03:12:10] <delewis> which is contrary to the situation with C++ -- no published ABI standards.
[03:12:32] <steleman> there are ABi standards for C++ published. very many of them. ;-)
[03:12:47] *** comay__ has quit IRC
[03:13:05] <delewis> steleman, every compiler has their own? :-)
[03:13:06] *** comay has joined #opensolaris
[03:13:15] <steleman> delewis: yes every compiler on every ISA has their own.
[03:13:27] <steleman> :-P
[03:13:46] <steleman> more precisely: try compiling ffmpeg and/or libmad or a52dec with Studio
[03:13:55] <steleman> or h264
[03:13:58] <steleman> it will compile
[03:14:01] <steleman> it just wont work
[03:14:16] <steleman> they only work if you rewrite all the structs
[03:14:25] <steleman> (and some of the asm in ffmpeg)
[03:14:33] <delewis> right
[03:14:42] <delewis> the code is pretty dependent upon how gcc packs structures from I've seen.
[03:14:53] <steleman> that code is messed up royally
[03:15:07] <delewis> don't get me started :-)
[03:15:31] * delewis had a bout with MPlayer's signal handling code for SPARC
[03:15:36] <steleman> and because ffmpeg and a52dec have been having a steaming affair ever since, you can either build both with gcc or both with studio but cant mix
[03:15:51] *** comay_ has quit IRC
[03:15:54] <delewis> enough that I've last interest in "trying" to maintain the code.
[03:15:58] <nachox> if there is a published standard that specifies the abi for a given isa and os then there is a bug in the non compliant compiler and you should report it
[03:16:10] <delewis> but when 99.9% of the developers (a) use Linux (b) use x86 and (c) haven't used anything else
[03:16:13] <delewis> things can be quite difficult
[03:16:16] *** comay_ has joined #opensolaris
[03:16:27] <delewis> nachox, I'm consulting the Sun Studio documentation at the moment to see what it complies with.
[03:16:38] <steleman> the problem in ffmpeg is that the actual rendering depends on the offset of struc telements
[03:16:44] <steleman> s/struc/struct/g
[03:17:01] <steleman> because they do some casting magic
[03:17:06] *** nachox has quit IRC
[03:17:09] *** nachox_ has joined #opensolaris
[03:17:10] <delewis> steleman, yes, and they're still under the impression they can write faster code than a decent compiler.
[03:17:11] *** nachox_ is now known as nachox
[03:17:25] <delewis> perhaps they can, given the compiler is gcc.
[03:17:29] <steleman> and they use float all over the place which they actually dont need
[03:17:56] <delewis> if you can write faster code than Sun Studio, its a bug.
[03:18:15] <steleman> it took me 3 months to get ffmpeg working with studio
[03:18:26] <steleman> and then of course when i was done they came out with a new version
[03:18:49] <delewis> steleman, they stopped doing versioning by the time I started working with the code
[03:18:55] <delewis> "just grab the SVN sources"
[03:19:12] <steleman> i stopped at 0.4.9.<something>
[03:19:44] <steleman> that's the one i have and i'm sticking to it
[03:21:29] *** coffman has quit IRC
[03:22:30] <delewis> steleman, most of the developers are under the impression that if you remove the #attribute's the code will work everywhere :-)
[03:22:50] <steleman> i removed them because studio 10 wont digest them anyway
[03:23:00] <steleman> but that wasn't enough
[03:23:08] *** comay has quit IRC
[03:23:13] <steleman> i had to rearrange the struct elements and then pad the structs
[03:23:38] <delewis> and get rid of all the inline asm?
[03:23:54] <steleman> yah studio doesn't digest gnu asm
[03:24:38] *** triplah_ has quit IRC
[03:24:39] *** comay has joined #opensolaris
[03:24:57] <delewis> hmm, I still can't find anything in the Sun Studio documentation to see if it complies with SCD (which encapsulates SVID)
[03:25:10] <delewis> only C89/C99 compliance
[03:25:15] *** edwardocallaghan has joined #opensolaris
[03:25:26] <steleman> i vaguely remember that studio C is based on pcc
[03:25:50] <edwardocallaghan> Hi all
[03:26:02] <edwardocallaghan> boyd: You around ?
[03:27:09] *** comay_ has quit IRC
[03:27:29] *** comay_ has joined #opensolaris
[03:27:32] <delewis> steleman, I doubt a lot of that code is still lying around :-)
[03:27:48] <steleman> delewis: lol well i'm sure sun has a copy of it somewhere
[03:27:49] * delewis learned about er_src a couple of weeks ago
[03:28:25] <steleman> :-)
[03:29:00] <delewis> and I never had a clue that writing column-wise multidimensional array code in C was not optimal
[03:29:07] <nachox> what does that do?
[03:29:16] <delewis> fortunately, Sun Studio is smart enough to convert the code to row-wise
[03:29:19] <nachox> er_src i mean
[03:29:34] *** MikeTLive has joined #opensolaris
[03:29:35] <delewis> nachox, compile code with debugging symbols, and you can use er_src to see what optimizations were done where in the code
[03:29:45] <delewis> its quite useful
[03:30:13] <nachox> the problem with column-wise are probably the cache misses right?
[03:30:21] <delewis> nachox, correct
[03:30:27] <steleman> yup
[03:30:41] <delewis> interestingly enough, Fortan does column-wise
[03:30:46] <delewis> Fortran*
[03:31:42] <nachox> it would deppend on how fortran allocates memory for the array i guess
[03:31:49] <delewis> right
[03:34:46] <steleman> __asm __volatile("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb));
[03:35:09] <steleman> ffmpeg beauty sample
[03:35:14] *** comay__ has joined #opensolaris
[03:35:19] <delewis> ugh.
[03:35:20] <nachox> youre sick :P
[03:35:27] <delewis> steleman, they're still stuck in the past.
[03:35:30] *** comay has quit IRC
[03:35:45] <steleman> well it just shows that gcc doesn't optimize that well
[03:36:05] <steleman> if you have to resort to this to get decent code
[03:36:19] <delewis> steleman, yes
[03:36:22] <delewis> precisely my point
[03:36:26] <steleman> this multiplies two 16-bit signed ints and += it into an unsigned 32-bit
[03:36:56] <steleman> now i am quite certain studio can do this very well in C
[03:37:33] *** comay_ has quit IRC
[03:37:59] *** comay_ has joined #opensolaris
[03:38:32] <delewis> http://pastebin.com/863330
[03:40:50] * delewis would like to see gcc do that
[03:41:03] <delewis> having an er_src-equivalent that is
[03:41:18] <nachox> those are the comments generated by er_src?
[03:41:31] <steleman> :-)
[03:41:56] <delewis> nachox, yes
[03:42:10] <nachox> nice
[03:42:41] <delewis> I assume gcc is capable enough to fix multi-dimensional code that's written in a non-optimal manner..
[03:44:54] *** Netwolf has quit IRC
[03:45:34] <nachox> compile that with o3 with gcc and disassemble it
[03:45:48] * lloy0076 sigh
[03:46:02] <steleman> or save it with -S
[03:46:03] <lloy0076> I didn't call my friend when he asked me to so what does he do? He calls me.
[03:46:22] *** comay__ has quit IRC
[03:46:23] <lloy0076> So I'm there basically just saying "Yes, no" and eventually he picks up that I don't want to talk.
[03:46:34] *** comay has joined #opensolaris
[03:46:44] <sniffy> lloy0076, Gotta love those moments
[03:47:02] *** abackos has left #opensolaris
[03:48:30] *** comay__ has joined #opensolaris
[03:48:46] *** comay_ has quit IRC
[03:50:19] *** mega has joined #opensolaris
[03:51:09] <delewis> hah.
[03:51:14] <delewis> its not interchanging the loops
[03:52:11] <delewis> http://pastebin.com/863334
[03:52:15] <delewis> or am I reading that wrong?
[03:52:32] <delewis> I assume the instructions in .LL9 are being executed first
[03:52:47] <delewis> no, wait. It is interchanging the loops
[03:53:34] <nachox> damn pastebin...
[03:54:03] <nachox> sparc asm?
[03:54:12] <delewis> yeah
[03:54:41] <mega> hi
[03:54:53] <delewis> actually its not interchanging them, unless I'm missing something. .LL9 then.LL8, and then it branches back to .LL9
[03:55:59] <nachox> ble is branch less or equal? and bpos?
[03:56:39] <delewis> nachox, yes, on ble, not sure about bpos.
[03:57:37] *** comay has quit IRC
[03:57:44] *** deedaw has quit IRC
[03:57:47] *** comay has joined #opensolaris
[03:57:58] <delewis> bpos is branch on positive
[03:58:03] <nachox> i'd try the same in x86 since that's what gcc is better with
[03:59:20] *** comay__ has quit IRC
[03:59:46] *** comay__ has joined #opensolaris
[04:02:29] <quants> I'll be back again later, I'm just feeling so sick the last few days
[04:02:47] <quants> see yoiu guys, and take care
[04:02:57] *** quants has quit IRC
[04:06:51] <jamesd__> remind me never to take a job writing sparc assembly that i can't cheat and make the c compiler do all the work
[04:07:13] <steleman> jamesd__: as long as you use studio you'll be fine :-)
[04:07:24] *** peteh has quit IRC
[04:07:50] <delewis> SPARC assembly isn't *that* bad, though, it can be tricky reading it if you don't know how delayed branch instructions work.
[04:07:55] <nachox> sparc assembly is that bad? never used it in my life
[04:08:08] <delewis> nachox, no
[04:08:13] <delewis> the instruction set is quite simple
[04:08:26] <nachox> the closest i have been was using arc at the university
[04:08:27] *** peteh has joined #opensolaris
[04:08:38] <delewis> but like I said, if you don't know how delayed branch instructions work, it can make life interesting. :-)
[04:08:42] *** comay has quit IRC
[04:08:48] <jamesd__> i was refuring to the  5 lines of code to  do    a = 0x43434343434
[04:09:36] <edwardocallaghan> I love RISC
[04:09:55] *** peteh has quit IRC
[04:10:10] *** comay_ has joined #opensolaris
[04:10:33] *** peteh has joined #opensolaris
[04:10:45] *** comay__ has quit IRC
[04:10:54] <delewis> nachox, SPARC assembly used to be taught in most undergraduate architecture courses with good reason.
[04:10:57] *** lloy0076 has quit IRC
[04:11:08] <delewis> (though, nowadays, MIPS seems to be the standard, but for the same reason)
[04:11:27] <nachox> we have x86
[04:11:31] <delewis> ugh.
[04:11:38] <delewis> that's the *worst* thing you could possibly teach.
[04:11:59] <delewis> "Which addressing mode should we use today class?"
[04:12:20] <nachox> not really, when the world is dominated by x86 desktops, it's just easier to have students practice
[04:12:40] *** Theoden-Nexenta has joined #opensolaris
[04:12:47] <delewis> nachox, no, as even x86 is taking the RISC route, nowadays.
[04:13:04] <nachox> amd64 fixed that these days
[04:13:25] <delewis> nachox, you still have different addressing modes.
[04:13:28] <delewis> and that's just silly.
[04:13:35] <delewis> and makes things way too complicated for students.
[04:13:36] <nachox> only for compatibility
[04:13:49] <delewis> its much easier to write a compiler for MIPS or SPARC than it is to write one for x86.
[04:14:00] <delewis> nachox, no the addressing modes are still there.
[04:14:09] <nachox> you can bet on that, yes :P
[04:14:11] <delewis> segmented memory management is not, however.
[04:15:17] <delewis> and don't get me started on the complexities of the orthogonal register sets..
[04:17:33] <nachox> i never said it was easy :)
[04:18:04] <Error_e^ipi> i liked SPARC asm...
[04:18:21] <Error_e^ipi> PPC was a pain, and x86 is stupid
[04:18:26] <jamesd__> all assembly language is easy... just code in C and  execute   gcc -S  .. and then convert stack based variables to register based ones.
[04:18:45] <nachox> when i read about register window i got scaared
[04:18:48] <delewis> PowerPC is a pain because it doesn't have named registers, from what I understand.
[04:18:49] <nachox> :P
[04:18:50] <edwardocallaghan> little endian
[04:18:57] <delewis> nachox, its not bad.
[04:19:14] <delewis> and makes a compiler-writer's life easier
[04:19:30] <delewis> "Do we have to rewrite this compiler to add 5 more GP registers? NO!"
[04:20:55] <delewis> register windows on pre-SPARCv9 were a bit of a detriment to performance, as all the registers had to be flushed back to memory when a context switch occured.
[04:21:10] *** comay_ has quit IRC
[04:21:13] <delewis> so most of the criticisms you'll find for register windows are only relevant for pre-SPARCv9
[04:21:13] <edwardocallaghan> I hate the endianness of x86, I don't go any where near it and I'm a student
[04:21:16] *** comay_ has joined #opensolaris
[04:22:27] <delewis> registers on SPARCv9 have a dirty bit, so only modified registers need to be flushed back to memory.
[04:24:11] <edwardocallaghan> You know this BrandZ thing, can you install windows in that ?
[04:24:16] <delewis> edwardocallaghan, no
[04:24:30] <delewis> all BrandZ does is remap Linux system calls to Solaris system calls.
[04:24:34] <edwardocallaghan> What's the difference with that and containers ?
[04:24:43] <edwardocallaghan> Oh
[04:24:45] <nachox> a completely different thing
[04:25:09] <nachox> brandz uses special containters
[04:25:22] <edwardocallaghan> Is it possible to do something similar for a Windows app ?
[04:25:43] <delewis> edwardocallaghan, no.
[04:25:59] <delewis> lots of the Windows windowing code is stuffed in the kernel-level.
[04:26:06] <edwardocallaghan> I know Microsoft don't tell about there API's but I was wondering how to run something on Solaris x86
[04:26:13] <nachox> linux and solaris have fairly similar system calls, solaris and windows, do not
[04:26:19] <Doc> win4solaris
[04:26:26] <delewis> edwardocallaghan, also true.
[04:26:33] <delewis> Windows system calls have never been publicly specified
[04:26:36] <Doc> http://blogs.sun.com/ThinGuy/entry/coming_soon
[04:26:44] <Doc> running windows on a solaris system
[04:26:45] <edwardocallaghan> does that exist ? win4solaris ?
[04:27:10] <Doc> no, i just made it up and pasted a blog entry about it to annoy you
[04:27:17] <nachox> and there is qemu too to run a windows guest
[04:27:22] <nachox> doc :)
[04:28:07] <nachox> Doc: btw, is it as slow as the flash video? sounds impossible to use...
[04:28:27] <Doc> yes, because flash is designed for 30 frame/second video streaming
[04:28:38] <edwardocallaghan> Thanks
[04:31:44] *** comay__ has joined #opensolaris
[04:32:13] *** comay_ has quit IRC
[04:33:26] <edwardocallaghan> Where do I find Win4Solaris ?
[04:33:43] <edwardocallaghan> I only know Win4Lin and BSD
[04:39:19] <edwardocallaghan> Do I just use the BSD version ?
[04:39:19] *** jsubl2 has quit IRC
[04:39:36] <edwardocallaghan> Or the Linux version in a BrandZ
[04:39:42] <edwardocallaghan> :/
[04:39:49] <Auralis> from the coming soon i would guess you wait till its released
[04:40:02] <edwardocallaghan> ?
[04:40:28] <Auralis> the name of the blog entry "coming soon"
[04:40:49] <edwardocallaghan> Oh yea sorry
[04:42:42] *** comay__ has quit IRC
[04:42:45] <edwardocallaghan> Does crossOverOffice work in BrandZ ?
[04:42:50] *** comay__ has joined #opensolaris
[04:43:14] <edwardocallaghan> I have not used Windows for years ! nor do I plan too
[04:43:40] <delewis> would you like a cookie now?
[04:43:53] <nachox> yes, please
[04:44:13] <nachox> but you pay for the shipping to argentina
[04:44:23] <nachox> :P
[04:45:27] <edwardocallaghan> lol
[04:45:32] <edwardocallaghan> I want one!
[04:45:51] <edwardocallaghan> stupid AVR32 IDE only works on Windows
[04:46:16] <edwardocallaghan> You know for the Atmel u-controllers
[04:46:27] <edwardocallaghan> I do lots with PIC and Atmel
[04:46:40] <edwardocallaghan> Moving to Atmel now though
[04:48:59] <edwardocallaghan> What would it take Sun to ask Atmel to port there IDE to Solaris ? http://www.atmel.com/dyn/products/tools_card.asp?tool_id=2725
[04:49:38] <Tpenta> some customers requesting it might be a start
[04:49:46] <nachox> more costumers interested in it
[04:51:09] *** Jiko_ is now known as Jiko
[04:51:19] *** triplah_ has joined #opensolaris
[04:52:21] <edwardocallaghan> How would I get people to request it ?
[04:52:32] <edwardocallaghan> And where do I make requests ?
[04:52:46] *** comay__ has quit IRC
[04:52:47] <edwardocallaghan> I am sure I seen the a page before ... ?
[04:53:18] *** comay__ has joined #opensolaris
[04:53:47] *** slowhog has left #opensolaris
[04:55:43] *** Gman has quit IRC
[05:04:16] *** comay__ has quit IRC
[05:04:44] <edwardocallaghan> Any one complied OpenSolaris for MIPS so I can try it on my O2 ?
[05:05:11] <nachox> it would have to be ported to mips first
[05:08:07] *** mega has quit IRC
[05:09:21] *** freakazoid0223 has quit IRC
[05:10:40] <delewis> which would take considerable effort
[05:11:03] <delewis> not to mention most (if not all) of the specs for the hardware in that O2 aren't publicly available.
[05:11:24] <delewis> SGI's hardware is highly proprietary and private.
[05:11:47] <nachox> porting it to mips would put solaris closer to the embeded world though
[05:11:52] <delewis> hene why Linux still lacks support for a lot of the SGI hardware.
[05:11:54] <delewis> hence, rather.
[05:12:04] *** freakazoid0223 has joined #opensolaris
[05:12:07] <delewis> I believe gdamore was interested in a MIPS port
[05:12:18] <delewis> unfortunately, that's not enough interest to do a port
[05:13:23] <nachox> interest as in cash? :)
[05:14:10] <delewis> nachox, I believe some of the difficultly came from (a) a lack of hardware (b) lack of man-power -- it'd take a huge development effort.
[05:14:29] <delewis> you see how far the PowerPC port has gotten and that had some help from Sun employees (not a lot, given)
[05:15:56] <richlowe> a mips port would also be very different from "something that ran on the O2".
[05:16:06] <richlowe> given how little of SGI's hardware actually has available documentation...
[05:16:39] <nachox> plus, since sgi abandoned mips there is probably 0 interest
[05:16:47] <delewis> less than 0 :-)
[05:17:00] <delewis> and finding people familiar with SGI MIPS internals would be difficult.
[05:17:17] <nachox> just hire all the devels sgi fired ;)
[05:17:40] <delewis> but yeah, to do a MIPS port, you'd have to port all the platform-dependent code, which includes a HAT layer, device drivers, etc.
[05:17:51] <delewis> interrupt handlers, system calls, etc.
[05:18:30] <delewis> fortunately, all that code is at least easily identifiable.
[05:18:31] <nachox> i thought solaris had less platform dependant code than linux, it probably has less developers too
[05:18:52] <delewis> nachox, yes and the platform-dependent code can be easily identified.
[05:19:09] <edwardocallaghan> http://www.opensolaris.org/jive/thread.jspa?threadID=4024&tstart=0 I was reading this
[05:19:22] <edwardocallaghan> OK I'll put NetBSD on there then
[05:27:58] <nachox> it's funny, ever since i saw niagra i thought it would make a great appliance to route and filter network traffic, an appliance with solaris
[05:29:08] <delewis> nachox, sure, given Cisco has shown an M68k makes a great packet filter and router, too.
[05:29:29] <nachox> iss uses plain x86 boxes for their own
[05:29:44] <delewis> Cisco uses x86, nowadays, too.
[05:29:48] <Doc> 15 years ago it was very, very common to use sun boxes as routers
[05:30:07] <Doc> hell, solaris still acts as a router by default if it's got more than 1 network interface
[05:30:12] <delewis> Doc, there's probably a lot of FW-1 deployments still out there.
[05:30:25] <Doc> yah.. that's a bit different tho 'cos it
[05:30:32] <Doc> yah.. that's a bit different tho 'cos it's not straight solaris/sunos
[05:31:15] <nachox> specially now with the new network stack
[05:32:51] <jbk> i've been strongly considering it
[05:33:03] <jbk> except need to find something cheap and low-heat that'll run solaris
[05:33:22] <nachox> that would be a really nice thing to test, how much network traffic would it take to saturate a niagra proc?
[05:33:33] <delewis> I've been using ipfilter/ipnat with Solaris 9 for the last 3 years.
[05:33:34] <delewis> with little trouble.
[05:33:44] <delewis> given I'm not doing anything terribly complicated -- just NAT and packet filtering.
[05:34:07] <edwardocallaghan> Going to bed
[05:34:11] <edwardocallaghan> Night night all
[05:34:54] *** edwardocallaghan has left #opensolaris
[05:39:15] <nachox> i meant something more complicated than just nat and firewall, i meant a full blown ips anyways
[05:55:13] <steleman> this cmake thing is quite nice
[06:00:27] *** Theoden-Nexenta has left #opensolaris
[06:03:55] *** uebayasi has joined #opensolaris
[06:05:21] *** freakazoid0223 has quit IRC
[06:07:45] *** jsubl2 has joined #opensolaris
[06:10:39] *** bondolo has quit IRC
[06:12:04] *** nachox has quit IRC
[06:20:08] *** Gman has joined #opensolaris
[06:30:07] *** tsoome1 has quit IRC
[06:33:34] *** Volkodav has quit IRC
[06:34:26] *** jsubl2 has quit IRC
[06:35:15] *** uebayasi has quit IRC
[06:35:21] <jteo> wb Gman.
[06:37:00] <Gman> g'day jeremy
[06:40:42] *** jsubl2 has joined #opensolaris
[06:58:24] *** uebayasi has joined #opensolaris
[07:02:06] *** jsubl2 has quit IRC
[07:07:58] *** gm152 has quit IRC
[07:15:38] *** freakazoid0223 has joined #opensolaris
[07:35:30] *** bitvector2 has joined #opensolaris
[07:36:07] <jteo> http://photomatt.net/2007/01/18/relevant-sun/
[07:38:44] <delewis> jteo, oh, my eyes :-(
[07:39:33] <delewis> I guess offering an enterprise-grade operating system for free isn't appropriate for start-ups..
[07:39:52] <delewis> so, let's investigate what a start-up has to do to get a decent storage solution for Linux.
[07:40:05] <delewis> (a) grab RHEL (b) purchase some hardware (c) purchase a Netapp.
[07:42:28] <jteo> delewis: i think the complaint was more along the lines of how inept Sun's sales department can be.
[07:44:12] <delewis> jteo, and then somehow the commenters jumped from "sales department sucks -> Sun's products suck"
[07:44:29] <delewis> amazing leap of logic there.
[07:44:42] <richlowe> There's a huge difference between an OS vendor and a hardware Vendor.
[07:44:52] <richlowe> given he's talking of the startup program, which is mainly hardware, the OS seems irrelevent.
[07:45:04] <richlowe> SE is just as there if you're going to run linux on the hardware, is it not?
[07:45:11] *** pogma has quit IRC
[07:45:28] *** pogma has joined #opensolaris
[07:45:43] <jteo> delewis: i ignored the comments. ;)
[07:46:01] <delewis> SE? startup expenses?
[07:46:36] <richlowe> the startup essentials thing referenced in the article.
[07:47:02] <delewis> oh, that. Probably so, but I'm not familiar with the startup program.
[07:47:05] <Error_e^ipi> the one not in canada, just to annoy me
[07:51:15] <nprice> Anyone got any ideas... Got a raidz1 with three new 250gb disks, each disk gives good performance in standalone, but when added to the raidz1 the array gets 0-9mbyte/sec write speed writing zeroes to a file, cpu and ram usage are nominal, each drive is on its own physical controller and nothing is sharing interrupts... Any ideas?
[07:52:23] <jteo> nprice: zfs-discuss.
[07:52:32] <jbk> they need to hire some sales guys from BMC -- those guys can sell ice to eskimos
[07:52:36] <jbk> :)
[07:52:42] <jteo> jbk: BMC?
[07:52:52] <jbk> yeah, at least around here
[07:53:15] <jteo> jbk: what's BMC?
[07:53:24] <jbk> software company
[07:53:27] <richlowe> baseboard management controller?
[07:53:45] <jbk> they sell a lot of software that's not really needed :)
[07:53:56] <jbk> but still manage to get people to spend lots of money to buy it
[07:54:19] <jteo> jbk: sounds brilliant (from a business perspective)
[07:54:34] <jbk> i hate their install stuff, instead of making native os packages
[07:54:53] <jbk> they use overly elaborate shell scripts that break easily
[07:55:36] <jbk> my favorite was the automated install process was failing when it was doing a certain piece.. their fix was to simply not have it do that anymore
[07:56:12] <jbk> the problem is they have this heirarchy with all their scripts, and screwed up the relative path to the response file in some of them
[07:56:37] *** Gman has quit IRC
[08:02:14] <jteo> if Sun actually had a decent sales team, every university here would be running Sunrays.
[08:03:04] <delewis> jteo, I'm trying to get my university interestedin Sun Rays, but apparently they have their interests elsewhere at the moment.
[08:03:31] <delewis> we have a large population of Windows workstations (with no ADC, etc.) and the maintenance required is a nightmare.
[08:03:37] <delewis> not to mention we upgrade every 2-3 years.
[08:03:50] <delewis> and 99% of the workload is word processing, etc.
[08:03:58] <jteo> precisely.
[08:04:26] <Jiko> its the etc that kills you
[08:04:26] <jteo> upgrade costs are a utter waste of money.
[08:04:40] <Jiko> if its anything like our uni, there's mandatory software that only runs on Windows
[08:04:45] <delewis> Jiko, for that, rdesktop and a Windows terminal server.
[08:04:54] <Error_e^ipi> our school has whatever the windows equivalent to sunrays are
[08:04:59] <delewis> and at least one or two Windows systems in every lab
[08:05:09] <Error_e^ipi> just little thin clients
[08:05:12] <delewis> Error_e^ipi, Wyse Winterminals, etc.
[08:05:17] <Jiko> delewis: doesn't work for things like autocad
[08:05:24] <Error_e^ipi> actually, i think they use exactly those delewis
[08:05:31] <Jiko> not saying runrays aren't good, there's just situations that they won't work in
[08:05:40] <jteo> what irks me the most was: the uni i graduated from DIDN'T have any windows apps students had to run.
[08:05:42] <delewis> Jiko, not every lab would be replaced.
[08:05:51] <delewis> we have a graphics/autocad lab that would be untouched.
[08:06:09] <delewis> only faculty systems and most student systems would be converted to Sun Rays.
[08:06:14] <jbk> we had that for pspice initially, before we got a license for mentor graphics
[08:06:28] <Error_e^ipi> our graphics lab is where the Origin lives
[08:06:41] <delewis> heh, we don't have anything that large here.
[08:06:54] <delewis> just a very, very small university ~ 2,000 students.
[08:07:07] * jbk was one of the few students that knew unix, and could run mentor remotely :)
[08:07:27] <jbk> so i didn't have to go to the lab to do assignments
[08:07:46] <delewis> right now, though, we swap out about 300 Windows workstations over a 3 year period.
[08:07:47] <jbk> (which weren't open 24/7)
[08:07:50] <delewis> so you can imagine the costs.
[08:08:26] <jteo> where do those used workstations go?
[08:08:39] <delewis> drop some Sun Rays in there (which wouldn't have to be upgraded for years) and two servers that have a large physical memory capacity (E4500, etc.) and that would take care of it.
[08:08:50] <delewis> jteo, re-sold, I think.
[08:08:53] <delewis> or donated.
[08:09:36] <richlowe> university surplus is always fun.
[08:09:51] <delewis> oh, I know where they go -- faculty gets them.
[08:09:58] <delewis> (faculty always gets everything last)
[08:14:39] <jamesd__> faculity usually get all the good stuff on there desks first when it arrives..
[09:05:53] *** lplatypus has joined #opensolaris
[09:14:37] *** dunc has quit IRC
[09:45:02] *** alobbs has joined #opensolaris
[09:49:02] *** lloy0076 has joined #opensolaris
[09:49:23] *** mazon is now known as Mazon
[09:57:32] *** richlowe has quit IRC
[10:01:01] *** alobbs has quit IRC
[10:01:28] *** lloy0076 has quit IRC
[10:04:03] *** kloczek has quit IRC
[10:06:07] *** richlowe has joined #opensolaris
[10:10:06] *** alanc-away has quit IRC
[10:10:15] *** alanc_away has quit IRC
[10:10:36] *** alanc-away has joined #opensolaris
[10:10:37] *** alanc_away has joined #opensolaris
[10:21:04] *** deedaw has joined #opensolaris
[10:24:43] *** LordKing has joined #opensolaris
[10:25:48] *** lloy0076 has joined #opensolaris
[10:26:32] <lloy0076> I downloaded the ON 55 DVD, ran the installer as "root" and it seems to have worked. It called itself "Live Upgrade". However, my /etc/motd still says it's ON 54...
[10:26:46] <lloy0076> How do I actually know *what* I'm running?
[10:27:09] * lloy0076 looks at the logs
[10:29:21] *** alobbs has joined #opensolaris
[10:31:03] *** Mazon is now known as mazon
[10:34:00] <jteo> cat /etc/release
[10:34:49] * lloy0076 sigh
[10:35:03] <lloy0076> This is confusing me: I'm trying to follow the instructions given at: http://www.sun.com/bigadmin/features/articles/using_lu.html
[10:35:26] <lloy0076> It's asking me for a "device" to give the / partition for the new boot environment...
[10:36:00] <lloy0076> I obviously can't give it any random name - but it implies that I need to make this partition in the slice and put a file system in it...
[10:36:03] <lloy0076> I *think*.
[10:38:14] <lloy0076> ...
[10:38:29] <lloy0076> Ok, I've worked out what it wants me to do. I need to do some reading on "format" and "fdisk" before I proceed.
[10:40:23] *** uebayasi has quit IRC
[10:41:49] *** lplatypus has quit IRC
[10:42:59] *** Fish has joined #opensolaris
[10:43:46] <Fish> hello
[10:44:39] *** alobbs has quit IRC
[10:45:07] <lloy0076> Fish: Hi
[10:45:49] <richlowe> the installer is the liveupgrade installer.
[10:46:04] <richlowe> you'd need to either boot the image, and upgrade normally, or use liveupgrade, rather than just installing liveupgrade.
[10:46:14] <lloy0076> Ah, ok.
[10:46:42] <lloy0076> I tried to boot the thing normally and it *looked like* it wanted to remake the root partition (i.e. /)...
[10:48:01] <richlowe> it should have given you an option to upgrade.
[10:48:42] <lloy0076> I'll have another go at it in a few minutes.
[10:50:12] <lloy0076> BBS
[10:50:27] *** lloy0076 has quit IRC
[11:06:41] *** thomsog has joined #opensolaris
[11:21:04] <Berny> what a sad day
[11:21:53] <Berny> after nearly 9 years i removed my old timberwolf fromthe server room :-(
[11:23:22] *** LordKing has quit IRC
[11:24:08] <lasseoe> Berny: what's replacing it?
[11:27:29] <Berny> it's been replaced about a year ago by a sashimi+wasabi
[11:27:56] <Berny> it just took about a year to move all old tapes to the new library
[11:28:37] <lasseoe> it's been replaced by a sushi roll?
[11:28:38] <Berny> today is cleanup day...
[11:28:57] <Berny> ack
[11:29:04] <lasseoe> :)
[11:29:17] <Berny> sashimi a.k.a L100 :-) and wasabi being a L25
[11:29:44] <Berny> .oO(i love them code names :-))
[11:29:49] <lasseoe> ah ok, wasn't aware of the codenames :)
[11:30:01] <Berny> though i would have rather had a new stk library
[11:30:08] <lasseoe> of course.
[11:31:58] <Berny> though at the time we bought that pack it was moneywise the better choice (and money rules everywhere :-\)
[11:34:45] <Berny> hmm the L8500 weighing nearly 2tons wouldn't fit in our server room though...
[11:36:19] <quasi> L8500 is not cheap
[11:36:43] <Berny> that's the next hiccup ;-)
[11:37:16] <Berny> maybe a L500 then
[11:37:20] <quasi> starting at around $200k from what I recall
[11:37:55] <Berny> that wouldn't be too bad though
[11:38:30] <quasi> the L500 is pretty cheap compared to that - but once you need a capacity that grows to 2 racks of L500s, L8500 is probably going to be cheaper
[11:38:45] <Berny> when we got the timberwolf back in '98 that one was something like $60k plus updates for 5 drives in 2000 for another $50k
[11:39:36] <Berny> a L500 with 50 slots lto is a little under 40k
[11:40:12] <Berny> the "empty" L8500 is on the list with 85k
[11:40:23] <quasi> oh?
[11:40:54] <quasi> list prices or are you getting very special discounts?
[11:41:08] <Berny> list sun.de
[11:41:50] <Berny> no drives though
[11:41:55] <quasi> ah
[11:42:11] <quasi> but still, there's a bit of way to $200k
[11:42:39] <Berny> add 5 drives and you're close to 200k
[11:42:40] * quasi needs to "have words" with a certain supplier ;)
[11:43:11] <quasi> http://www.sun.com/storagetek/tape_storage/tape_libraries/sl8500/ says from $195k
[11:43:13] <Berny> prices i told you being euro of course
[11:43:56] <quasi> yeah, yeah
[11:45:35] <Berny> but mind you the german list sometimes offers stuff that is not operational without further components
[11:45:49] <Berny> so in total you may end up with about the same price
[11:46:46] <quasi> yeah, perhaps missing the drives and a few extra bits
[11:47:06] <Berny> pretty sure
[11:47:14] <Berny> drives go extra anyway
[11:47:41] <Berny> sometimes they leave out unimportant things like power supplies and co
[11:48:38] <Berny> but even with the usual 30% edu discount that's still dreaming for me...
[11:51:49] <quasi> lot of room for expansion though - starting at 1448 tapes and expanding up to 300000
[11:52:34] <Berny> yeah that should do for the time i work here 8-)
[11:53:17] <Berny> the collection of the past 9 years is "just" 14TB
[11:54:16] <Berny> right time to grab some lunch
[11:54:33] <quasi> 300000*400G tapes - that should be enough for the mp3 collection ;)
[11:54:41] <Berny> real cleaning starts tonight...
[11:55:31] <Berny> take the server down, replace a broken raid, move some racks... workout or the next month should be done when i finish :-)
[11:56:58] <Berny> btw has someone tried to read stuff from disk that belonged to a raid5 (where the raid controller died and no replacement is available)?
[11:57:42] <Berny> in theory shouldn't it be possible to pull the stuff off the drives and emulate the raidcontroller in software?
[11:58:22] <lasseoe> hm
[11:59:05] <lasseoe> how he heck do I map a cXtXdXsX LUN to a WWN when using lpfc drivers?
[11:59:14] <lasseoe> *the
[12:01:15] <PerterB> dmesg? (from memory the lpfc spews out a whole load of messages as it binds to each WWN)
[12:01:32] <PerterB> s/lpfc/lpfc driver/
[12:01:38] <lasseoe> that's long gone :)
[12:02:17] <quasi> so you just reboot the box ;)
[12:02:37] <PerterB> there's also an lputil or something (used to upgrade firmware etc), does that have an option to list LUNs?
[12:02:42] <lasseoe> heh oh of course, I'll ask Emulex to put that in the docs :)
[12:02:53] <lasseoe> PerterB: not really, already checked
[12:03:07] <lasseoe> I can see which WWNN's it binds to, but not which device it binds to
[12:03:13] <quasi> lasseoe: ah, emulex ... they have a couple of tools with the driver iirc
[12:03:23] <lasseoe> yes, lputil being useless :)
[12:03:34] <quasi> isn't there one more?
[12:03:41] <lasseoe> hbacmd
[12:03:45] <PerterB> dunno then... the boxes I used to admin with lpfc also had powerpath, so we could use that
[12:03:50] <lasseoe> but that's no better
[12:04:10] <lasseoe> PerterB: thanks though.
[12:04:23] <Berny> pull the plugs, re-insert, cfgadm, dmesg?
[12:04:33] <lasseoe> Berny: not really an option
[12:04:40] <Berny> .oO(just being naive..)
[12:04:54] <Berny> wimp ;-)
[12:04:57] <lasseoe> :-)
[12:04:57] <Berny> no risk no fun
[12:05:04] <lasseoe> no pain no gain
[12:05:09] <PerterB> and luxadm doesn't work at all? (remember seeing you say that before)
[12:05:10] <Berny> ack
[12:05:31] <lasseoe> PerterB: correct, the lpfc driver doesn't support that :(
[12:05:52] <PerterB> that sucks
[12:06:25] <PerterB> as does the lpfc driver, but I'm assuming a short-term switch to emlxs is also out of the question
[12:06:26] <lasseoe> PerterB: yes, some of them are scheduled to be upgraded, in that process I'll install the newer emlx driver
[12:06:43] <PerterB> ah right
[12:07:01] <lasseoe> unfortunately I sort of need to find out soon :)
[12:07:03] <Berny> lunchtime for me... if someone has any idea about my broken raid controller query me ;-)
[12:07:11] <PerterB> ask again when jmcp is about, he knows this stuff inside out
[12:07:38] <lasseoe> PerterB: been waiting, but I've not seen him about the past days
[12:07:47] <PerterB> me either
[12:08:00] <lasseoe> The man's always around, then when ya need him, he's not ;-)
[12:08:24] <Berny> laters
[12:08:32] <lasseoe> enjoy your lunch
[12:08:47] *** tsoome has joined #opensolaris
[12:16:03] *** LordKing has joined #opensolaris
[12:23:28] *** LordKing has quit IRC
[12:31:43] *** nostoi has joined #opensolaris
[12:33:17] <Doc> http://img.thedailywtf.com/images/200701/pup2/addedsql.jpg
[12:33:28] *** tsoome1 has joined #opensolaris
[12:33:46] <lasseoe> haha
[12:35:40] *** Odin-LAP has joined #opensolaris
[12:43:33] *** tsoome2 has joined #opensolaris
[12:44:04] <PerterB> in the same vein: http://bingweb.binghamton.edu/~tony/cake.jpg
[12:44:19] <PerterB> (linked from http://community.livejournal.com/good_eats/507059.html)
[12:45:52] <lasseoe> :-)
[12:46:57] *** LordKing has joined #opensolaris
[12:47:34] *** tsoome has quit IRC
[12:54:22] *** Gman has joined #opensolaris
[12:59:10] <jmcp> gdvnng vrybdy
[12:59:32] *** tsoome1 has quit IRC
[13:01:31] *** nprice has quit IRC
[13:02:54] <lasseoe> jmcp: sober? :)
[13:02:56] *** lloy0076 has joined #opensolaris
[13:03:04] <lloy0076> So much for a short time.
[13:03:09] <lloy0076> Why does upgrading take *so long*?
[13:03:09] <jmcp> lasseoe: of course
[13:03:16] <jmcp> lasseoe: the party was *last night*:)
[13:03:30] <jmcp> lloy0076:  because the pkg db has to be interpreted
[13:03:37] <lasseoe> jmcp: so you're saying you didn't  have enough to drink then ;-)
[13:03:50] <jmcp> lloy0076: and let's not get into a discussion about wtf the packaging system is still broken after all these years
[13:03:59] <jmcp> lasseoe: nono ... I had plenty. but I've recovered
[13:04:18] <lasseoe> jmcp: good good
[13:04:21] <lloy0076> jmcp: I wasn't planning to. But a 4 hour upgrade for a DESKTOP system is worse than Windows.
[13:04:23] * lloy0076 sigh
[13:04:34] <jmcp> lloy0076: meh ... that seems about standard to me
[13:04:53] <lloy0076> jmcp: It would have taken me an hour to get a full Debian Stable to Unstable...
[13:05:02] <lloy0076> jmcp: I like Solaris, but that just sucked.
[13:05:24] <jmcp> lloy0076: like I said ... (11:05:01 PM) jmcp: lloy0076: and let's not get into a discussion about wtf the packaging system is still broken after all these years
[13:05:33] <lloy0076> heh
[13:05:42] <lloy0076> I need to get another monitor and keyboard and mouse...
[13:05:51] <lloy0076> So then I can watch pr0n while Solaris takes its time upgrading :P
[13:05:54] * lloy0076 rofl
[13:06:19] <lloy0076> At least gaim hasn't crashed yet after my upgrade.
[13:06:28] <lloy0076> It has periods of just randomly crapping out because of IRC.
[13:06:49] <jmcp> I'm expecting to be able to upgrade to 56 this week
[13:07:09] <lloy0076> Kewl
[13:08:01] <lloy0076> I've been looking into Java Server Faces and been saying to myself: OMG, why didn't I find out about this before?
[13:08:02] <lloy0076> heh
[13:08:24] <jmcp> JSF is pretty cool, that's for sure
[13:08:55] <lloy0076> I've wanted something like "Swing on HTML" and "JSF" seems to answer most of my complaints about the tediousness of writing web apps.
[13:10:16] <lloy0076> Interestingly, the thing I've found about Java and Solaris is that without the necessary resources (hardware) they can really run like crap, but give them the resources and they railroad other technologies.
[13:10:31] <jmcp> that's true of most OSen and apps, in my experience
[13:10:48] <lloy0076> hmmm
[13:11:20] <lloy0076> I can run the same set of apps (well same name so to speak) on Solaris and Debian GNU/Linux...
[13:11:37] <lloy0076> I can get reasonable performance from Debian with about 1-1.2 GB of RAM.
[13:11:40] <thomsog> hello there, if your into JSF you might want to try sun java studio creator
[13:11:46] <lloy0076> Solaris will crawl at that rate.
[13:12:05] <lloy0076> But if I throw 2GB at the same system, Solaris runs it better.
[13:12:06] <delewis> with 1GB of memory? I wouldn't call it "crawling"
[13:12:07] *** a_pavlenko has joined #opensolaris
[13:12:10] <delewis> constraining, yes.
[13:12:21] <jmcp> lloy0076: are you doing a true apples-to-apples comparison? same jvm rev on both?
[13:12:27] <lloy0076> delewis: Solaris will go swap at about that rate. For some reason, Linux doesn't.
[13:12:46] <lloy0076> jmcp: Same machine, JVMs downloaded from java.sun.com at the same time.
[13:13:15] <trygvis> I've got to say I have the same experience as lloy0076
[13:13:24] *** jmc1 has joined #opensolaris
[13:13:41] *** jmc1 has quit IRC
[13:13:44] <lloy0076> From what I can tell, Solaris *knows* what to do with extra memory, whereas Linux basically twiddles its fingers.
[13:13:55] <lloy0076> It seems to *know* what to do with more than one cpu too.
[13:14:04] <delewis> it should.
[13:14:13] <Doc> that's because it's an enterprise OS, rather than a toy OS
[13:14:16] <delewis> otherwise E25K customers wouldn't be very happy with their purchase.
[13:14:19] <lloy0076> Indeed.
[13:14:41] <delewis> this is the sort of thing that happens when your standards are set by real hardware, and not consumer peecees.
[13:14:53] <lloy0076> I'm using a consumer PC :P
[13:14:58] <delewis> though, I did learn Linux seems to have MPO and large page sizes, nowadays.
[13:14:59] * trygvis too
[13:15:00] *** lasseoe has quit IRC
[13:15:01] *** lasseoe has joined #opensolaris
[13:15:08] <Doc> MPO? on what hardware?
[13:15:36] <delewis> Doc, from what I read -- the IBM stuff (x440, and the pSeries gear)
[13:16:09] <jmcp> delewis: that doesn't seem like common-or-garden PC hw to me
[13:16:12] <delewis> http://lse.sourceforge.net/numa/
[13:16:28] <delewis> jmcp, definitely not. :-)
[13:17:05] <delewis> you can apparently do MPSS, nowadays, too.
[13:18:06] <delewis> though, its a bit more constraining than what Solaris does. You have to allocate a "pool" of pages with a certain size upon boot, which is not resizable.
[13:19:03] *** Gman has quit IRC
[13:21:54] <lloy0076> Cra[
[13:21:57] <lloy0076> Crap
[13:22:02] <lloy0076> Eclipse still craps out on 0N55
[13:22:03] * lloy0076 sigh
[13:22:28] <trygvis> eclipse sucks .. yay intellij idea
[13:22:31] <delewis> is it because of the fact that SXCR ships JSE 1.6 nowadays?
[13:22:42] <lloy0076> delewis: It causes the JVM to fall out underneath it.
[13:22:44] <delewis> (which breaks a lot of Java-based installers, including Sun Studio)
[13:23:19] <lloy0076> It's rather annoying.
[13:23:33] <lloy0076> I have a Brand LX Centos 3.8 zone just for Eclipse.
[13:23:36] * lloy0076 sigh
[13:27:05] *** bunker has joined #opensolaris
[13:27:24] <lloy0076> http://lse.sourceforge.net/numa/status/ seems messy.
[13:28:20] <lloy0076> Whilst I know Solaris Internals skips a LOT of the detail, how can Linux be at the stage it is when there's documentation from other operating sytems that have already done it...
[13:28:27] <lloy0076> Maybe I'm being unfiar.
[13:29:30] <Doc> why do hyuandi continue to make cars that aren't ferraris?
[13:29:50] <jmcp> lloy0076: I don't understand your question
[13:30:01] <Doc> easy enough to pull apart a ferrari and see how it works, and just copy it
[13:30:33] <dlg> lloy0076: i dont understand what you're trying to ask either
[13:30:44] <lloy0076> I'm not really asking anything.
[13:32:32] <jmcp> hm.... maybe I should wait for snv_57 rather than 56
[13:33:14] <lasseoe> jmcp: any idea how I can display cXtXdXsX to WWN mappings when using an older lpfc driver ?
[13:33:25] <jmcp> lasseoe: not really
[13:33:33] <lasseoe> doh :)
[13:33:42] <lasseoe> you were my last hope :)
[13:33:43] <jmcp> well ... not with a Solaris util. You'd need something from emulex '
[13:34:02] <dlg> isnt there a variant of an inquiry that lets you ask the disk what it thinks it is?
[13:34:07] <lasseoe> I've tried Emulex' lputil and hbacmd, but they can't do what I want
[13:34:12] <jmcp> yes, you can write some code to do it
[13:34:50] <jmcp> lasseoe: if you do an ls -l /dev/dsk/cXtYdZs2 of the device you're interested in, does that give you a path which has a wwn in it?
[13:34:51] <lasseoe> I've heard EMC has an 'inq' tool, but I can't find the download page anywhere
[13:35:06] *** SunTzuTech has quit IRC
[13:35:06] <lasseoe> jmcp: nope, if only it did!
[13:35:14] <lloy0076> I'm kind of thinking that Linux is a good OS compared to Windows, however now that there's at least one Real Unix (tm) that is open source with many more years of experience in development and governance than Linux, is Linux's strength really its market position (much like Microsoft's strength is its market position - if its customers actually were told: "This new OS you're buying has a defaul e-mail program which allows you're machin
[13:35:15] <jmcp> I've got a little inquiry tool that I'm trying to get released under CDDL
[13:35:18] <lasseoe> and lpfc doesn't support luxadm
[13:35:25] <jmcp> lasseoe: it's the other way around
[13:35:30] <lasseoe> ah ok :)
[13:35:34] <jmcp> luxadm doesn't support non-leadville drivers
[13:35:43] <lasseoe> makes sense.
[13:36:14] <sickness> # uname -a
[13:36:15] <sickness> SunOS vintage 5.11 snv_55b i86pc i386 i86pc
[13:36:21] <sickness> why that "b" ?!?
[13:36:26] <jmcp> respin
[13:36:31] <sickness> oh, so is it ok? :)
[13:36:38] <jmcp> I believe so
[13:36:40] <sickness> I just downloaded the iso and liveupgraded...
[13:36:41] <sickness> tnx :)
[13:37:04] <jmcp> I installed snv_55b yesterday, it appears fine
[13:37:05] <lloy0076> I've got the 'b' too. List was saying it was respun twice. A...B.
[13:37:10] <lasseoe> hm maybe I can hack up my own inq tool.. not done anything with C in a decade
[13:37:20] <sickness> cool
[13:37:55] <Doc> 55 will be released under more names than you can possibly imagine, so what's an extra "b" matter...
[13:38:52] <lloy0076> Incidentally, the ISO I downloaded has a GRUB that decided that the Solaris Developer Release (or some similar name) was the default thing to install.
[13:39:14] <lloy0076> This doesn't appear to be the same as what I have...(which is Solaris Express Community Edition).
[13:39:27] <lloy0076> So hence, it wanted to WIPE OUT my system as opposed to upgrade it.
[13:39:28] <Doc> like i said...
[13:39:33] <jmcp> lloy0076: the developer release ... yes
[13:39:35] <lloy0076> Rather baffling at first.
[13:40:14] <sickness> yeah, it also rebooted in the keyboard selection thingie to me :/
[13:40:19] <sickness> that was annoying :)
[13:40:38] *** triplah_ has quit IRC
[13:40:47] <sickness> I was trying to ssh in and it didn't worked out until I gone to the rack and gave him the damn US setting ghgh :P
[13:41:26] <lloy0076> sickness: I take it that you're first language isn't English :P ?
[13:41:37] *** gm152 has joined #opensolaris
[13:41:39] <sickness> lloy0076: yeah =)
[13:42:11] <lloy0076> Sorry, it's just that in English, "he", "him" always refer to human beings or sentient animals.
[13:42:24] <sickness> eheh, I use that for machines too :P
[13:42:26] <lloy0076> So when you said "him" it made me think your Solaris boxen were robots.
[13:42:29] <lloy0076> :P
[13:42:40] *** Arnald has joined #opensolaris
[13:43:03] <lloy0076> "I just installed ON 66 and now my Solaris based maid is hitting on all my gerbils!"
[13:47:50] * Doc puts "President of Norway" on his resume
[13:48:55] <Doc> although if anyone checks i could be in trouble, because norway apparently has a prime minister, not a president, so they will know i'm lying... hrmmm
[13:49:49] <andersmo> And there are Norwegians here, watching you right now. Step carefully. ;)
[13:50:08] *** SunTzuTech has joined #opensolaris
[13:52:16] <lloy0076> Anyway, off to play Diablo II
[13:52:17] *** lloy0076 has quit IRC
[13:53:34] * steleman is getting annoyed with Joerg
[13:53:46] <dlg> schilling?
[13:53:55] <steleman> precisely :-)
[13:54:14] <dlg> c'est la vie
[13:54:25] <steleman> La vie en rose.
[13:54:46] <jteo> et tu, brutu?
[13:54:52] <steleman> :-)
[13:55:08] <Doc> joerg is still alive?
[13:55:24] <steleman> he is alive, kicking, and babbling about C++ templates
[13:55:25] <Doc> i would have thought someone would have had him put out of our misery by now
[13:55:49] <steleman> C++ Template ABI to be exact
[13:56:55] <steleman> (which makes no sense)
[13:57:29] *** LordKing has quit IRC
[13:58:36] <Doc> right, but you need to remember one very important fact
[13:58:39] <Doc> he's an idiot
[13:59:25] <dlg> steleman: doc is right
[13:59:33] <dlg> so take a deep breath
[13:59:34] <dlg> and move on
[14:07:53] *** Qapf has quit IRC
[14:08:26] <delewis> oh, joy. The OpenSolaris Developers conference this year looks fun. They're going to put Martin, Joerg, and Roland in the same room together.
[14:08:49] <delewis> http://www.guug.de/veranstaltungen/osdevcon2007/abstracts.html#1_2_2
[14:09:18] *** FastJack has joined #opensolaris
[14:09:52] *** FastJack has left #opensolaris
[14:10:09] <jmcp> delewis: take a video camera, flameproof suit, and popcorn
[14:11:08] <steleman> Doc: i have trouble reconciling that with the fact that he writes software which works :-)
[14:12:49] <steleman> delewis: I WANT THE VIDEO :-)
[14:12:56] <steleman> on YouTube
[14:13:01] <delewis> steleman, yeah, me too :-)
[14:13:34] <delewis> though, Joerg at a Debian developers conference might be a bit more entertaining.
[14:13:54] <steleman> Terminator 4 ?
[14:16:08] <Doc> steleman: he also released the first non-Sun release of OpenSolaris
[14:16:32] <Doc> and in the process he _deliberately_ broke backwards compatibility with Solaris
[14:17:11] <Doc> so perhaps he's a smart idiot, but he's still an idiot
[14:30:35] *** LordKing has joined #opensolaris
[14:34:49] <steleman> Doc: lol
[14:35:06] <steleman> Doc: but Sun's OpenSolaris was not POSIX compliant!!
[14:35:28] <steleman> Zat is a mistakke vich musst be korrekted
[14:36:18] * steleman has espresso + cranberry orange scone for breakfast
[14:39:47] *** Fish- has joined #opensolaris
[14:45:48] <jteo> steleman: not to mention he has already solved <insert problem here> 20 years ago.
[14:45:59] <steleman> since 20 years ago
[14:46:00] <jmcp> jteo: .. since 20 years ago
[14:46:05] <jteo> oops.
[14:46:34] * steleman will be a good boy and will not reply to joerg's latest ramble
[14:48:32] <quasi> steleman: did someone post a "don't feed the trolls" sign?
[14:48:53] <steleman> quasi: i thought it was posted implicitly :-)
[14:54:27] * steleman imagines a wonderful world where i could watch tv on my solaris boxes
[14:55:49] *** Fish has quit IRC
[14:55:52] <quasi> steleman: don't hold your breath ;)
[14:56:14] <steleman> yeah i know
[14:56:28] <steleman> although haupaugge has linux drivers
[14:56:35] <steleman> which means they are willing to release specs
[14:56:44] <jteo> that's not always true. -sigh-
[14:57:11] *** klocze2 has joined #opensolaris
[14:57:52] <quasi> most of the linux drivers for tv cards are evil hacks (and that's putting it politely)
[14:58:32] <steleman> quasi: yah but we wouldnt be able to use linux drivers anyway (and not necessarily because of the license)
[14:58:43] <steleman> it's the willingness to release specs that's usually the problem
[14:59:18] <quasi> and the whole driver infrastructure for video grabber cards
[14:59:58] *** gnu2it2 has joined #opensolaris
[15:00:56] * jmcp hits the hay ... it's 0100 here
[15:00:57] <jmcp> gnite all
[15:01:51] <gnu2it2> what is the command to get solaris to download and install updates?
[15:02:02] <quasi> smpatch update
[15:02:13] <gnu2it2> thanks
[15:03:17] *** gallium has joined #opensolaris
[15:04:15] *** hspaans has quit IRC
[15:06:04] *** hspaan1 has joined #opensolaris
[15:06:38] *** hspaan1 is now known as hspaans
[15:11:37] *** tekops has joined #opensolaris
[15:11:49] <tekops> I'm confused and need slight clarification
[15:12:31] <tekops> if opensolaris builds have been tested, and sun offers support for it (even in enterprise settings), then what is the benefit to using solaris 10?
[15:13:12] <quasi> since when do they offer support for opensolaris builds?
[15:13:27] <tekops> hold on, let me find the page
[15:13:29] <jteo> tekops: oracle will ask you to bugger off if you go to them for support for running oracle on anything else other than certified solaris versions.
[15:13:40] <jteo> replace oracle with your favourite vendor.
[15:13:50] <Stric> microsoft? :)
[15:14:34] <tekops> jteo, as in oracle (or any other vender) will say use the certified release? or are you saying replace oracle with another and i won't have the issue
[15:14:52] <Stric> no, other vendors will say the same
[15:15:05] <tekops> i understand now
[15:15:25] <jteo> yes.
[15:15:42] <jteo> and major vendors are very very picky about what is supported and what isn't.
[15:16:04] <Stric> security fixes etc come to official releases
[15:16:11] <jteo> when S10 was released, the most common question i heard from potential customers was: is Oracle10g certified? when?
[15:17:21] <tekops> stric are you saying security patches aren't plugged into open solaris as well?  it would make sense to me to just test them in opensolaris before putting them in sol 10
[15:19:11] *** mrdeviant has joined #opensolaris
[15:19:42] <Stric> tekops: they are.. but it's up to you to get them onto your machine vs getting an official patch that you just install
[15:20:12] <tekops> That makes perfect sense, thank you
[15:20:50] <Stric> opensolaris isn't the perfect "stable system" exactly.. it's a moving target
[15:22:44] *** uebayasi has joined #opensolaris
[15:22:53] <tekops> well, is there a safe way to utilize the advances in ZFS seen in OpenSolaris, while using Solaris 10?
[15:23:12] <jteo> tekops: no.
[15:23:18] <jteo> wait for U4.
[15:23:56] <gallium> when is U4?
[15:25:11] <quasi> if they keep the current pace, then sometime around june or july probably
[15:26:57] <tekops> I'm confused
[15:27:04] <tekops> is U4 the next release of solaris 10?
[15:27:24] <Stric> update 4, yes
[15:28:04] <gallium> update X is the internal name. the release name is more like Solaris 11/06 (a.k.a. Update 3)
[15:28:15] <tekops> it will include all the ZFS updates seen in Open Solaris, i.e. the Raid 6 stuff
[15:28:27] <Stric> that's already in update3
[15:29:15] <jteo> U3 includes the all important fsync fix.
[15:31:07] *** LordKing has quit IRC
[15:32:14] <tekops> is there a changelog publically available for Solaris 10?
[15:35:37] *** nachox has joined #opensolaris
[15:35:58] *** rmorse has joined #opensolaris
[15:36:22] *** rmorse has quit IRC
[15:36:31] *** rmorse has joined #opensolaris
[15:38:19] <kimc> trying to add a whole root zone with contents of /usr/local.. can I get it with 'add inherit-pkg-dir' ?
[15:38:48] <kimc> thats the contents of /usr/local in the global zone..
[15:40:00] <kimc> or use 'add fs' -?
[15:40:03] <quasi> or you could just add it later as a regular
[15:40:25] <kimc> how would i add it later ?
[15:40:39] <quasi> add fs works, although add inherit-pkg-dir probably does as well
[15:40:46] <quasi> zonecfg ...
[15:41:33] *** tekops has quit IRC
[15:41:57] <kimc> how about changes made in /var such as directories -is there a way to recreate /var as it was on the global zone ?
[15:43:35] <kimc> also, what does set special=/something do ?
[15:45:50] <kimc> how about
[15:45:53] <kimc> add fs
[15:46:02] <kimc> set dir=/usr/local
[15:46:07] <kimc> set dir=/var
[15:46:11] <kimc> ?
[15:46:56] *** Fish- has quit IRC
[15:47:01] *** Fish has joined #opensolaris
[15:48:11] <quasi> nope, you need set dir and set special, not duplicate set dir
[15:48:36] <kimc> set dir=/usr/local
[15:48:47] <kimc> set special=/usr/local
[15:48:58] <quasi> bingo
[15:49:21] <kimc> then again set dir=/var;set special=/var
[15:49:51] <kimc> does set dir 'clone' the global zone's dir ?
[15:50:06] <jteo> it loopback mounts it.
[15:50:10] <kimc> ahh
[15:50:34] <kimc> is there a way to duplicate the global zone's /usr/local ?
[15:51:06] <kimc> if its a whole root zone
[15:55:39] <kimc> add inherit-pkg-dir
[15:55:52] <kimc> set dir=/usr/local ?
[16:06:26] *** nachox has quit IRC
[16:06:31] *** Arnald has quit IRC
[16:14:41] *** shalini has joined #opensolaris
[16:15:49] <shalini> where can i get the OpenSolaris admin help
[16:16:47] *** shalini has left #opensolaris
[16:24:03] <quasi> kimc: duplicate as in copy? and no, you don't want loop of /var
[16:26:57] <kimc> quasi: yes i meant copy.. understand the no loopback mount of /var ..soo how about a copy of /var from the global zone ?
[16:27:22] <quasi> cp -rp
[16:28:09] <kimc> ok.. so no way to set it up to be a copy when creating the whole root zone ?
[16:28:31] <quasi> or whatever else takes your fancy - but not necessarily a good idea to because of /var/sadm
[16:28:46] <quasi> sure, just don't touch it at all
[16:28:46] *** mrdeviant has left #opensolaris
[16:28:55] <quasi> then you'll have a copy
[16:29:39] <kimc> one reason is i have installed smf methods which are not 'standard' with solaris and i want them to be accessable from the 'virtual host' in the new zone
[16:29:41] *** gallium has quit IRC
[16:30:02] <Tekni> if you have zfs hot spares defined and start getting read/csum errors on a disk, shouldn't a hot spare kick in immediately?
[16:34:18] <quasi> there's probably a threshold after which it will take over
[16:35:20] <Tekni> c5t600D0230006C737E0C4F700E250D2700d0  ONLINE   1.78K   923     0
[16:35:31] <Tekni> this disk is currently not in the disk array :)
[16:35:40] *** tsoome has joined #opensolaris
[16:38:51] *** hiran3005 has joined #opensolaris
[16:39:08] <hiran3005> hi
[16:39:24] <hiran3005> i recently placed for an order
[16:39:42] <hiran3005> for the solaris 10 dvd & media kit
[16:40:07] <hiran3005> my pc is a pIII 800mhz old machine
[16:40:16] <hiran3005> will solaris run smoothly on it
[16:40:25] <hiran3005> id like to try this out
[16:40:32] <hiran3005> will it be possible
[16:40:33] <tsoome> try and you will see
[16:40:44] <hiran3005> ok
[16:40:58] <hiran3005> does it support old hardware!!!
[16:41:06] <tsoome> snoothly is quite "flexible" term;)
[16:41:13] *** uebayasi has quit IRC
[16:41:27] <hiran3005> mm
[16:41:58] <hiran3005> is it possibe to compile & run linux apps
[16:42:03] <hiran3005> ex xine??
[16:42:35] <quasi> hiran3005: http://www.sun.com/bigadmin/hcl/ - (and the Device Detection Tool) is a good place to start
[16:42:50] <hiran3005> cool thanks
[16:42:55] <hiran3005> ill try it now
[16:43:06] <hiran3005> cant wait till i get my hands on the dvd
[16:43:24] <lasseoe> you could just download it
[16:43:39] <jteo> thanks to the massive marketing blitz, we now have a whole generation of people who expect Solaris to replace Linux and Windows. gotta love marketing.
[16:43:48] <hiran3005> well im from sri lanka
[16:43:54] <jteo> after all, Solaris 10 is the "world's most advanced operating system"
[16:44:07] <hiran3005> ya i went throuh the demos
[16:44:14] <hiran3005> and presentations
[16:44:30] <quasi> jteo: they should have added server to that
[16:44:33] <hiran3005> iwann have zfs installed on my small hard drive
[16:45:10] <hiran3005> heard its dam good
[16:45:10] <quasi> but of course, then it wouldn't be marketing ;)
[16:45:14] <jteo> quasi: you give marketing too much credit.
[16:45:38] <hiran3005> solaris will take time to hit the linux mkt
[16:45:53] <hiran3005> it needs popularity
[16:46:46] <hiran3005> zfs seems to be very complex
[16:47:07] <hiran3005> i watched the presentation by the senior engineer
[16:47:11] <hiran3005> it was cool
[16:48:31] <andersmo> hiran3005: http://blogs.sun.com/eschrock/entry/ufs_svm_vs_zfs_code - regarding zfs complexity
[16:49:45] *** hiran3005 has quit IRC
[16:50:46] <jteo> ah. PJIRC explains how so many folks find their way in here.
[16:51:10] <tsoome> http://blogs.sun.com/eschrock/entry/ufs_svm_vs_zfs_code
[16:51:18] <jteo> hence throwing my argument that "if you know about IRC in this era, you know how to actually google" out of the window
[16:51:26] <tsoome> ups, sorry:)
[16:51:39] <quasi> heh
[16:54:39] <quasi> I have the same opinion of web irc clients as of web interfaces to mailinglists - they should be banned because people who are too daft for the real thing is not really the kind of people you want making noise
[16:55:07] <jteo> web interfaces to mailing lists are good for only 1 thing: archives, and drive-by participation
[16:55:11] <jteo> uh. 2 things
[16:55:12] <jteo> :(
[16:55:50] <quasi> the drive by bit seems to be spamming more often than anything else
[16:56:41] <Odin-LAP> quasi: There are occasions where knowledgeable users use them to bypass locks on IRC...
[16:56:43] <lasseoe> depends on the nature of the mailing list
[16:57:06] <jteo> and i really need to get a dvd drive so i can enjoy not having to insert multiple discs during upgrade. (at least before Solaris requires > 1 dvd)
[16:57:27] <lasseoe> jteo: jumpstart ?
[16:57:32] <quasi> jteo: with sfw it already takes two
[16:57:46] <jteo> lasseoe: i have but 1 boxen.
[16:57:51] <lasseoe> jteo: fair enough :)
[16:58:21] <jteo> lasseoe: i live in remote island off the equator where ebay is uneconomical.
[16:58:51] <quasi> jteo: if the postage wasn't going to kill it, you could have had one of my spare dvd drives
[16:59:01] <lasseoe> jteo: just noticed your location - although, with all the big business there, surely there must be a market for used hardware
[16:59:06] <lasseoe> liquidation companies?
[16:59:20] <Kmays> {there's no 128-bit integer type in standard C language), and globally optimal I/O sorting and aggregation.}
[16:59:41] <jteo> lasseoe: i'm a penny pincher.
[16:59:42] <jteo> ;)
[16:59:44] <lasseoe> hehe
[16:59:55] <lasseoe> nothing wrong with that
[17:00:10] <jteo> Kmays: it glosses over all the nice performance bugs folks are still wringing out of ZFS.
[17:03:26] <jteo> lasseoe: i've progressed to the unfortunate stage in life where i can no longer convince myself to live on ramen just so i can have more boxen. it is a sad thing, this thing called growing up,
[17:03:36] *** tsoome2 has quit IRC
[17:05:06] <lasseoe> jteo: haha - moved on to takeaway ? ;-)
[17:06:47] <Stric> "grown up is when you do the dishes after you eat instead of before" ;)
[17:14:57] *** mazon is now known as Mazon
[17:15:19] *** dunc has joined #opensolaris
[17:17:56] <Kmays> Stric <- ha!
[17:20:12] <Kmays> "growning up is when the pizza delivery companies in your area no longer call you by your 'Internet' name...'
[17:20:58] <lasseoe> and when they stop sending christmas cards
[17:26:32] <jteo> oh fsck i was wrong. there is sparc hardware on the market here.
[17:27:55] <lasseoe> E10k's :)
[17:29:01] <jteo> enterprise 3500s. -sigh-
[17:29:52] *** tsoome1 has joined #opensolaris
[17:30:39] <Kmays> http://www.opensolaris.org/os/community/on/flag-days/pages/2006102901/
[17:33:05] *** rmorse has quit IRC
[17:34:56] *** broadcast has joined #opensolaris
[17:35:40] <Kmays> For a fun read: http://blogs.sun.com/tpenta/ (The attack of Stevel's beagles in peature butter)
[17:35:51] <Kmays> peanut butter
[17:38:18] *** LordKing has joined #opensolaris
[17:49:01] <BadKarma> CON error sighted on the hroizon!
[17:49:35] <BadKarma> arg! skip my jibbers break on QSYSOPER SEV 25 arg!
[17:51:45] *** estibi has joined #opensolaris
[17:54:45] *** tsoome has quit IRC
[17:57:18] *** tsoome1 is now known as tsoome
[18:07:53] *** deather_ has joined #opensolaris
[18:11:39] *** Kmays has quit IRC
[18:16:30] *** deather has quit IRC
[18:22:49] *** adlpaf has joined #opensolaris
[18:34:22] * tsoome is away - bbl - [AcidJazz - Log ON]
[18:39:29] <mv> maybe dumb question, but ... why there is no on-src-DATE.tar.bz2 in http://dlc.sun.com/osol/on/downloads/current/ ?
[18:42:49] <Triskelios> mv: see the hg link at the beginning of the page
[18:45:36] *** coffman has joined #opensolaris
[18:46:01] <mv> Triskelios: thx, i'm blind or something..
[18:48:26] <Triskelios> hm, why is the current SXCR release labelled "b55b" this time?
[18:48:48] <oxygene> second respin
[18:49:25] <Triskelios> what happened to the first one?
[18:50:32] *** LordKing has quit IRC
[18:55:32] *** Theoden-Nexenta has joined #opensolaris
[18:56:42] *** kim0 has joined #opensolaris
[18:56:52] *** kim0 has left #opensolaris
[19:10:51] *** rodrickbrow_n has joined #opensolaris
[19:22:56] *** CosmicDJ has joined #OpenSolaris
[19:35:35] *** blkad has joined #opensolaris
[19:42:15] *** drio has quit IRC
[19:45:31] *** Kmays has joined #opensolaris
[19:46:40] *** Kmays has quit IRC
[19:47:44] *** switch has quit IRC
[19:47:59] *** switch has joined #opensolaris
[19:52:08] *** swa_mobil has quit IRC
[19:59:49] *** swa_mobil has joined #opensolaris
[20:08:55] *** schily has joined #opensolaris
[20:16:03] *** nrubsig has joined #opensolaris
[20:16:03] *** ChanServ sets mode: +o nrubsig
[20:16:38] <nrubsig> Tpenta: ping!
[20:16:57] <nrubsig> Tpenta: Could you check the triage queue whether my new bug arrived ?
[20:18:13] <Error_e^ipi> hey nrubsig
[20:19:19] <nrubsig> Error_e^ipi: Hi!
[20:19:27] <nrubsig> Error_e^ipi: waiting for the list getting set up...
[20:19:32] <Error_e^ipi> *nod*
[20:19:53] <Error_e^ipi> that's cool that it was accepted without real issue
[20:20:01] <nrubsig> umpf
[20:20:08] <Error_e^ipi> big bitch-fest, but the project got set up without any problems
[20:20:12] <nrubsig> Error_e^ipi: the real prolbem is to get it running.
[20:20:45] <steleman> and the first step towards getting it running would be to stop bickering about C++ on public mailing lists because it achieves nothing ;-)
[20:20:45] <Error_e^ipi> steleman's working on that probably as we speak
[20:20:53] <Error_e^ipi> oh, there you are...
[20:20:57] <steleman> :-)
[20:21:00] <Error_e^ipi> hey
[20:21:06] <nrubsig> but I think I found a hole which we can use to convince even sun to accept it.
[20:21:06] <steleman> he hey hey
[20:21:27] <steleman> what hole would that be ?
[20:21:46] <nrubsig> steleman: right. We use Sun Studio C++ and nothing else. Point. Otherwise we'll run into a deep hell where the walls are pained with blood.
[20:21:49] <Error_e^ipi> nrubsig: the "potential german customers" ?
[20:21:51] <nrubsig> steleman: secret for now.
[20:21:56] <nrubsig> no.
[20:22:12] <Error_e^ipi> yeah, i'm of the opinion that we deliver for studio only
[20:22:14] <nrubsig> actually customers are interested, too.
[20:22:23] <steleman> we use Sun Studio.
[20:22:33] <Error_e^ipi> if someone needs GCC compatible, they can use blastwave
[20:22:34] <steleman> Also Sprach KDE e.V.
[20:22:41] <nrubsig> could we postpone that discussion for the list.
[20:22:49] <steleman> lol
[20:23:17] <steleman> yup im building QCA for QT 4.2
[20:23:23] <steleman> (which i believe is required)
[20:23:38] <nrubsig>  /forcelistname opensolaris opensolaris_kde
[20:23:54] <steleman> we should ask for a kde-solaris IRC channel
[20:24:03] * nrubsig crosses his eyes
[20:24:09] <nrubsig> agghhhrrrll
[20:24:12] <Error_e^ipi> ?
[20:24:15] <steleman> ?
[20:24:28] <nrubsig> erm
[20:24:40] <nrubsig> w don't even have a list yet. Please be patient...
[20:25:01] <steleman> i just dont feel comfortable spamming this channel with minute details about partial template specializations noone but us would care about
[20:25:34] <Berny> .oO(don't always blame the germans)
[20:25:57] <Error_e^ipi> who's blaming anything?
[20:26:08] <Berny> Error_e^ipi nrubsig: the "potential german customers" ?
[20:26:23] <Error_e^ipi> oh... naw, they're helping
[20:26:24] * Berny is not a potential customer ;-)
[20:26:51] * Berny is just happily using stefans kde build ;-)
[20:27:08] <steleman> :-)
[20:27:57] <Berny> though is still haven't managed to build exiv2 with the studio compilers, so digikam is still built with gcc against blastwave's kde :-\
[20:28:16] <steleman> Berny: i havent had the time to fix exiv2 (thanks to PHP)
[20:28:18] <steleman> i have it here
[20:28:20] <steleman> it's half way fixed
[20:28:29] <Berny> you're the man!
[20:28:40] <Berny> .oO(maybe i should learn c++ one day ;-))
[20:28:59] <steleman> the problem is yet again with differences between how studio and how gcc handle partial template specializations
[20:30:10] <Berny> yeah and the way you can do stuff like " int x[n];" in gcc
[20:30:26] <steleman> well that's a studio booboo
[20:30:42] <steleman> if you pass -xlang=c99 to SunProCC you should be allowed to do that
[20:30:43] <Berny> well
[20:31:02] <nrubsig> steleman: in C++, too ?
[20:31:09] <Berny> yes
[20:31:13] <steleman> nrubsig: -xlang=c99 is a SunProCC flag
[20:31:20] <steleman> it's supposed to enable c99 features in C++
[20:31:23] <steleman> (and it doesn't)
[20:31:26] <Error_e^ipi> heh
[20:31:34] *** a_pavlenko has quit IRC
[20:32:36] <Berny> oh well the stuff i've seen in the past days is much worse than all that...
[20:32:44] <steleman> Berny: do tell :-D
[20:32:51] <Berny> nice commercial product with fda approval
[20:33:14] <Berny> does some dipole fitting on magnetic-encephalograph data
[20:33:21] *** estibi has quit IRC
[20:33:48] *** estibi has joined #opensolaris
[20:34:03] <Berny> there linux version does a mem=sysconf (SC_PAGESIZE) * sysconf (SC_PHYS_PAGES) to retrieve the memory size
[20:34:22] <Berny> so far this is a nice way, but mem is declared as unsigned long
[20:34:40] <steleman> and sysconf returns a signed long
[20:35:03] <nrubsig> heh
[20:35:09] <Berny> so on my box with 8GB ram it just barfs out an "Unknown error from unknown source (error code: -8857552)"
[20:35:17] <Berny> no i just overflows
[20:35:58] <Berny> so their bright piece of software think i have a negative amount of ram available and refuses to do anything
[20:36:31] <steleman> can you cast it ?
[20:36:58] <Berny> cast to what? unsigned long long? :-)
[20:37:08] <steleman> you could do that
[20:37:14] <Berny> i don't have the source code for their lib
[20:37:17] <steleman> ohh
[20:37:34] <Berny> i just happened to have one of their guys here to install that stuff
[20:37:42] <Berny> he checked the source on their box
[20:37:45] <steleman> oh
[20:38:08] <steleman> i guess their software can only run on systems where total amount of memory <= ULONG_MAX
[20:38:31] <Berny> and funny thing even though we know what the problem is and even though he built a test lib which worked, we will not get a fixed version until the next release is approved by the fda
[20:38:34] <rodrickbrow_n> go to 64bit leenox
[20:38:42] *** swa_mobil has quit IRC
[20:38:59] <Berny> so for now i just removed 4gb from that box 8-)
[20:39:06] <steleman> help m e
[20:39:20] <Error_e^ipi> steleman: help you do what?
[20:39:33] <steleman> help me understand why fda won't allow them to fix a bug :-)
[20:39:38] <steleman> (doesn't surprise me)
[20:39:51] <Error_e^ipi> maybe the fix introduces new bugs?
[20:39:57] <Error_e^ipi> potentially, i mean
[20:40:05] <Berny> fda would, but that would require a new complete approval process
[20:40:10] <steleman> ohh
[20:40:45] <Berny> so for now 4gb has to be enough
[20:41:08] <Berny> i should check if it is possible to limit the reported phys mem inside a branded zone 8-)
[20:41:52] *** swa_mobil has joined #opensolaris
[20:41:58] <Error_e^ipi> you can limit the amount it can use
[20:42:25] <Berny> yeah but would that change the value reported by sysconf (SC_PHYS_PAGES) ?
[20:42:30] <Error_e^ipi> as for reporting, i dunno
[20:43:03] <nrubsig> make a *.so which overlays |sysconf()| and pokes this value.
[20:43:11] <nrubsig> and LD_PRELOAD it.
[20:44:57] <Error_e^ipi> yay for haphazardly fucking with medical equipment code
[20:45:17] <steleman> as long as it's not connected to someone who must be kept alive ... :-)
[20:45:24] <nrubsig> Error_e^ipi: what is better - such a hack ir using linux ? =:-)
[20:45:25] * Berny is a little wimp in that case :-)
[20:45:37] <nrubsig> s/ir/os/
[20:45:40] <nrubsig> er
[20:45:42] <nrubsig> "or"
[20:45:47] <nrubsig> stupid typos
[20:46:02] *** a_pavlenko has joined #opensolaris
[20:46:05] <Error_e^ipi> linux *IS* a hack, so i guess they're about equal
[20:46:56] <Berny> other choice is run it on sparc and require about 5 times the time to analyze data
[20:47:03] <Error_e^ipi> though I guess you can't really blame Torvalds for it... it was just a school project, nothing big & professional
[20:47:04] <nrubsig> uhm
[20:47:26] <nrubsig> Berny: what do you use ? Ultra1 with 143MHz or what ?
[20:47:33] <Berny> or keep on nagging them for a solx86 build
[20:47:56] <Berny> nrubsig: blade 1500 1.2GHz USIII, 4GB
[20:48:15] <Berny> compared to a 2.4GHz opteron with 4GB
[20:48:26] <nrubsig> grumpf
[20:48:39] * nrubsig wishes Sun wouldn't have dropped the UltraSPARC-4i
[20:49:03] * steleman wishes QCA::Logger actually had an implementation
[20:49:25] <steleman> i has a header file, but no implementation
[20:50:46] * nrubsig wishes the stuff from solaris.kde.org would be compiled as XPG6 application.
[20:50:49] <Berny> well maybe i'm lucky and i get a solx86 build one day
[20:51:00] <steleman> nrubsig: i think it was
[20:51:05] <nrubsig> no.
[20:51:06] <steleman> (i dont remember right now)
[20:51:28] <nrubsig> steleman: AFAIK |popen()|&co. use /sbin/sh and not /usr/xpg4/bin/sh
[20:51:47] <steleman> solaris.kde.org doesnt use either of the sh's ;-)
[20:52:00] <Berny> they use /bin/bash? *duck*
[20:52:05] <steleman> Berny: yes
[20:52:18] <nrubsig> steleman: if |_xpg4| is set to 1 (for a XPG6 application this is true) then popen() uses /usr/xpg4/bin/sh
[20:52:25] * nrubsig adds an item to his todo list
[20:52:46] * Berny takes some servers down
[20:53:03] * steleman wishes the best of luck to nrubsig
[20:53:45] * nrubsig crosses his eyes
[20:53:46] <Berny> bugger
[20:54:11] <steleman> nrubsig: you are going to have to create patches which wil not be accepted upstream
[20:54:52] <Berny> remember: down take the nis+-master down, when you still need to login to another box to shutdown
[20:55:13] <nrubsig> Berny: you don't have replicas ?
[20:55:20] <nrubsig> steleman: patches for what ?
[20:55:44] <steleman> nrubsig: for all the /bin/bash stuff hardcoded inside *.cpp files
[20:55:49] <steleman> (expecting /bin/bash)
[20:56:00] <Berny> nrubsig: down already ;-P
[20:56:17] <steleman> nrubsig: remember: this is written for Linux
[20:56:44] * Berny just loves scripts that say #! /bin/sh and expect it to be bash
[20:57:01] <steleman> that's life on Linux
[20:57:16] <steleman> ./configure sez #!/bin/sh when in fact it's bash
[20:57:30] <nrubsig> steleman: I don't intend to mess around with that initially, but in the long term a POSIX shell would be usefull.
[20:57:47] <steleman> nrubsig: i agree.
[20:58:00] <Error_e^ipi> when things work consistently as they are, we can worry about that
[20:58:13] <nrubsig> steleman: which is about portability, not anti-bash or pro-ksh93
[20:59:14] <steleman> nrubsig: i agree. however, you have to keep in mind that any change we make has the potential of affecting everyone, including Linux, FreeBSD, SUSE, RedHat, etc. so we can't impose a minority view on the majority. and the reality is that on Linux, /bin/sh is in fact bash.
[20:59:57] <oxygene> steleman: but not on freebsd
[21:00:04] <nrubsig> as I said, postponed, please
[21:01:33] *** swa_mobil has quit IRC
[21:01:43] *** swa_mobil has joined #opensolaris
[21:01:55] * steleman hunts for logger.cpp in the worldwide svn
[21:01:56] *** swa_mobil has joined #opensolaris
[21:02:53] *** svoboda has quit IRC
[21:03:14] *** svoboda has joined #opensolaris
[21:04:20] <Error_e^ipi> steleman: btw, your code on the CVS repo is broken
[21:04:35] <steleman> Error what did i break ?
[21:04:50] <Error_e^ipi> it depends on your /home/steleman, and it won't recognize solaris-cc64
[21:05:03] <steleman> where ?
[21:05:10] <dinolinux> hi
[21:05:17] <steleman> oh in the script files ?
[21:05:23] <steleman> solaris-cc-64
[21:05:30] <Error_e^ipi> yes
[21:05:38] <Error_e^ipi> the script files all just fail for me
[21:05:41] <steleman> lemme look
[21:06:53] <Error_e^ipi> and where was the idea to deliver
[21:06:56] <steleman> oh you have to change it or i have to fix it
[21:07:01] <Error_e^ipi> /usr/qt3 and /usr/qt4 ?
[21:07:23] <steleman> we don't know if that's where it will end up
[21:07:40] <steleman> you can change that you have to create the env script files
[21:08:24] <steleman> besides the XEvent handling in QT3 needs to be rewritten
[21:08:31] <steleman> because it's thread-unsafe right now
[21:09:10] <steleman> ok logger.cpp is nowhere to be found
[21:09:24] <nrubsig> XInitThreads() and fix Solaris libX11...
[21:09:33] <Error_e^ipi> qt3 i was never a fan of working with
[21:09:37] <Error_e^ipi> but QT4 is a dream
[21:09:46] <steleman> nrubsig: i am surprised it works at all the way it is right now
[21:09:48] <Error_e^ipi> it works natively on my mac too...
[21:09:53] <nrubsig> steleman: http://www.google.com/codesearch?q=logger.cpp&hl=en&btnG=Search+Code
[21:10:07] <Error_e^ipi> signals/slots... brilliance.
[21:10:26] <Error_e^ipi> <3 qt4
[21:11:25] <steleman> nrubsig: great. which one of these belongs to QCA2 ? i see VLC and a bunch of other stuff
[21:12:52] <steleman> i'll email brad hards
[21:13:57] <steleman> Error: ohh i know what is wrong with solaris-cc-64: you need to symlink one of the qmake.conf.XO? to qmake.conf in the same dir
[21:14:16] <steleman> (QT's configure expects qmake.conf)
[21:14:16] <Error_e^ipi> that'd do it then
[21:14:40] <steleman> and you probably have to edit the qmake.conf.XO? because it depends on my paths
[21:14:52] <steleman> (like /usr/gnu and other things)
[21:16:46] * nrubsig gets to hunt down some food...
[21:16:47] <nrubsig> bye
[21:16:51] *** nrubsig has quit IRC
[21:18:43] <Berny> someone stole my philips screwdriver :-/
[21:19:15] <Error_e^ipi> when you find it
[21:19:20] <Error_e^ipi> take it back from them
[21:19:23] <Error_e^ipi> and stab them with it
[21:19:31] <Error_e^ipi> that'll teach 'em
[21:21:03] <quasi> evil rumors - http://www.theinquirer.net/default.aspx?article=37086
[21:21:12] <steleman> ok i see what happened. very recent checkin for logging capability so i guess logger.cpp didnt make it
[21:22:28] <rodrickbrow_n> is ufs from 8 and ufs in 10 compatible?
[21:22:40] <rodrickbrow_n> ie. if I rebuilt a system on 8 not touch a drive with ufs and some data
[21:22:47] <rodrickbrow_n> it would mount after I upgraded to 10 ?
[21:22:58] <Error_e^ipi> rodrickbrow_n: yes
[21:23:12] <Error_e^ipi> it's 2 versions away... *everything* will work
[21:23:27] <Error_e^ipi> or sun owes you money or something? how does that guarantee work?
[21:24:28] * tsoome is away - bbl - [AcidJazz - Log ON]
[21:26:30] <Auralis> it means its a bug in solaris
[21:27:45] *** deedaw has quit IRC
[21:27:48] <jteo> the nice binary compatibility guarantee.
[21:28:54] * steleman takes a break from QCA2 for now
[21:33:37] <coffman> if i have a zfs on nv54, can i use that under 10u3(11/06 or whatever to call)?
[21:34:50] <jteo> coffman: did you create the zpool on nv54?
[21:34:57] <coffman> yeah
[21:35:46] <jteo> it might not work, since they only guarantee forward compatiblity
[21:36:01] <jteo> "zpool upgrade" will give you the version used for the pool.
[21:36:04] <bitvector2> wouldn't the zfs version indicate that?
[21:36:10] <jteo> if they match, it should mount.
[21:36:25] <coffman> but it could if create the pool under 10u3?
[21:36:33] <Error_e^ipi> s10u3 uses ZFS v3 does it not?
[21:36:34] <coffman> and the use it und b54?
[21:37:09] <jteo> Error_e^ipi: i don't know. I've never used U3/
[21:37:18] <jteo> coffman: correct.
[21:37:59] <coffman> jteo: how about putting it then back to 10u3?
[21:38:44] <coffman> like create the pool under 10u3, write it under b54 and put it back to 10u3?
[21:38:47] <jteo> coffman: no issues there.
[21:39:07] <jteo> version of the pool remains the same unless you explicitly upgrade it.
[21:39:59] <coffman> ic
[21:40:01] <coffman> nice
[21:40:48] *** Theoden-Nexenta has left #opensolaris
[21:45:38] *** lplatypus has joined #opensolaris
[21:49:07] *** comay has joined #opensolaris
[21:55:10] <Error_e^ipi> solaris still doesn't support cpu frequency scaling, correct?
[21:55:15] *** digix has joined #opensolaris
[21:55:16] <Error_e^ipi> (like cool&quiet )
[21:55:30] <digix> hi everyone :)
[21:56:52] <digix> i have a question about how solaris handles arrays and volumes...
[21:57:53] <digix> my server is up and running, and ive configured my array in the controller, but how can i find out where that array is mapped to in solaris? (ie. /dev/dsk/...)
[21:58:41] <Error_e^ipi> format(1M)
[21:58:43] <Error_e^ipi> ?
[21:59:42] <digix> its not listed in format...  crap
[22:00:03] <delewis> try doing a devfsadm
[22:00:09] <delewis> what kind of array is this?
[22:00:53] <digix> raid 6 on an areca controller
[22:00:58] <digix> 2.7TB
[22:01:04] *** Plaidrab has joined #opensolaris
[22:01:32] <digix> i know the controller works, because solaris is installed on a raid 1 on it
[22:04:27] <digix> as i understand, the 32bit kernel only supports up to 1TB. when the system boots, it says that its loaded the 64bit kernel, but a uname still lists i386... could this be the problem?
[22:05:15] <delewis> digix, doubtful. do an 'isainfo -k'
[22:07:34] *** ojpitre has joined #opensolaris
[22:08:24] <digix> ahh, ok. the 64bit extensions are all listed
[22:08:30] <digix> hrm..
[22:09:47] <Berny> touch /reconfigure && init 6
[22:09:50] <Berny> :-)
[22:11:07] <digix> ive already tried the -r kernel arg and it didnt seem to do anything
[22:11:25] <digix> ill try this hough
[22:11:32] <digix> though*
[22:13:19] <delewis> all -r does is run devfsadm
[22:13:26] <digix> ahh
[22:13:26] <delewis> which if you followed me instructions, you've already run it.
[22:13:31] <delewis> s/me/my/
[22:13:42] <digix> i have
[22:14:00] <digix> and the 'touch /reconfigure' reboot didnt seem to do anything either
[22:16:03] <coffman> Error_e^ipi: there is some like this
[22:17:05] *** kimc has quit IRC
[22:17:06] <coffman> Error_e^ipi: frkit delivers powernowadm
[22:17:19] <coffman> http://www.opensolaris.org/os/community/laptop/power/
[22:18:36] <Error_e^ipi> interesting
[22:20:03] <coffman> but its amd only i think
[22:20:56] <coffman> and it does not work perfect
[22:23:54] *** _william_ has joined #opensolaris
[22:24:21] <_william_> hi all
[22:31:23] *** coffman has quit IRC
[22:36:03] *** Berny has quit IRC
[22:40:36] <digix> well, thats not reassuring at all... i open SMC, go to disks, and the system reboots
[22:42:53] <Error_e^ipi> solution: don't use SMC
[22:43:48] <digix> so... command line is the only real way to administer a solaris system at the current i take it?
[22:44:13] <digix> i have about 5 years of linux experience, but am completely lost in solaris right now
[22:44:24] <Error_e^ipi> it's not so much "use the command line" as much as it is "SMC sucks balls"
[22:44:31] <digix> lol, fair enough
[22:44:50] <Auralis> Super Memory Crapper
[22:45:05] <digix> any third party utilities you would suggest to this solaris noob?
[22:45:07] *** Fish- has joined #opensolaris
[22:45:22] <lasseoe> Error_e^ipi: heh, can I quote you on that? :)
[22:46:07] * tsoome is away - bbl - [AcidJazz - Log ON]
[22:50:43] <Plaidrab> Most of the commandline tools behave better. A always doubletake at the coworkers of mine who insist that they need to use admin tool.
[22:51:29] <Plaidrab> digix: screen is your friend. Learn it, love it. :)
[22:52:16] *** jsubl2 has joined #opensolaris
[22:52:17] <lasseoe> admintool? :)
[22:52:32] <lasseoe> you must be kidding!  hah
[22:56:08] *** LeftWing has quit IRC
[22:58:48] <delewis> admintool was marginally better than smc :-)
[22:59:08] <Auralis> at least it did not used half a gig ram for changing a passwd :)
[23:00:47] *** rockpool has joined #opensolaris
[23:00:54] *** Fish has quit IRC
[23:08:18] *** kimc has joined #opensolaris
[23:10:45] <Plaidrab> neither does passwd. :)
[23:11:26] <Plaidrab> Though at some point I have to get annoyed enough to deterine why passwd -r ldap doesn't reliably work in our config...
[23:11:36] <digix> Plaidrab, i learned to use screen long ago :P
[23:11:45] <Plaidrab> Good man.
[23:11:56] <Plaidrab> (well, presumptive man. heh.)
[23:12:01] <digix> lol
[23:12:31] <digix> your presumption is correct
[23:13:16] <Plaidrab> Half of my team didn't start SAing until after GUIS were basicly useful. They think my commandline habits are weird. They look at me funny when I say things like "There is no valid reason for this particular server to have X on it."
[23:13:17] *** Berny has joined #opensolaris
[23:13:36] <Berny>  re
[23:13:53] <Plaidrab> "What about when I log in at the console" "What about it?:)"  They're weird
[23:15:07] <digix> haha
[23:16:23] <Error_e^ipi> I didn't start seriously administering computers even @ home  until gui's were around & useful
[23:16:38] <Error_e^ipi> i still think that most of the time gui admin tools create more problems than they solve
[23:16:55] <Error_e^ipi> gui is good for doing graphical tasks, and that's about it
[23:17:09] <Error_e^ipi> web browsing, movie watching, graphic design, etc
[23:17:32] <digix> i second that... i just feel so lost in solaris since most of my cli experience is with linux
[23:19:31] <delewis> smitty isn't bad in AIX, but in most cases, what you get is an operating system that is biased towards using a GUI for administration, and that's certainly true of AIX (ODM, most of the CLI tools are  PITA to use, as they're meant to be scripted from smitty)
[23:19:46] <Error_e^ipi> docs.sun.com is your friend
[23:21:18] <lasseoe> and google, but of course, if you don't understand your problem, then neither are any good
[23:22:00] <digix> and thats my impasse... i know that solaris isnt seeing my array, but ive no idea why
[23:23:06] <lasseoe> driver, HBA, or array vendor issue, so search on those
[23:23:17] <Auralis> does iostat -En sees the array?
[23:23:46] <Plaidrab> I have used useful gui tools. There are some things that can be expressed better that way. I just can't think of one on Solaris. Well, The Sunray admin  thingy is useful, but it's web based. Does it count? :)
[23:24:01] *** myrkraverk has joined #opensolaris
[23:24:28] <digix> i have a raid 1 on the same controller that solaris is installed to, so i assume that driver and card itself is working fine
[23:24:34] <lasseoe> netbackup and legato guis aren't all that bad
[23:24:38] <delewis> Plaidrab, dhcpadm isn't bad.
[23:24:57] <Plaidrab> I don't use dhcp, so I'm not surtprised I didn't know that. :)
[23:25:25] <digix> iostat -En also shows 1 soft error and 1 illegal request... not sure what those are about
[23:26:01] <Plaidrab> I think the problem is a lot of the GUI tools get designed by people who either dislike guis or just don't understand good interface design and as a result, you get weird crap tools. :)
[23:26:08] *** jopi has joined #opensolaris
[23:26:15] <jopi> hi all
[23:26:33] <Berny> bugger
[23:26:57] <Plaidrab> What kind of raid is this? Something like a D1000?
[23:26:58] <jopi> is this the right channel to ask for doubts at newbie level?
[23:27:03] <digix> areca 1230
[23:27:08] <Error_e^ipi> jopi: couldn't hurt
[23:27:08] <Auralis> jopi: just ask away
[23:27:13] <jopi> Error_e^ipi: thx
[23:27:19] <Berny> .oO(while reorganising your server room, don't fall over fc/al cables they don't like that :-\)
[23:27:34] <Plaidrab> Try not to step on them too. :)
[23:27:53] <Berny> could you tell me before i kill a cable next time :-)
[23:27:55] 
[23:27:56] *** Tpenta has quit IRC
[23:28:07] <lasseoe> Bearny: most people don't like falling either, so it's a tie :)
[23:28:26] <Plaidrab> You installed your locale?
[23:29:09] *** _william_ has quit IRC
[23:29:28] <jopi> Plaidrab: not consciently, I just installed with default options
[23:29:33] *** hspaans has quit IRC
[23:29:51] <jopi> I see the application menus correctly in spanish, but keyboard is in english format
[23:31:32] <Plaidrab> I think you need the locale for the proper keyboard. I forget if there's a package just for that bit of stuff.
[23:32:56] <jopi> Where do I find the locale package?
[23:33:38] <Plaidrab> Presming the layout of the OE disk matches my memory, you should find it in the Solaris_10/Products/ directory on your media.
[23:33:50] *** nostoi has quit IRC
[23:33:51] <Plaidrab> That should contain all the relevant patches.
[23:33:58] <Plaidrab> packages, rather.
[23:35:19] 
[23:35:22] <Plaidrab> http://developers.sun.com/dev/gadc/techtips/sol9_locale_pkgs/index.html
[23:35:38] <Plaidrab> That's Solaris 9, but most likley the package names are the same.
[23:36:18] <jopi> thx again, now searching
[23:43:18] *** jsubl2 has quit IRC
[23:48:38] *** ojpitre has quit IRC
[23:52:21] <jamesd__> how to do i make  format in solaris 10 not to bitch about the disk having mounted partitons... trying to create a metadb slice on a drive that is mounted as  / and swap.
[23:53:10] <Doc> Dear customer,  The reason your cluster broke is because you broke it.  Have a good day.
[23:54:03] <Plaidrab> Hee!
[23:54:24] <jamesd__> doc any guesses?  trying to shrink  /export/home and giving  50MB to a metadb, and i don't have a serial console setup only ssh.
[23:54:26] <Doc> well they wanted it in email!
[23:54:36] <Doc> yes. dont.
[23:54:37] <Plaidrab> You can take it from the swap slice
[23:54:55] <delewis> Plaidrab, that requires booting into single-user mode.
[23:55:00] <Plaidrab> no
[23:55:09] <tomww> jamesd__: htere is a environment variable to shut up format.
[23:55:10] <jamesd__> Plaidrab, i just need to know how to  let format  relabeling the disk... with mounted partitions.
[23:55:30] <Plaidrab> You can create a swap file and add it. Then unmount the swap slice and repartition
[23:55:55] <Plaidrab> Then remount the swap slice, remove the swap file, and delete it.
[23:56:10] <jamesd__> but root will still be mounted...  so it will bitch
[23:56:23] <Plaidrab> Why?
[23:56:31] <Plaidrab> I've done this before, it's never been an issue.
[23:56:56] <tomww> NOINUSE_CHECK=1 format
[23:57:11] <Doc> do NOT shrink home without newfs'ing it
[23:57:23] <Plaidrab> You are talking about creating a slice 7 metadb for SVM?
[23:58:05] * BadKarma is watching CNN and kinda freaking out
[23:58:23] <jamesd__> partition> la
[23:58:23] <jamesd__> Cannot label disk while it has mounted partitions.
[23:58:23] <jamesd__> partition>
[23:58:46] <Plaidrab> Huh.
[23:58:52] <Plaidrab> 8 and 9 don't do that.
[23:59:09] <tomww> Plaidrab: yes, and this is not the best :)
[23:59:10] *** adlpaf has quit IRC
[23:59:37] <BadKarma> hillary clinton announced her candicy for president





top