November 11, 2006 
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30
November 11, 2006 [00:06:15] *** jamesd_ has joined #opensolaris [00:07:52] <gisburn> booooooring [00:11:13] <gisburn> Ok... [00:11:44] <gisburn> ... I am yellow, covered with slime, climb trees and like mushrooms - what am I ? [00:12:06] <asyd> give me some mushroom [00:12:14] <dwc-> banana slug? [00:12:23] <gisburn> ugh [00:12:24] <gisburn> dwc-: yes [00:12:32] <jamesd_> hmm some one that we keep in hte dark and fee shite. [00:12:34] <dwc-> oops [00:12:35] <jamesd_> er feed [00:12:47] <gisburn> dwc-: WTF did you figure that out that fast ? [00:13:08] <dwc-> because there's banana slugs in the mountains to the south not a half hour or so away [00:13:27] <dwc-> and the college on the far side of the mountains have banana slugs as their mascot [00:13:36] <gisburn> yeah.... University of California. [00:13:50] <dwc-> http://www.ucsc.edu/about/campus_mascot.asp [00:14:18] <dwc-> it was a guess really, but that was the only yellow slimy thing I could think of [00:14:19] <gisburn> Ok... [00:14:48] <Error_404> yum... banana [00:14:49] <gisburn> ... I am the old silver, the very old, poisonous but often loved and used together to get gold. [00:14:52] <gisburn> What am I ? [00:15:16] <dwc-> this is another easy one for me [00:15:24] <dwc-> so I'll not answer [00:15:55] <gisburn> anyone else ? [00:15:58] <Error_404> lead? or mercury? [00:16:02] * dwc- nods [00:16:07] * dwc- points to the latter [00:16:10] <gisburn> mercury [00:16:12] <Error_404> thought so [00:16:21] * dwc- lives near a quicksilver park [00:16:30] <gisburn> dwc-: grrrr [00:16:32] <dwc-> where they used to have ... quicksilver mines [00:16:41] <Error_404> i just know how to extract gold from ore for some reason [00:17:28] <gisburn> Ok... [00:17:53] <gisburn> ... I am not rome. All ways lead away from me. And I smell like the whahoonie fruit. What am I ? [00:18:00] *** jamesd has quit IRC [00:18:38] <gisburn> (large city with many dwarfs) [00:18:54] <gdamore> emerald city [00:18:58] <gisburn> ?! [00:18:59] <gisburn> No. [00:19:33] <gdamore> wtf is whahoonie fruit? [00:19:48] <dwc-> google knows not, but suggests wahoonie [00:20:12] <Error_404> paris [00:20:15] <quasi> tyou fruit ;) [00:20:22] <quasi> typo [00:20:53] <gisburn> Error_404: no. [00:20:56] <Error_404> perhaps berlin? [00:21:10] <Error_404> (those wacky facists and their effective transportation engineering) [00:21:11] <dwc-> which looks to be something similar to durian (yum), from fiction [00:21:12] <gisburn> heh [00:21:18] <gdamore> only large city with lots of dwarfs i can think of is moria [00:21:26] <gisburn> gdamore: wrong world. [00:21:49] <dwc-> howondaland? [00:21:50] <gisburn> It's ANOTHER world. [00:22:02] <gisburn> Not this one, not Tolkies one. [00:22:15] <gisburn> And not related to Illuvatar&co. [00:22:16] <gdamore> some WoW thing? [00:22:24] <gisburn> gdamore: no. [00:22:25] <clee> hooray ironforge [00:22:28] <dwc-> The Wahoonie is a large, spiky, smelly fruit, resembling the Roundworld durian, but even larger and elongated. The city of Ankh-Morpork is sometimes affectionately called the Big Wahoonie as New York is called the Big Apple. [00:22:42] *** jcea has joined #opensolaris [00:22:43] <dwc-> is that the right one? [00:22:52] <dwc-> (I have no idea what they're talking about) [00:23:02] <Error_404> heh [00:23:03] <gisburn> dwc-: you didn't say the name of the town yet. [00:23:06] <alanc> you need to read more Terry Pratchett then [00:23:23] <dwc-> or so it appears.... [00:23:30] *** ircminer has joined #opensolaris [00:23:46] <gisburn> dwc-: what am I ? [00:24:10] <dwc-> ksh93 fanatic? [00:24:17] <gisburn> heh [00:24:20] <gisburn> dwc-: no. [00:24:23] <gdamore> komodo dragon fetisher? [00:24:30] <gisburn> gdamore: no. [00:25:12] * gisburn points out that dwc-'s quote already contained the name. [00:25:17] <gisburn> A.M. [00:25:30] * gdamore points out that nobody else seems to care. :-) [00:25:48] * gisburn throws gdamore in the Ankh river. [00:25:57] <dwc-> apparently alan's the only one that reads terry pratchett [00:26:12] <gisburn> Ok... [00:26:22] <gdamore> i've read _some_ terry pratchet, but i don't recall the Ankh or the city... [00:26:39] <alanc> I should actually be working after spending the morning at the ubuntu summit [00:26:42] *** slowhog has quit IRC [00:26:50] <gisburn> ... I am the old land of men... where once the people of gondor lived... the five-pointed star. What am I ? [00:26:53] <Error_404> hmm... this polaris checkout always hangs on troff.d/dev.h [00:26:58] <gdamore> numenor [00:27:09] <gisburn> gdamore: correct... ;-/ [00:27:26] <alanc> btw gisburn: did you see the mail to xprint at mozdev dot org last week asking if anyone had seen you? [00:27:45] <dwc-> did someone suggest that the komodo dragon had eaten him? [00:27:56] * gisburn made a note to answer this weekend or next week [00:28:38] <alanc> "Roland has been the main person that pulled this project, he's also very kind and responsive." - they must not know about the komodo dragons [00:28:51] <gdamore> heh. [00:29:44] <gdamore> gisburn: I am the essence of Evil, Sauron was my protoge. who am i? [00:30:16] *** craigmohrman has joined #opensolaris [00:30:29] <dwc-> c++? [00:30:52] <gdamore> hehe. well, maybe in a different world... :-) [00:30:54] <sickness> gisburn: hi! :) [00:31:18] <craigmohrman> Hello hello [00:32:05] <gisburn> craigmohrman: Hi! :-) [00:32:18] <gisburn> gdamore: Melkor [00:32:21] <craigmohrman> Hello Roland. Finally made it. [00:32:26] <gdamore> yes, aka Morgoth. [00:36:18] *** craigmohrman has quit IRC [00:38:03] <gisburn> erm [00:38:14] <gisburn> !summon craigmohrman [00:38:52] *** uebayasi has joined #opensolaris [00:39:43] * jamesd_ listens to the room echos with "REQUREST DENIED" and RMS pops into gisburn's lap. [00:40:14] *** mercykiller has joined #opensolaris [00:40:53] *** slowhog has joined #opensolaris [00:46:06] *** ircminer has quit IRC [00:47:43] *** Fish- has quit IRC [00:59:28] *** ircminer03 has joined #opensolaris [01:11:38] *** sommerfeld has quit IRC [01:15:20] *** nbkk6fo has joined #opensolaris [01:15:23] *** nbkk6fo is now known as rodrickbrown [01:16:44] *** ircminer03 has quit IRC [01:17:47] *** ircminer03 has joined #opensolaris [01:25:35] *** bunker has quit IRC [01:32:34] *** yarihm has quit IRC [01:47:57] *** mercykiller_ has joined #opensolaris [01:52:31] *** CommGuru__ has joined #opensolaris [01:52:41] *** CommGuru__ is now known as CommGuru [01:56:41] <gisburn> Ok... [01:57:30] <gisburn> .... I am brown and in ancient times I was consumed with pepper after my fermantation. And I am a nut. What am I ? [01:57:57] <Kitty> gisburn [01:58:21] <gisburn> Kitty: try again. [01:58:24] <alanc> lol, here I was just thinking "gisburn is a bit of a nut".... [01:58:32] * gisburn kicks alanc [01:58:53] * gisburn grabs a taster and tortures alanc for fun [01:58:56] <gisburn> er [01:58:59] <gisburn> s/taster/taser/ [01:59:03] <gisburn> stupid typos [01:59:12] <gisburn> come on [01:59:19] <gisburn> your kids drink it all the time [01:59:36] <Error_404> whiskey? [01:59:48] <Tpenta> that's almost quotable Roland... gisburn: ... [A]nd I am a nut. :) [01:59:57] * Tpenta ducks [02:00:06] <Tpenta> soudns licke chocolate to me [02:00:07] <gisburn> Error_404: sounds you should never get kids. NEVER [02:00:22] <gisburn> Tpenta: yes [02:00:49] <gisburn> The quiz items are too easy, right ? [02:01:39] *** mercykiller has quit IRC [02:03:14] <gisburn> Ok... [02:03:50] *** jamesd_ is now known as jamesd [02:03:57] *** ChanServ sets mode: +o jamesd [02:04:24] <jamesd> gisburn, next quiz... how long till you get kicked for off topic annoying jokes ;-p [02:04:40] <gdamore> i take 10 minutes. :-) [02:05:00] *** gisburn has left #opensolaris [02:05:19] <gdamore> heh. off by a factor of 60. [02:05:52] <alanc> so what prize does gdamore win? getting his code putback before ksh93 is? [02:05:52] *** gisburn has joined #opensolaris [02:05:57] <gisburn> jamesd: funny [02:06:04] * gisburn sits in his edge and sulks [02:06:13] <gdamore> actually, i've already had code putback before ksh93. several times. [02:06:33] <jamesd> takes a nap those stupid riddles were too much for me .... [02:07:00] * gdamore sends gisburn a copy of the oss-bite-size bugs list, since he seems to be so bored... [02:07:15] <Error_404> do all of them [02:07:39] <alanc> careful what you wish for, you might find all of them fixed by replacing them with libcmd wrappers [02:07:50] <gdamore> oh, nuts, didn't think about that. [02:07:57] <alanc> 8-) [02:08:21] <gdamore> mv /kernel/unix /kernel/unix.old; ln /usr/bin/ksh93 /kernel/unix [02:09:15] * gisburn starts sulking harder [02:09:18] <alanc> don't forget ln -s /usr/bin/amd64 /kernel/amd64/unix [02:09:32] <alanc> whoops, /usr/bin/amd64/ksh93 [02:09:43] <gdamore> there were some scary bugs in what _should_ have been a simple command line parser for xargs.... i wound up fixing bugs introduced by XPG6 xargs! :-) [02:10:46] <gdamore> anyway, whodda thunk there would be _bloat_ in something as simple as xargs? :-) [02:11:03] <LeftWing> heh [02:11:18] <alanc> this should cheer gisburn up: usr/X11/bin/Xorg: ELF 64-bit MSB executable SPARCV9 Version 1, UltraSPARC1 Extensions Required, dynamically linked, not stripped [02:11:24] <gisburn> gdamore: too good that xargs becomes obsolete in the future [02:11:39] <gisburn> alanc: which compiler did you use ? [02:11:41] <alanc> now if only I had a driver to go with it [02:11:44] <gdamore> alanc: thats great! [02:11:51] <gdamore> oh, no driver, not so great. [02:12:03] <gdamore> gisburn: why do you think xargs will become obsolete? [02:12:25] <LeftWing> <insert ksh93ism here> [02:12:25] <gdamore> you can't replace it by libcmd. trust me, i've looked at the hideous compatibility argument parsing in it.... [02:12:27] <alanc> jay has 32-bit Xorg on SPARC working with m64 using /dev/xsvc, so it shouldn't be hard to get a driver [02:12:30] <gisburn> gdamore: ksh93 builtins do not have an argument list length restriction. [02:12:41] *** rodrickbrown has quit IRC [02:12:42] <gdamore> its not a length limitation. [02:12:45] *** rodrickbrown has joined #opensolaris [02:13:08] <gdamore> its how the -X (or whatever) args are processed. [02:13:15] <gdamore> read the code. [02:13:39] <alanc> and I've had scary thoughts about the wsfbdev driver from BSD, which looks suspciously like the FBIO* ioctl interface [02:13:42] <gdamore> there are a lot of undocumented compatibility hacks and such that can't be replaced, even though they aren't in man pages, because they will break something somewhere. [02:13:56] <gdamore> alanc: it pretty much is. [02:14:59] <gdamore> at one point i was seriously thining of adding support for a DMA/shared memory segment (kernel to userland) that would do 2D (and maybe some 3D?) acceleration for wsfb. This would have been NetBSD. But then I lost interest in NetBSD. [02:15:18] <alanc> so I wonder if that could get free unaccelerated Xorg support for all the drivers Linda's team hasn't ported [02:15:23] * gdamore still believes that userland programs should not be mapping PCI device registers. [02:15:38] <gdamore> well maybe not free, but close to it. [02:16:25] <gdamore> i suspect some of the cursor ioctls and such may need tweaking. [02:16:56] * gisburn is away for 10mins [02:17:15] <gdamore> (is that to avoid getting kicked? :-) [02:17:32] <alanc> I should publish my silly console mode test program for hwc/FBIOCURSOR to let you move the mouse pointer around the sparc console in text mode [02:17:38] * gdamore wishes sun would give me a job offer.. :-) [02:17:57] <gdamore> i have one that does the same thing for netbsd. [02:18:31] <alanc> wow, that was a long time ago: -r--r--r-- 1 alanc staff 6673 May 18 1999 /home/alanc/hwc/hwctest.c [02:18:59] <gdamore> heh. i wrote mine just a few months ago. as part of the work I was doing for radeonfb on netbsd. [02:19:28] <alanc> bah, mine requires Xsun server internal headers to build [02:19:53] <gdamore> www... i can't find mine right now. but it used netbsd headers. :-/ [02:20:22] <alanc> hmm, I should opensource the hwc kernel module one of these days since the ON guys haven't responded to my whines that they should take it from us [02:20:23] <gdamore> found it, wstest.c [02:21:19] <gdamore> hmm... it uses a lot of WSDISPLAYIO_xxx ioctls, rather than FBIO ioctls. i can share if you want it. [02:22:53] <gdamore> alanc: and yes, any thing you can do to opensource it would probably be good. if we could just get the sparc graphics group to open source... [02:23:06] <gdamore> btw, Linda never got back to me. should I ping her? [02:23:52] <alanc> hwc lives in X, but it's just a small streams module that gets pushed onto the mouse stream to make the FBIO calls directly to move the mouse in the kernel without having to context switch into and out of the X server first [02:24:01] <delewis> some guy (Michael Plant of Sun, or so he claimed) contacted me a week or so ago (I imagine after my rant on osol-discuss about SPARC graphics), and asked if I had any questions regarding "Sun's IP initiatives" [02:24:12] <delewis> he was horribly unprepared and didn't even know why he was contacting me (or so he claimed) [02:24:22] <clee> funny, I also got a call from Sun last week [02:24:30] <clee> guy who was calling me also seemed to have no idea why he was calling me [02:24:37] <delewis> clee: yeah, that's it. [02:24:38] <gdamore> alanc: that sounds "weird". so how is the plumbing between the framebuffer and the hwc managed? [02:25:18] <clee> delewis: my call was from somebody in Services, though. apparently it was a followup to some survey or something I filled out on Sun's site [02:25:22] <clee> but he didn't know which one [02:25:29] <delewis> clee: ah [02:25:34] <gdamore> sounds like a clueless marketdroid. [02:25:37] <clee> yeah. [02:25:39] <delewis> was it a CO # by any chance? [02:25:44] <clee> actually, yes [02:25:49] <delewis> yeah, that's it. :-) [02:25:50] <clee> 303 area code [02:25:50] <alanc> not sure - it's been too long since I've looked at hwc [02:25:55] <delewis> clee: bingo. [02:26:06] <alanc> I'll have to publish the source just so you can figure it out 8-) [02:26:25] <gdamore> heh, it wouldn't be the first sun code i've had to do that for. :-) [02:26:40] <gdamore> (can you say "Sun Ray"? :-) [02:27:34] <alanc> hmm... maybe I can just mark PSARC 1993/334 as open - don't think hwc is particularly secret [02:27:39] <gdamore> i sometimes wish bugtraq had a way to record the bugs that have been fixed by patches I sent in while _not_ a Sun employee... :-0 [02:28:12] <alanc> gives very little detail though [02:28:14] <gdamore> alanc: heh. i'm surprised it isn't patented. :-) [02:28:29] <gdamore> (or is it?) [02:28:36] <alanc> "With the SunWindows driver, the driver intercepted the mouse [02:28:36] <alanc> events and sent the new location to the cg6 at the same time it [02:28:36] <alanc> sent the events up to the OpenWindows server. In the MIT [02:28:36] <alanc> server, we read the mouse event into the server and then send [02:28:36] <alanc> an ioctl() to the cg6 to have it move the pointer for us. The [02:28:37] <alanc> problem with this is that there could be some delays before the [02:28:39] <alanc> server processes the mouse events which will cause the screen [02:28:42] <LeftWing> clee: Be careful. I got a call like that, and it turned out to be scammers looking for credit card numbers. [02:28:43] <alanc> pointer to stutter." [02:28:45] <alanc> no clue if it's patented [02:28:56] <delewis> LeftWing: !? [02:28:58] <delewis> interesting. [02:29:05] <delewis> I figured this Michael Plant guy might've been suspect. [02:29:12] <LeftWing> I also got a string of dodge-arse e-mails. [02:29:15] <clee> LeftWing: well, I had alanc look him up on the intranet :) [02:29:17] <alanc> you know it's old when it talks about SunWindows... [02:29:18] <delewis> I did a reverse phone search on his # and it came back as a personal #. [02:29:24] <gdamore> alanc: i wonder if this is the source of mouse pointer stutter i sometimes notice with Xorg on x86. [02:29:32] <delewis> clee: what were the results? [02:29:40] <LeftWing> clee: Ahh cool. I had my 'boss'/mate talk to our Sun marketting contact about it. [02:29:45] <clee> delewis: well, I told him that I couldn't really give him much feedback [02:29:56] <clee> (given that he didn't even know which survey it was that he was calling me in response to) [02:29:59] <delewis> clee: because you had no clue what he was talking about? :-) [02:30:03] <clee> exactly [02:30:07] <delewis> yeah, I was stumped, too. [02:30:15] <clee> but I did mention that, overall, Sun's site generally does have the info I'm looking for when I go there [02:30:18] <alanc> gdamore: we haven't ported it to Xorg yet since we've had no x86 kernel drivers that recognized FBIO cursor ioctls [02:30:21] <LeftWing> I was happy to answer his generic "So you've downloaded Solaris 10" questions. But when it got to the point where he'd asked for my address because of 'database corruption', I stopped it there. [02:30:24] <gdamore> but then again, i used ioctls from userland on our Alchemy based NetBSD box (Sun Ray soft client) and it worked out fine [02:30:49] *** Cybernd has quit IRC [02:31:14] * LeftWing wants a Java Soft Ray. [02:31:26] <delewis> no, you don't. :-) [02:31:29] <gdamore> alanc: the approach, intra-kernel communication between the mouse and the framebuffer device, is still a good idea. Xorg should adopt it. :-) [02:31:42] * gdamore thinks LeftWing needs his head examined. [02:31:49] <alanc> if only we had framebuffer devices in the kernel... [02:31:51] <LeftWing> Would be handy for administrative purposes. =P [02:32:00] <LeftWing> I'm not thinking for regular use. [02:32:11] <gdamore> LeftWing: "ssh" :-) [02:32:26] <gdamore> and if you need an X server, then Hummingbird or somesuch. :-) [02:32:31] <LeftWing> I mean to attach to existing Sun Ray sessions from things that aren't Sun Ray. [02:32:35] <LeftWing> *Rays [02:32:42] <gdamore> oh, yeah, session mobility would be nice. [02:32:55] <delewis> in that case, you want SSGD. [02:32:58] <gdamore> that was one of the reasons i wanted a soft ray for Solaris for so long. [02:33:05] <gdamore> SSGD? [02:33:19] <alanc> SSGD == Tarantella [02:33:20] <delewis> Sun Secure Global Desktop (a.k.a. Tarantella) [02:33:29] <delewis> sort of a Citrix-clone, if you will. [02:33:32] <LeftWing> delewis: No, not really. I want to be able to attach to an existing Sun Ray session from, say, a wifi laptop when I realise there's not a Sun Ray around right now. [02:33:35] <delewis> both do basically the same thing. [02:34:05] <LeftWing> I use SGD as well, but it doesn't fill that niche. [02:34:12] <gdamore> SSGD won't help if you already have a Sun Ray session though. [02:34:18] <LeftWing> gdamore: Exactly. [02:34:25] <gdamore> and if you want to use actual Sun Rays, you need to use Sun Ray protocol. [02:34:32] *** sahafeez has quit IRC [02:35:05] * gdamore is right this minute creating an alternate remote USB driver stack for the Soft Ray software. [02:35:32] <LeftWing> heh [02:35:49] <gdamore> the sun ray protocols are a bit twisted. [02:36:38] <jbk> heh [02:36:40] <LeftWing> Well at least the units are built so that the protocols can be replaced. [02:36:51] <gdamore> sort of. :-) [02:43:13] *** salamanders has joined #opensolaris [02:43:52] *** oooo has joined #opensolaris [02:46:39] *** salamanders has joined #opensolaris [02:48:47] <gisburn> Tpenta: http://projects.info-pull.com/mokb/MOKB-04-11-2006.html [02:50:08] <Tpenta> so, has a bug been logged and a call been made? [02:50:09] <alanc> just another reason you should be using zfs instead of ufs 8-) [02:50:28] <gisburn> alanc: Dont ask for zfs bugs. [02:50:34] <gisburn> Tpenta: I don't know. [02:50:39] <gisburn> Tpenta: I just saw it. [02:50:44] <Tpenta> ahh [02:50:54] <alanc> zfs bugs don't bother me, I don't have to fix them [02:50:57] <gisburn> Solaris, the most buggy OS on the planet. See http://projects.info-pull.com/mokb/MOKB-04-11-2006.html [02:51:05] <gdamore> i'm not sure this is a bug in ufs or not. will on-disk corruption create it? [02:51:28] <gdamore> gisburn: bwahahaha.. [02:51:35] <gisburn> Tpenta: Linux guys just posted it around as "proof" how bad solaris is. [02:51:37] <alanc> wow, one bug makes it "the most buggy OS"? ever seen Windows 98? [02:51:54] <Tpenta> If I read this correctly, they corrupt the on disk filesystem and then bitch that it panics ufs? [02:52:11] <jbk> that's what it sounds like [02:52:18] * Tpenta shakes his head [02:52:19] <Tpenta> sad [02:52:20] <gisburn> alanc: I don't agree with them. I just saw the discussion in one of our lists. [02:52:25] <gdamore> bad data on disk (apart from the kernel) "shouldn't" panic the kernel [02:52:26] <lplatypus> is the zfs code smart enough to make two mirrored drives perform better than a single drive? [02:52:27] <jbk> of course linux has never been one for data integrity [02:52:36] <lplatypus> at least for random reads... [02:53:19] <jbk> lplatypus: if not, it's probably in the works [02:53:30] <Error_404> jbk: esp. since a bunch of distros push ricerfs [02:53:33] <gdamore> linux is so terribly full of holes it isn't even funny. [02:53:40] <Tpenta> I'm tempted to download the filesystem, fsck it and see if it's fixed [02:53:47] <Error_404> which has this wonderful property of just obliterating data randomly [02:53:56] <gisburn> Error_404: [02:54:01] <gisburn> Error_404: well [02:54:14] <gisburn> Error_404: I am typing from a SuSe 10.1 system right now. [02:54:31] <gdamore> anyway, if on-disk corruption can panic the kernel (apart from kernel module data being corrupt), then it is probably a real bug. [02:54:32] <gisburn> Error_404: which is running stable since 50 days with two VMware VMs running. [02:54:41] <g4lt-U60> gisburn, redmondlinux? ;P [02:54:46] <gisburn> suse linux [02:54:48] <gdamore> gisburn: 50 days. hah. [02:54:56] <gisburn> gdamore: this is a laptop [02:55:07] <gisburn> gdamore: and solaris on a laptop is no match for linux [02:55:16] <g4lt-U60> gisburn, google for novell+microsoft [02:55:18] <gdamore> you haven't been running one of _our_ laptops. [02:55:35] <gisburn> g4lt-U60: I know that. And I know the Sun+M$ deal. [02:55:36] <g4lt-U60> yeah, gdamore's run solaris pretty well [02:55:44] <jbk> some of us don't want to sell our first born for them :) [02:55:55] <gisburn> gdamore: I would begg for a sparc laptop. [02:55:57] <gdamore> yeah, well, i can't help the price. sorry guys. [02:56:11] <gisburn> gdamore: right now it seems we're doing the ksh93 putback for i386/AMD64 only [02:56:14] <gdamore> but low end 500MHz systems should be available for ~$3k. [02:56:19] <g4lt-U60> yeah, the used market is even getting obscene again, I mean like $500 for a laptop equivalent to a SS%? [02:56:31] <g4lt-U60> SS5 even [02:56:40] <delewis> g4lt-U60: more like an SB100 [02:56:43] <delewis> which is still utter shit. [02:57:01] <g4lt-U60> delewis, no, even 3gx's are running $500 nowadays [02:57:13] <gdamore> delewis: you're not still hung up on the CMDK are you? the sb100s work pretty well. [02:57:38] <delewis> gdamore: not compared to an Ultra 2 or SB1000, which is what I'm using at the moment. [02:57:41] <g4lt-U60> delewis, if you tell that to hte andrea-doria, I'll have to kill you [02:57:46] <gdamore> and the SPARCLE 500 is a pretty decent system. not a performance machine. but nice. [02:57:59] <gdamore> yeah, but try putting a U2 or SB1000 in your "lap" :-) [02:58:10] <delewis> hah. [02:58:18] <g4lt-U60> gdamore, that would be the sparcbook 5000 ;P [02:58:29] <Tpenta> it would be nice f the crashdump for that ufds thing was available somewhere [02:58:30] <gisburn> gdamore: I think the SB1000 fans are sufficient enougth to turn this machine into a hoovercraft [02:58:34] <gdamore> sparcbook 5000 and sparcle 500 are nearly identical inside. [02:58:49] <gdamore> heh. you should see the fans we have on the Bullfrog 2. [02:58:49] <g4lt-U60> ahh, then what was the US2 sparcbook? [02:59:11] <gdamore> we had Ultrabook IIi, and prior to that original Ultrabook. [02:59:18] <gdamore> I think the original Ultrabook used a US-1. [02:59:24] <gdamore> but it was before my time. [02:59:35] <g4lt-U60> gdamore, let me guess, you merged with general dynamics so you could get their turbofan technology? ;P [02:59:42] <gdamore> heh. [02:59:58] <gisburn> well, reheat may be nice [03:00:03] <gdamore> anyway, we have a US3i based 15" laptop that is nice. [03:00:11] <gisburn> argumented turbofan zsed for cooling! [03:01:00] <gdamore> gisburn: why do you say ksh93 is only getting putback for x86? that shouldn't be permitted according to ON rules. [03:01:14] <gisburn> Tpenta: is there any precedent for doing a putback for only one arch (e.g. i386) ? [03:01:28] <Tpenta> roland, you will not be allowed to do that [03:01:30] <gisburn> gdamore: due lack of resources we may not have a way to handle sparc [03:01:41] <gdamore> there is precdent, but it requires extra paper work, and lots and lots of justification. [03:01:44] <Tpenta> the only acceptable reason for doing that is architecture specific drivers [03:01:48] <gdamore> your resources complaint won't be acceptable. [03:02:09] <gdamore> Tpenta: IMO, even that excuse has been "overused". some drivers that should be common are not. [03:02:16] <Tpenta> gdamore +1 [03:03:39] <gdamore> gisburn: there used to be waivers that people could get to commit features for sparc but not x86. but now x86 is supposed to have feature parity, and therefore all features that make sense for all ports must be supported for all ports. [03:04:03] <gdamore> Tpenta: can someone at Sun just ship a SB1000 loaner or some such to gisburn? [03:04:24] <alanc> hmm, might have a spare U10 [03:04:46] <gdamore> i think the problem is that gisburn is whining about wanting something faster than 500MHz. [03:04:58] <gdamore> i'm not sure why thought. he shouldn't have to do full builds all the time. [03:05:03] <alanc> but shipping from US to Germany sounds like customs nightmare [03:05:21] <gdamore> Sun has a presence in germany. [03:05:26] <alanc> last I saw public build machines were planned for Q1 2007 [03:05:33] [03:05:35] <gdamore> that will be a big help. [03:05:40] <delewis> what about setting up a community development server? [03:05:53] <delewis> oh, well, I guess that wouldn't help ksh93 too much. [03:06:30] *** hell` has quit IRC [03:06:33] <gdamore> if the servers can be configured with LDOMS with different revs of Solaris, that will be cool. [03:06:48] <alanc> of course these days, we see more waivers to ship stuff for x86 and not SPARC, like Xorg 8-) [03:06:51] <delewis> maybe with the Niagra hyerpvisor [03:06:59] <delewis> hypervisor, rather. [03:07:03] <gisburn> hype-vistor [03:07:26] <gisburn> virtual hype manager [03:07:29] * gdamore wants a laptop built with a pair of niagra-2's. [03:07:37] <jbk> heh [03:07:51] * gisburn wants a machine with one niagara-1 right now [03:08:13] <gdamore> hmmm..... store.sun.com? [03:08:16] <jbk> i've been chomping at the bit to get some niagras in [03:08:17] <alanc> what would you do with that many cores in a laptop? give each firefox tab it's own dedicated core? [03:08:20] <gdamore> or sun.com/store [03:08:31] <gdamore> alanc: dmake :-) [03:08:32] <gisburn> alanc: that will not work [03:08:45] <jbk> however the people that decide sizing/etc have said that their group is too stupid to know where it should be deployed, therefore we won't use it [03:08:45] <astinus> alanc: pfft, try and make JDS run faster ;) [03:08:46] <gisburn> alanc: my patch to make gecko multitthreaded was rejected [03:09:14] <alanc> does JDS have enough going on in parallel for that to help? [03:09:24] <gisburn> alanc: not really. [03:09:25] <gdamore> probably not. [03:09:42] <gdamore> isn't the Xserver still basically single threaded/poll driven? [03:11:03] <gdamore> what i like about niagra2 is reduced power consumption, with a lot of cycles for parallel making available on demand [03:11:21] <gdamore> we also have customers that use some laptops as mobile sun ray servers. [03:11:30] <gdamore> (niagra 2 would be _very_ nice for that.) [03:12:07] <alanc> gdamore: yes, though we've had some wild ideas about rewriting the Xorg select() loop using threads and POSIX AIO [03:12:27] <gdamore> that would be "interesting". [03:12:39] <Tpenta> oooooo optimising X for multi cpus, what a novel concept :-D [03:12:44] <clee> alanc: keithp is against it though [03:13:00] <alanc> would have to keep the threads limited to the I/O handling though - can't re-enter the DDX/driver layers [03:13:04] <clee> because, you know, in 1986 somebody tried it and it didn't show any significant benefits then. so obviously that's still true today. [03:13:07] <clee> ;) [03:13:09] <gdamore> i'm not sure X should be MT-hot. [03:13:47] <alanc> I've still got a copy of the "MultiThreaded X11R5" research paper about how all the threads just hit the hardware lock [03:13:50] <clee> it would be nice if input was in a separate thread. [03:13:51] <gdamore> if you think about it, the typical X server has only one or two programs that are doing much IO at any given point in time. [03:14:09] <gdamore> s/IO/screen IO/ [03:14:32] <alanc> of course, 99.9% of desktops back then were single cpu/core, so multithreading wouldn't help much [03:14:41] <clee> alanc: exactly [03:14:46] <clee> the hardware landscape has changed significantly [03:15:08] <gdamore> even so: how much time do you think your CPU spends in the X server? my guess is not much. [03:15:17] <gdamore> even less so with modern systems handling a lot of the acceleration. [03:15:18] <clee> gdamore: depends on how well your driver accelerates things. [03:15:33] <alanc> and I wonder if composite managers change the software landscape too - if all the programs render to offscreen pixmaps, and only one is writing to actual screen... [03:15:37] <clee> gdamore: on a modern Athlon64 system, using 'nv', simply scrolling web pages that have large images is still enough to bring the system to its knees [03:15:37] <astinus> 4133 root 5 -10 402m 297m 8156 S 0.3 14.7 496:17.26 Xorg [03:15:49] <clee> (X server spikes to 99% CPU usage, stays there until the scrolling stops) [03:16:02] <astinus> I'd argue that 496:17.26 is quite a bit of time :X [03:16:02] <gdamore> clee: that's bad. a blitter should do that smoothly in hardwar.e [03:16:14] <astinus> I have big issues with multiple nVidia cards :/ [03:16:27] <clee> gdamore: switching to 'nouveau' makes everything happy :) [03:16:27] <alanc> Red Hat's use of a font server as your only source of fonts is effectively splitting font handling into it's own thread [03:16:32] <gisburn> BTW: Konqueror is multithreaded [03:16:35] <clee> (this is on Linux, of course.) [03:17:10] <gdamore> X clients can be MT. that makes sense. E.g. each firefox tab having its own thread, or doing mt downloads. but screen rendering? I'm not sure. [03:17:29] <clee> gdamore: might be nice to have a separate thread per card or per screen [03:17:30] * clee shrugs [03:17:36] <clee> won't know until somebody does it and benchmarks it [03:17:45] <gdamore> clee: run multiple servers and use Xinerama. Problem solved. [03:17:50] <clee> gdamore: true. [03:17:55] <clee> (except that Xinerama is ass.) [03:17:58] <alanc> the biggest advantage of MT-ing I/O is cleaning up the huge spaghetti code mess that is the Xserver's select loop [03:18:12] <gdamore> i agree with that statement. [03:18:21] <clee> alanc: a few people were talking about writing an input daemon. [03:18:24] <alanc> SPARC OpenGL actually uses a thread-per-head to render in Xinerama mode [03:18:25] <clee> I think krh was one of them. [03:18:26] <gdamore> i.e. cleaning up the code would be nice. but there might not be a performance improvement. [03:18:50] * gdamore is still sore that SPARC OpenGL is 100% closed. [03:19:04] <gdamore> even for ancient hardware like creator graphics boards. [03:19:36] <gdamore> and by 100% closed, not just the drivers, but even information so that someone else could build say, a radeon driver, is not available. [03:19:55] <alanc> pick up your spare change from under the couch cushions, buy SGI, and release their sources then [03:20:00] <clee> hahahaha [03:20:08] <clee> alanc: why haven't you guys bought them yet? [03:20:15] *** darkcmd has joined #opensolaris [03:20:18] <gdamore> alanc: where do I sign up? :-) [03:20:21] <alanc> SPARC code is encumbered [03:20:24] <jbk> do they have anything worth buying still? [03:20:24] <alanc> by that [03:20:51] <jbk> or owning [03:20:54] <gdamore> alanc: yes, but you can't even get _specs_ from the group, even under NDA! [03:20:56] <clee> jbk: a lot of OpenGL-related code/specs/IP [03:21:02] <gnu2it2> is using the SUNWjet package ok for current jumpstart or is there a better wat? [03:21:07] <jbk> i thought they sold that stuff to M$ [03:21:12] <clee> they sold a license to it to MS [03:21:15] <clee> (IIRC) [03:21:35] <alanc> the problem with buying SGI is you get all the IRIX support contracts and the cost of fulfilling them [03:21:48] <gdamore> SGI sold their most important division to Sun a long time ago. It resulted in the E10k, and landed Sun in the data center. [03:21:57] <gdamore> Sun should just complete the acquisition. :-) [03:22:17] <gisburn> gdamore: that was Cray, not SGI [03:22:31] <jbk> did sgi buy cray before or after sgi bought them though? [03:22:34] <gdamore> actually, it was SGI's Cray acquisition. [03:22:40] <darkcmd> is solaris on x86 becoming just as stable to use as on sparc? [03:22:49] <alanc> SGI bought Cray and immediately sold Cray's SPARC division to Sun [03:22:56] <jbk> darkcmd: it's almost all the same code [03:22:57] <gdamore> I.e. the "BSD" (Business Systems Division) of Cray was sold to Sun by SGI after SGI bought Cray [03:23:10] <alanc> since SGI didn't want to be in the business of shipping SPARC/Solaris machines [03:23:14] <gdamore> the E10k group were SGI employees for ~1 day [03:23:15] <astinus> darkcmd: 'becoming' .. its pretty much the same in my (admittedly miniscule) experience :) [03:23:28] <darkcmd> jbk, how is almost all the same code? [03:23:52] <gdamore> hmm... netbsd folks looking for gem sources.... too bad. [03:24:14] <jbk> i mean, that aside from some low level stuff specific to the platform, it's the same thing running on either, so the stability should be similar [03:24:45] <jbk> the biggest thing is just for a long time, x86 lacked adequate driver support, but that's been changing for the better (quite a bit) [03:25:08] <darkcmd> is there any isa support? [03:25:38] <delewis> I hope not. [03:25:41] <gdamore> there used to be. [03:25:48] <darkcmd> what happened to it? [03:25:50] <delewis> if you still have ISA devices that you regular use, I'm scared. [03:25:53] <gdamore> we have isa in some sparc laptops. :-) [03:25:54] <delewis> regularly* [03:25:59] <delewis> gdamore: interesting [03:26:07] <gdamore> can you say ebus? [03:26:49] <darkcmd> delewis, I have an ISA SCSI card [03:27:02] <astinus> delewis: time to get scared :P [03:27:09] <gdamore> yeah, but can you put it in a system that runs solaris? [03:27:16] <darkcmd> yeah [03:27:19] <gdamore> i.e. Solaris now requires a 486, IIRC. [03:27:24] <darkcmd> 2.8 MHz P4 [03:27:26] <gisburn> 586 [03:27:34] <delewis> somehow [03:27:41] <darkcmd> it has two ISA slots, I have 2 scsi drives in the machine [03:27:44] <darkcmd> and 1 sata [03:27:45] <gdamore> wow. [03:27:50] <delewis> running an ISA SCSI card in a "somewhat" modern Intel architecture seems.. wrong. [03:27:56] <delewis> [sic] [03:28:01] <darkcmd> it was free [03:28:04] <darkcmd> so i used it [03:28:07] <delewis> darkcmd: rightly so. [03:28:09] <gdamore> those ISA SCSI cards were mostly for CDROMs, IIRC. :-) [03:28:12] <darkcmd> the drives are 20gb a piece [03:28:37] <rydis> You can probably find a decentish PCI SCSI card for ten bucks. [03:28:39] <alanc> gdamore: Solaris 10 and later require at least a Pentium [03:28:48] <rydis> (Used, that is. New, probably not.) [03:28:56] <gdamore> alanc: Not surprised. Too bad though -- that rules out AMD Geode, IIUC. [03:29:01] <delewis> uh. you can even by FC HBAs for $20, nowadays. [03:29:13] <delewis> finding a POS PCI SCSI card should be easy. [03:29:26] <g4lt-U60> gdamore, one word: AHA1542CF [03:29:27] <alanc> guess we won't run on the OLPC laptops then [03:29:31] <darkcmd> well, I'm getting an Ultra-320 card soon, as well as two 18gb seagate cheatahs [03:29:53] <gdamore> alanc: no loss, there. :-) the OLPC people are _exclusively_ interested in Linux. [03:30:03] <darkcmd> that should be decent, don't you agree delewis? [03:30:04] <astinus> I have a pair of 73GB cheetahs on an Adaptec U320 card, and highly recommend 'em [03:30:07] <jamesd> darkcmd, sounds like massive overkilll [03:30:19] <darkcmd> jamesd, my buddy is giving them to me used [03:30:21] <delewis> darkcmd: you'd be fine with Ultra160 in any case [03:30:45] <astinus> and you can get a decent U160 card on eBay for peanuts [03:30:52] <astinus> look for an Adaptec 19160, 29160 or 39160 [03:30:55] <gdamore> nobody has yet told me if the Via C3 onboard SATA is supported directly or only in PATA mode yet [03:31:10] <darkcmd> would they be supported under x86 solaris? [03:31:49] <delewis> darkcmd: yes [03:31:57] <delewis> don't try throwing them into a SPARC, though. [03:31:58] <darkcmd> cool [03:32:03] <astinus> yeah, I think Adaptec SCSI card support is excellent under x86 [03:32:09] <darkcmd> delewis, I also have an Ultra 2 workstation [03:32:12] <darkcmd> with 2GB of ram [03:32:30] * astinus thinks that's akin to squashing a cockroach with a nuclear weapon [03:32:33] <darkcmd> dual 400mhz ultrasparc-II cpus [03:32:56] <delewis> darkcmd: why are you telling me this? [03:33:17] *** fik has joined #opensolaris [03:33:18] <darkcmd> sorry, I'm excited about getting it, it was a birthday gift [03:33:32] <gdamore> astinus: nuke it from orbit. Its the only way to be sure. [03:34:14] <astinus> gdamore: "What should you do if you accidentally run over a chav while driving? .... Reverse over him a few times to make sure :P" [03:34:33] <gdamore> chav? [03:34:42] <astinus> http://en.wikipedia.org/wiki/Chav [03:34:44] <astinus> "Bling Bling" [03:35:27] <gdamore> ah. [03:35:42] * astinus smiles [03:37:43] <jbk> much like you'd do to a lawyer [03:37:48] <jbk> +what [03:38:54] <steleman> evening [03:39:35] <gdamore> anyone here have any idea what it would take to open up the GEM driver? [03:40:09] <gdamore> apparently there are some sbus details that are not "common" knowledge (i.e. Linux lacks support for GEM on sbus) [03:44:31] * boyd grumbles about more internal URLs appearing on the OpenSolaris site [03:46:01] *** dmick has joined #opensolaris [03:47:24] <LeftWing> delewis: He's probably telling you because he can sense you have an Ultra 2. ;P [03:47:47] <delewis> LeftWing: :-) [03:48:00] <delewis> mine only has 2x300MHz procs and 1.5GB of memory, though :-( [03:48:14] <delewis> 2x400MHz procs are quite rare. [03:48:17] <alanc> I thought SBus was an IEEE standard? [03:48:56] <LeftWing> delewis: What form factor are the CPUs in for the U2? [03:49:26] <darkcmd> MBus [03:49:29] <delewis> LeftWing: slot-based, same procs the U10, U5, U60 use [03:49:32] <darkcmd> I think, they're slot [03:49:44] <LeftWing> Are they US-II or US-IIi? [03:49:50] <delewis> US-II [03:49:55] <delewis> 2MB cache/processor [03:49:56] <LeftWing> Arr. [03:50:01] <darkcmd> it's a pretty nice computer [03:50:14] <LeftWing> I have US-IIi's w/ 2MB cache. =) [03:50:16] <LeftWing> 440Mhz [03:50:25] <delewis> unfortunately, there's not a lot of graphics options for the Ultra 2 [03:50:34] <delewis> you can throw a Creator in it [03:50:41] <LeftWing> A range of cards that cost more than the computer will cost you? =P [03:50:54] <delewis> the only SBus framebuffer I happen to have around at the moment is a cg6 (TGX, at that) [03:51:09] <delewis> LeftWing: nah, Creator cards are cheap, nowadays. [03:51:13] <delewis> I just haven't bothered with one. [03:51:19] <LeftWing> Aye. [03:51:22] <delewis> I'm still using my Blade 1000 as my workstation [03:51:36] <LeftWing> A Sun Ray is probably one of the better frame buffers you can get for an Ultra 2 anyhow. =P [03:52:43] <astinus> hah [03:52:46] <astinus> lu just segfaulted [03:52:47] <g4lt-U60> didn't U2's have horizontal UPA? [03:52:47] <astinus> genius. [03:53:53] <jbk> heh [03:54:06] <jbk> i can make the standard solaris installer segfault pretty easily [03:54:20] <jbk> or at the very least, act *really* strange [03:54:55] <LeftWing> The Solaris 10 installer has been pretty good for me, actually. Not that I use anything but JumpStart these days. [03:55:23] <jbk> try doing a flash archive install booted off cdrom over a network with a duplex mismatch :) [03:56:34] <jbk> though learning rm /tmp/.jump* as proven useful :) [03:56:43] <LeftWing> heh [03:56:56] <LeftWing> I don't think that problems with that scenario are entirely the installer's fault. =P [03:57:08] <jbk> our network group still lives in the stone age and disables autonegotiation on all 100mb ports [03:57:21] <LeftWing> ... [03:57:33] <jbk> so if you don't learn how to exit, force the mode, rm /tmp/.jump*; restart the installer.. things can prove frustrating [03:58:05] <jbk> well they'll hopefully go away [03:58:30] <jbk> due to an unrelated incident involving this policy, our director finally was able to get them to relent [03:58:40] <astinus> fmli seems to be the cause of my segfault [03:58:47] <astinus> it just infinitely loops [03:58:56] <astinus> totally pwns the CPU until you pkill it [03:59:11] *** dmesg has joined #opensolaris [04:00:05] <dmesg> hi [04:00:29] <dmesg> does opensolaris is similar to solaris 9 or 10? [04:00:49] <delewis> dmesg: the code bases are separate.. [04:01:03] <delewis> features integrated into OpenSolaris occasionally get backported to 10 [04:01:48] <alanc> opensolaris is the source code to the version of solaris after 10, which has not yet been named [04:01:51] <dmesg> ahh :) [04:02:18] <dmick> fmli?? [04:02:18] <dmick> wow. [04:02:23] <astinus> alanc: You guys should call it Solaris XI to piss off the guys at Apple ;) [04:02:29] <jbk> though they don't rewrite everything from scratch for every release, so it can sometimes still be useful for figureing things out with older versions, just no guarantees [04:02:40] <dmesg> do u think opensolaris can suport an sun enterprise 450? [04:02:40] <dmick> astinus: Solaris Vista :) [04:02:48] <Error_404> dmesg, yeah, why not? [04:02:51] <astinus> dmick: I fixed my issue by using lucreate manually *shrug* [04:02:52] <delewis> dmesg: I should certainly hope so. [04:02:57] <boyd> They could call it Sun Java Enterprise Platform Environment to piss of everyone [04:02:58] <Error_404> i've had it on an ultra2 [04:03:01] <boyd> s/of/off [04:03:01] <jbk> i did that to figure out an rather annoying interaction on solaris 10 with ldap + nss [04:03:10] <astinus> dmick: But yeah, something about my setup the more intuitive 'lu' doesn't like much :( [04:03:18] <jbk> dmesg: should be able to [04:03:21] <dmick> I'm surprised it uses fmli [04:03:25] <dmesg> Error_404 delewis i will try then :) [04:03:34] <dmesg> jbk k >( [04:03:35] <dmesg> :) [04:03:38] <alanc> lu is about the only reason fmli hasn't been removed from Solaris [04:03:57] <dmick> I should probably run lu once in my life [04:04:05] <alanc> even "faces" is gone now [04:04:06] <astinus> dmick: well, telltale signs were I tried lu three times .. then heard my fans spin up, so I checked load on the box, and there were three fmli processes eating 33% each and really raping the server :X [04:04:29] <alanc> I tried it once, went back to using the lu* command line tools instead [04:04:30] <astinus> alanc: Isn't lu about due for an overhaul? [04:04:49] <dmick> heh [04:04:51] <alanc> install group is busy overhauling the entire Solaris install system [04:04:52] <dmick> I guess this would be a clue: [04:04:56] <dmick> NAME [04:04:57] <dmick> lu - FMLI-based interface to Live Upgrade functions [04:05:09] <astinus> alanc: Is this part of the move to support ZFS as rootfs? [04:05:16] *** Tpenta has quit IRC [04:05:23] <boyd> Interesting to see that there are no plans to support in-place upgrades in future [04:05:35] <alanc> no, it's part of the move to make Solaris installable without requiring years of experience with Solaris first [04:05:38] <boyd> astinus: Separate, but concurrent [04:05:51] <astinus> alanc: Hah, years of experience .. I managed (just about!) [04:06:07] <astinus> alanc: I've only been using it for 3-4 weeks since switching from Linux... [04:06:11] <alanc> see the install community on opensolaris.org for the current proposal [04:06:13] <jbk> boyd: well once root is on zfs, i don't think it should be that big of an issue [04:06:26] <dmick> actually i didn't even know the 'lu' command existed [04:06:30] <boyd> jbk: I agree [04:06:31] <dmick> I've only ever used lucreate et all [04:06:34] <dmick> er, et al [04:06:36] <jbk> heh [04:06:42] <jbk> all our docs still reference lu [04:06:49] <jbk> we're gonna need to switch to the cmdline [04:06:50] <astinus> dmick: This is my first live upgrade tonight :) I'm going from b48 to b51 [04:07:01] <dmick> astinus: and god speed [04:07:03] <jbk> which actually shouldn't be too bad.. [04:07:28] <jbk> i found lu to be useful to work around our lack of infrastructure.. if they can get pfinstall fixed [04:07:33] <dmesg> is secure to use opensolaris for a webserver? :) [04:07:49] <astinus> dmick: Well, I actually have questions for you guys :) I accidentally managed to hose sshd and NFS when I removed some packages .. is there any way I can get a 'factory' install of b51 into my other slice, but keep settings and stuff? [04:08:12] <jbk> live upgrade should manage most of that i think [04:08:26] <jbk> and you can mount up the new environment from the old to manually copy stuff over if needed [04:08:27] <astinus> that'd be nice :) [04:08:32] * astinus nods [04:08:41] <jbk> for things that lu doesn't know about [04:08:46] <jbk> it works pretty well [04:08:55] <astinus> It was mostly just lack of experience, I wanted to close down some services and remove unnecessary packages - but I didn't check dependencies properly, so managed to hose quite a few things :X [04:08:59] <jbk> we used it to move 500 boxes from 2.6->8 with few issues [04:09:05] <jbk> way back in the day [04:09:28] <astinus> But yeah, I *really* like ZFS and the whole concept behind Live Upgrade [04:09:29] <jbk> the other groups were doing full reinstalls... we just laughed :P [04:09:42] <alanc> recent opensolaris turns off all remote ports except ssh by default [04:09:52] <astinus> Makes me seriously wish I'd found Solaris about 3 years ago :/ [04:10:07] <alanc> we weren't trying to hide it! 8-) [04:10:16] <astinus> Lots of things really do come across as "Linux done properly" if you know what I mean... [04:10:28] <dmick> astinus: listen to these guys, they've got 100% more experience with LU than me [04:10:36] <LeftWing> boyd: No plans to support in-place upgrades in the future? [04:10:47] <dmick> astinus: preach it brother [04:10:59] <boyd> LeftWing: With the new installation architecture, yes [04:11:10] <astinus> alanc: Yeah, I know :) But I'm still a young 'un - finding out this stuff is an education unto itself :) [04:11:21] <LeftWing> boyd: in-place being boot from CD and upgrade over your existing install? [04:11:28] <boyd> LeftWing: correct [04:11:33] <LeftWing> Interesting. [04:11:38] <boyd> vs LU style [04:11:44] <LeftWing> Aye. [04:12:02] <dmesg> is posible to use pkg-get on opensolaris? [04:12:09] <astinus> presumably I can mount my b51 ISOs and luupgrade without burning them? [04:12:16] <astinus> dmesg: yes, if you install it [04:12:35] <boyd> astinus: Yes [04:12:47] <dmesg> astinus ah ok thanks :) [04:12:56] <boyd> astinus: CD or DVD (since you say "ISOs" I assume CD) [04:13:13] <astinus> boyd: Either really, I'm about to grab 'em [04:13:45] <boyd> astinus: If you're not going to burn them get the DVD... no simulated swapping reqd [04:13:50] * astinus nods [04:13:55] <astinus> Good advice :) [04:15:07] *** fik has quit IRC [04:15:25] <boyd> Oh, and rather than the slow unzip, unzip, unzip, cat thing I recommend something like this: [04:15:29] <boyd> for file in sol-nv-b49-x86-dvd-iso-* do [04:15:32] <boyd> print Doing $file [04:15:35] <boyd> unzip -p $file >> sol-nv-b49-x86-dvd.iso [04:15:36] <boyd> done [04:15:46] <boyd> (ksh or zsh or the print will fail) [04:16:03] <astinus> hah, downloading at 4500kB/second [04:16:16] <astinus> looks like someone upgraded the mirror's bandwidth :P [04:16:16] <boyd> Since it saves the intermediate unpacked segments [04:16:35] <boyd> 4.5MB/s! sheesh [04:16:51] <Error_404> i don't even get that on my local network [04:17:26] <jbk> i was getting close to that at work [04:17:34] <jbk> well after i had upgraded [04:17:39] <astinus> Error_404: I can push 90-95MB/s on my local network, between workstation and NFS fileserver [04:17:41] <jbk> at the time i had a sb100 & sb150 [04:18:10] <jbk> and i had solaris 9 on the 150, and had put solaris express (before 10 was out) on the 100 [04:18:17] <jbk> then wanted to update the 100 [04:18:43] <jbk> on s9, i was getting about 500kb/s [04:18:47] <Error_404> astinus, my network sucks... [04:18:50] <Error_404> 10/100 ftl [04:18:57] <jbk> on the 150, i was getting around 1000kb/s [04:19:06] <jbk> on each segment [04:19:16] <jbk> downloading all 5 segments simultaneously [04:19:21] <jbk> i was impressed :) [04:19:34] <astinus> Error_404: 10/100/1000 here, although my fileserver has 2 * Gig-E links [04:21:26] *** knightblader has joined #opensolaris [04:21:37] <astinus> i was watching our university SAN replicate the other day :O [04:22:05] <astinus> it was 2 * 2Gbit fibre connections linking it to its mirror on the other side of campus [04:22:30] <astinus> that was rather scary, it was pushing >300MB/second [04:23:47] <dmick> fast nets are awe-inspiring [04:23:51] <dmesg> to install opensolaris in a ultrasparc, martux is ok? [04:24:02] <astinus> dmick: yeah, no kidding :X [04:24:13] <delewis> dmesg: or Solaris Express [04:24:24] <dmesg> delewis ok :) [04:24:25] <astinus> I was just thinking back 10 years when my hard drive wasn't 300MB :S [04:25:29] <gisburn> ok, I hate this. Sun has all the resources we need but can't share it. F*CK. [04:25:49] <dmick> gisburn: and a good evening to you too [04:26:25] <gisburn> dmick: good evening. [04:26:47] <gisburn> I just figured out that I need around 1200 euro for a machine from ebay [04:28:34] <astinus> now >300GB is common, its definitely Moore's Law [04:28:51] <astinus> boyd: I think that unzip snippet above works great - they should recommend it :) [04:28:53] <astinus> boyd: In one line though, why would you recommend ksh/zsh over bash? [04:29:07] <astinus> gisburn: upgrading? [04:29:19] <boyd> I wouldn't recommend ksh but zsh [04:29:50] <gisburn> astinus: no, I just want to finish my small little opensolaris.org project and don't have a decent sparc to get it done. [04:29:53] <alanc> oh no, now he's done it! [04:29:56] <boyd> For me it's all about completion, a few shell options and interactive features [04:30:24] <astinus> alanc: shell wars? :P [04:30:25] * astinus hides [04:30:32] <boyd> astinus: Try this: man bash | wc -l; man zshall | wc -l [04:30:37] <boyd> All of this is IMHO [04:30:52] *** alanc is now known as alanc-away [04:31:00] <dmick> boyd: is more lines or fewer lines better? [04:31:01] <jbk> so... should echo allow any commandline arguments? [04:31:03] * jbk runs [04:31:22] <boyd> dmick: Depends on your POV I suppose... [04:31:27] * dmick points at his nose [04:31:43] <jbk> heh [04:31:58] <boyd> If you're Melanie Griffith I suspect fewer lines is better [04:32:19] * dmick hisses [04:32:39] <dmick> so you're saying zsh has trout lips? [04:32:43] <boyd> Hehe [04:32:56] <jbk> someone once told me that one of IBM's selling points for AIX was that they took the AT&T source and like double the size of the codebase :) [04:33:11] <dmick> it goes to 11 [04:33:23] <boyd> You mean by changing the font size? [04:33:24] <boyd> :) [04:37:33] *** salamanders has quit IRC [04:40:47] *** dmesg has quit IRC [04:52:27] <elektronkind> I can easily double the size of any code base [04:52:49] <elektronkind> whether that new code does anything useful is up for debate [04:53:42] <dmick> details [04:55:15] <Error_404> page after page of if (1) {} [05:04:14] *** darkcmd has quit IRC [05:17:39] *** sahafeez has joined #opensolaris [05:35:35] *** edp has joined #opensolaris [05:43:02] *** OnkelSchorsch has joined #opensolaris [05:56:34] *** OnkelSchorsch_ has joined #opensolaris [05:59:03] *** mercykiller has joined #opensolaris [06:05:07] *** OnkelSchorsch has quit IRC [06:08:10] *** [1]Cyl has joined #opensolaris [06:12:00] *** mercykiller_ has quit IRC [06:16:27] <sahafeez> hey, i am trying to boot b51 on a p3. i get to the choice about install method. no matter what i get after that i get "ERROR: Cannot find windowing software, Exiting to shell" [06:16:59] *** Cyl has quit IRC [06:17:00] *** [1]Cyl is now known as Cyl [06:18:56] <Triskelios> sahafeez: bad burn? md5sum the CD/DVD [06:19:07] <sahafeez> did that. [06:19:10] <sahafeez> no issues [06:19:17] <Triskelios> weird [06:19:36] <sahafeez> yup. [06:19:51] <Triskelios> I've only seen that on Ultra5/10s which don't quite support installing from DVDs properly [06:20:06] <dmick> can't you select something to install in text mode so it doesn't try to look for windowing sw? [06:20:11] <dmick> or does it do it even then? [06:20:19] <Triskelios> it's probably not going to find the packages anyway [06:20:40] <sahafeez> i get it on cd1 or dvd [06:20:46] <Triskelios> oh, weird [06:20:50] <sahafeez> did the textmode same ms [06:20:52] <sahafeez> msg [06:20:53] <Triskelios> never got it on a CD [06:24:00] *** darkcmd has joined #opensolaris [06:26:55] *** LeftWing has quit IRC [06:28:29] *** gisburn has quit IRC [06:29:39] <sahafeez> is there a quick simple howto on jumpstart [06:30:06] <Error_404> heh, w00t. [06:30:18] <Error_404> I conned a guy in to helping w/ polaris [06:31:30] *** LeftWing has joined #OpenSolaris [06:31:31] <OnkelSchorsch_> sahafeez, maybe this will get you started. there's a german and an english version http://wiki.c0t0d0s0.org/index.php/Hauptseite [06:31:42] <sahafeez> thanks. [06:31:49] <sahafeez> looking. [06:32:01] <OnkelSchorsch_> np. it's sparc centered though. [06:32:42] <Error_404> that's quite the convoluted URL [06:33:27] *** AbeFroman has quit IRC [06:34:24] *** AbeFroman has joined #opensolaris [06:35:36] <OnkelSchorsch_> hehe. yep. but I like it. [06:36:19] *** sahafeez has quit IRC [06:36:24] <Error_404> me too [06:37:09] *** sahafeez has joined #opensolaris [06:39:17] *** sparkleytone has quit IRC [06:39:28] <OnkelSchorsch_> sahafeez, quasi blogged about jumpstarting x86 boxes. haven't tried that though http://soulfood.dk/archives/2006/09/20/T23_43_37/index.html [06:39:43] <sahafeez> thanks. [06:40:03] *** delewis_ has joined #opensolaris [06:41:31] *** delewis has quit IRC [06:41:35] *** sparkleytone has joined #opensolaris [06:43:21] <sahafeez> that works! setting up now. [06:47:11] *** deedaw has quit IRC [06:59:49] *** oooo has quit IRC [07:23:52] *** Jiko has joined #opensolaris [07:37:04] *** dmick has left #opensolaris [07:50:30] *** jpipkin has joined #opensolaris [07:57:16] *** jpipkin has left #opensolaris [08:11:43] *** karrotx has quit IRC [08:12:37] *** karrotx has joined #opensolaris [08:47:28] *** slowhog has left #opensolaris [08:51:52] <quasi> yeah, jumpstarting usually does work ;) [09:09:04] <Error_404> where the blazes is someone who works on polaris, this is silly already [09:20:55] <sahafeez> anymore good links of jumpstarting.. [09:24:44] *** sahafeez has quit IRC [09:31:32] <Burana> can i jumpstart an x86 box without pxe/dhcp? by inserting dvd and then choosing custom jumpstart? [09:34:35] <OnkelSchorsch_> "Net install: The standard procedure for setting up net install images remains the same. Clients are assumed to boot via the Preboot eXecution Environment (PXE) mechanism. Clients not capable of PXE boot can use a GRUB floppy" [09:34:37] <OnkelSchorsch_> http://www.sun.com/bigadmin/features/articles/grub_boot_solaris.html [09:34:52] <OnkelSchorsch_> you may also use an install dvd/cd [09:38:22] *** Ezechial has joined #opensolaris [09:44:32] *** miffe has joined #opensolaris [09:50:09] <Burana> is it possible to move the dvd image to an usb stick and boot from there? [10:02:54] *** sil3nt|warri0r has joined #opensolaris [10:06:45] *** uebayasi has quit IRC [10:10:06] *** GmanAFK has quit IRC [10:10:14] *** alanc_away has quit IRC [10:10:58] *** GmanAFK has joined #opensolaris [10:13:10] *** alanc has joined #opensolaris [10:13:57] *** miffe_ has joined #opensolaris [10:18:56] *** miffe_ has left #opensolaris [10:23:47] *** bougie has joined #opensolaris [10:31:21] <bougie> hello :p [10:32:17] <Error_404> yo [10:33:48] <sickness> morning all [10:34:18] *** _schily__ has joined #opensolaris [10:35:51] <quasi> hey hey [10:36:02] <asyd> hello quasi [10:36:07] <quasi> hey asyd [10:45:37] *** Fish- has joined #opensolaris [10:45:46] *** schily___ has quit IRC [10:46:54] <Fish-> hello [10:52:29] *** axxl has joined #opensolaris [10:54:21] *** neoxed has quit IRC [10:55:20] *** neoxed has joined #OpenSolaris [11:09:36] *** Fish- is now known as Fish [11:11:59] *** OnkelSchorsch_ has quit IRC [11:30:31] *** logic has joined #opensolaris [11:33:32] *** mercykiller has quit IRC [11:44:37] *** logic_ has quit IRC [11:50:29] <lasseoe> ld.so.1: useradd: fatal: libc.so.1: version `SUNW_1.22.2' not found (required by file /usr/sbin/useradd) [11:50:36] <lasseoe> on S10U2 with all the latest patches [11:50:39] <lasseoe> weird [11:50:42] *** mercykiller has joined #opensolaris [11:51:02] <g4lt-U60> lasseoe, yeah, it's a known issue with NV51 [11:51:02] <boyd> someone else had that prob a day or so ago... just a minute [11:51:13] <boyd> Oh, nm [11:51:49] <Triskelios> g4lt-U60: that's not b51, though [11:51:58] <Triskelios> it's a hugely important difference [11:52:08] <lasseoe> Triskelios: indeed [11:53:23] <lasseoe> I'm guessing some patch broke it :-/ [11:54:14] <lasseoe> # ldd /usr/sbin/useradd [11:54:19] <lasseoe> libc.so.1 => /lib/libc.so.1 [11:54:19] <lasseoe> libc.so.1 (SUNW_1.22.2) => (version not found) [11:55:52] <quasi> ouch [11:56:32] <asyd> ah yeah, a friend of mine had the same problem with useradd [11:57:28] <Triskelios> shouldn't this affect a lot more binaries? I would think useradd is rebuilt as often as any other important things are [11:58:04] <lasseoe> Triskelios: so far everything else works [11:58:10] <quasi> some bad patches are slipping through the cracks lately - I had 122663-06 (x86) breaking zones recently [11:58:59] <lasseoe> /usr/sbin/role* doesn't work either [11:59:17] <boyd> they are hardlinks to user* [11:59:28] <lasseoe> ah didn't notice :) [12:07:12] *** mercykiller has quit IRC [12:21:52] *** miffe_ has joined #opensolaris [12:30:42] *** mercykiller has joined #opensolaris [12:32:42] *** LordKing has joined #opensolaris [12:34:37] *** miffe_ has quit IRC [12:35:25] *** miffe has quit IRC [12:38:26] *** miffe has joined #opensolaris [12:38:51] *** bunker has joined #opensolaris [12:39:00] *** miffe has quit IRC [12:39:09] *** miffe has joined #opensolaris [12:43:23] <Burana> Hi Miffe. How popular is solaris in finland at the moment? [12:49:13] *** mercykiller has quit IRC [12:50:42] *** mercykiller has joined #opensolaris [12:52:42] <sickness> i'm back [12:56:07] *** GmanAFK has quit IRC [13:07:07] *** mercykiller has quit IRC [13:10:46] *** mercykiller has joined #opensolaris [13:15:34] *** mlh has quit IRC [13:20:39] *** laca has quit IRC [13:26:25] *** mlh has joined #opensolaris [13:27:00] *** deedaw has joined #opensolaris [13:29:12] *** mercykiller has quit IRC [13:30:25] *** OnkelSchorsch has joined #opensolaris [13:30:47] *** mercykiller has joined #opensolaris [13:32:00] *** Burana has quit IRC [13:32:01] *** Burana has joined #opensolaris [13:40:38] *** Burana has quit IRC [13:41:22] *** soultan has quit IRC [13:41:27] *** soultan has joined #opensolaris [13:43:51] <myrkraverk> I'm back too ;) [13:47:01] *** piwi has joined #opensolaris [13:49:10] *** mercykiller has quit IRC [13:50:01] *** kill9 has quit IRC [13:50:46] *** mercykiller has joined #opensolaris [13:51:07] *** kill9 has joined #opensolaris [13:51:59] *** DataStream has quit IRC [13:55:27] *** uebayasi has joined #opensolaris [13:57:34] *** miffe has quit IRC [14:01:14] *** DataStream has joined #opensolaris [14:09:19] *** mercykiller has quit IRC [14:25:12] *** rodrickbrown has quit IRC [14:33:10] *** oxygene has joined #opensolaris [14:34:33] *** alanc_ has joined #opensolaris [14:34:37] *** alanc_work has joined #opensolaris [14:34:49] *** jacotton_ has joined #opensolaris [14:34:58] *** linma` has joined #opensolaris [14:35:11] *** stevel- has joined #opensolaris [14:35:11] *** ChanServ sets mode: +o stevel- [14:35:56] *** dduvall_ has joined #opensolaris [14:37:13] *** sickness has quit IRC [14:37:35] *** axxl has quit IRC [14:39:04] *** alanc-away has quit IRC [14:39:04] *** alanc_work has quit IRC [14:42:00] *** dduvall has quit IRC [14:42:06] *** dduvall_ is now known as dduvall [14:42:51] *** jacotton has quit IRC [14:42:54] *** alanc has quit IRC [14:43:05] *** stevel has quit IRC [14:43:18] *** alanc_ has quit IRC [14:44:19] *** sickness has joined #opensolaris [14:45:50] *** estibi has joined #opensolaris [14:48:04] *** smoco has joined #opensolaris [14:50:51] *** linma has quit IRC [14:52:34] *** sickness has quit IRC [14:53:15] *** sickness has joined #opensolaris [15:00:28] *** sickness has quit IRC [15:09:09] *** smoco has left #opensolaris [15:17:01] *** pikapika has quit IRC [15:23:43] *** kill9 has quit IRC [15:26:25] *** Somethingelse has joined #opensolaris [15:27:27] *** oxygene has quit IRC [15:29:31] *** bougie has quit IRC [15:30:59] *** loke has joined #opensolaris [15:38:09] *** delewis_ is now known as delewis [15:40:22] *** kill9 has joined #opensolaris [15:53:58] *** loke has quit IRC [15:55:42] *** bougie has joined #opensolaris [16:01:32] *** DataStre1m has joined #opensolaris [16:04:00] *** DataStream has quit IRC [16:04:18] *** PosixC has joined #opensolaris [16:05:22] <PosixC> I tried to prepare an install server and ran "boot net - install" from a T2000 machine and it hangs [16:05:40] <PosixC> is there a way to pull the DVD drive from t2000? [16:06:03] <quasi> pull the dvd drive? [16:06:04] <PosixC> It seems to me that there are problems with this DVD and I want to try another for installation [16:06:21] <lasseoe> it's probably your setup and not the dvd [16:06:27] <quasi> pretty much a standard drive [16:06:46] <PosixC> lasseoe, what do you mean "my setup"? [16:07:05] <lasseoe> your install server [16:07:16] <PosixC> lasseoe, this is a different problem [16:07:29] <PosixC> I tried the install server because the dvd drive is not ok [16:07:43] <lasseoe> well that wasn't obvious from what you wrote above [16:07:53] <PosixC> and causes panic when "boot cdrom" of installtaion [16:07:58] <PosixC> lasseoe, ok, sorry [16:08:38] <PosixC> so I want to try a dvd which I know that it works with other machines [16:09:08] <PosixC> but it seems to me that there is no easy way to pull it out [16:09:36] <PosixC> I though that there should be 2 or for screws which I should open but [16:09:45] <PosixC> it seems to me that it is not simple [16:09:48] <Doc> just disconnect the cable to it [16:09:53] <lasseoe> just use your install server, it's faster and better [16:10:38] *** onlinebacon has joined #opensolaris [16:10:43] <onlinebacon> hey [16:10:43] <PosixC> lasseoe, should there be any message logs in the install server which can help me debug the problem [16:10:57] *** DataStre1m has quit IRC [16:11:10] <onlinebacon> is it possible to be able to install solaris express community edition from 1 cd? [16:12:10] <onlinebacon> dw sorted it [16:12:14] *** onlinebacon has left #opensolaris [16:13:32] *** DataStream has joined #opensolaris [16:18:25] <PosixC> Doc , is the T2000 dvd is an ordinary IDE device (40 pins)? [16:19:02] <PosixC> I just don't have a comfortable access to it and I am afraid some damage will be done when trying to pull it out [16:19:59] <PosixC> are there anywhere on the install server logs which can help me detect the problem? I see that the T2000 send DHCP request for it's MAC address but does not get an answer [16:21:38] <PerterB> you could always use bootp rather than dhcp, then run in.rarpd and rpc.bootparamd in debug mode on the install server to see whatthey're receiving and sending [16:25:44] *** kill9 has quit IRC [16:26:23] <PosixC> PerterB, Thnks; [16:26:28] <PosixC> One last question: [16:26:37] <PosixC> I pulled the ide cable out [16:26:48] *** kill9 has joined #opensolaris [16:26:55] <PosixC> The DVD power in still connected [16:27:06] *** kill9 has quit IRC [16:27:11] <PosixC> suppose I will connect another DVD which has power from other source [16:27:26] <lasseoe> PosixC: which Solaris DVD is it? 3/05, HW1, 1/06 or 6/06 [16:27:47] <PosixC> will the fact that the original DVD is connected to power and not to IDE can cause any problems? [16:27:59] *** kill9 has joined #opensolaris [16:28:00] <PosixC> lasseoe, it is SXCR50 [16:28:26] <PosixC> Any ideas about my question regarding power? [16:29:09] *** g4lt-U60 has quit IRC [16:29:12] *** galt has joined #opensolaris [16:29:55] <PerterB> leaving the original dvd drive connected shouldn't be a problem, but plugging a dvd powered from another source into the IDE bus would worry me... If the T2000 and the other power source have different grounds then there might be an issue (but I am no electronics expert) [16:32:01] <PosixC> It could be ; I tried it with other NON T2000 machines and it was ok [16:32:34] <PosixC> Can the power chord be pulled out from the motheboard at all ? did anybody tried it? [16:34:40] <PerterB> never even seen the insides of a T2000 :) but a DVD drive should be a field replacable unit, so there must be a way... Is there maybe a hardware manual on docs.sun.com that explains how? [16:35:08] <PosixC> and the power connection to the DVD seems like "SATA" whereas my dvd ha usual wide power plug [16:36:07] *** polpol has joined #opensolaris [16:37:44] <OnkelSchorsch> did you check out the service manual? http://www.sun.com/products-n-solutions/hardware/docs/html/819-2548-12/ [16:37:58] <PosixC> looking now... [16:38:21] <OnkelSchorsch> there are instructions how to remove the dvd drive [16:38:33] <OnkelSchorsch> and of course other parts [16:39:15] <PosixC> thnks , I located it , and there are indeed instructions; might you know: is there a way to conenct non sata dvd drive to t2000? [16:42:49] *** cypromis has quit IRC [16:44:05] <lasseoe> There isn't [16:44:25] *** luxh has quit IRC [16:45:45] <OnkelSchorsch> maybe you can get some adapters. but there is no direct way [16:46:40] *** cypromis has joined #opensolaris [16:46:41] <lasseoe> I don't know why you won't fix your jumpstart server instead of messing around with DVD drives [16:49:41] *** Somethingelse has quit IRC [16:50:46] <PosixC> lasseoe, you mean debigging it with bootp rather than dhcp, as Peter suggested here before? [16:50:53] <lasseoe> yes [16:50:58] <lasseoe> did you even setup a DHCP server? [16:51:01] <PosixC> yes [16:51:15] <PosixC> I add a macro called T2000 with the required options [16:51:18] <lasseoe> ok, well RARP/BOOTP is a lot easier to use [16:51:22] <lasseoe> imho [16:51:29] <delewis> it is [16:51:40] <delewis> with DHCP you have to setup a macro [16:52:23] <PosixC> I assume there is no chance to tell niagra to boot from USB [16:52:30] <delewis> PosixC: no :-) [16:52:48] <PosixC> delewis, I had set up such macro named "T2000" with the options [16:53:01] <delewis> it does not want a macro called "T2000" [16:53:16] <delewis> if you had bothered to use snoop to see what the DHCP requests were like, you'd probably see "SUNW,T2000" [16:53:20] <PosixC> delewis, when I ran "add_client" [16:53:39] <PosixC> a moment [16:54:53] <PosixC> delewis, i ran it thus: ./add_install_client -d -s myInstallServer:/installServerSparc T2000 sun4v [16:55:00] <PosixC> and emitted some messages [16:55:18] <PosixC> isn't the T2000 a macro it is expecting after running this command? [16:55:43] <PosixC> because the message says: [16:55:44] *** Somethingelse has joined #opensolaris [16:55:44] <delewis> PosixC: use snoop -v dhcp to see [16:55:59] <delewis> most likely, it's SUNW,T2000 [16:56:13] <delewis> all of the Sun systems I've Jumpstarted request a prefix'd SUNW, [16:57:03] <PosixC> "to enable T2000 in the DHCP server,...add a macro to the server named T2000". [16:57:26] * delewis sigh [16:57:32] <delewis> if you don't follow my directions, I cannot help you [16:57:44] <PosixC> delewis, ok , I will try now [16:57:50] <delewis> you're probably feeding add_install_client the wrong parameter to begin with [16:58:18] <PosixC> I will run the same but with SUNW,T2000 instead of T2000 [16:58:27] <delewis> no, you meat-head. [16:58:32] <Somethingelse> hahaha [16:58:34] <astinus> w00t! I managed to get my first Live Upgrade done; well kinda... [16:58:35] <delewis> read the fucking snoop -v dhcp output if you want to be sure [16:58:36] <PosixC> ?? [16:58:43] <PosixC> ok [16:59:05] * lasseoe chuckles and gets some popcorn [16:59:30] <astinus> I've still got issues related to SUNWsshdr and SUNWsshdu - like the Live Upgrade didn't restore them. If I've got the DVD mounted to /mnt/cdrom, can I pkgadd them off it somehow? [17:06:06] *** sickness has joined #opensolaris [17:11:07] *** trs81 has quit IRC [17:18:59] *** boro has joined #opensolaris [17:24:17] *** ndroux has quit IRC [17:26:32] *** LordKing has quit IRC [17:27:14] *** cherry has joined #opensolaris [17:29:37] <myrkraverk> is it possible to dig up James Carlsons email somewhere? [17:30:48] <myrkraverk> or, can I translate b.s.c links to emails? [17:31:04] *** mercykiller has joined #opensolaris [17:31:09] <twincest> sun people are almost always firstname.lastname at sun dot com [17:32:31] *** cherry has quit IRC [17:32:38] <myrkraverk> twincest: ok - the (opensource) version of pppd mentions James as the Solaris maintainer, and that bug reports should go to him, but neglects an email, aparently ;/ [17:35:01] *** boro has quit IRC [17:36:20] <PosixC> delewis, are you listening? [17:37:15] <PosixC> anybody can help with the install from server problem I have ? [17:37:34] <quasi> myrkraverk: he's often mailing on networking-discuss - you should be able to find his addy there [17:37:53] <myrkraverk> quasi: ah, thanks [17:39:23] <lasseoe> posixc: google for solaris jumpstart, there are hundreds of tutorials [17:39:56] <PosixC> lasseoe, I had acted according to SUN tutorial [17:40:21] <quasi> http://jet.maui.co.uk/ has good info as well [17:40:32] <PosixC> I see in a sniffer in a machine near the T200 I see that the t200 sent [17:40:33] <PosixC> RARP request: RARP who is 00:03:ba:d8:b8:da? [17:40:43] <delewis> ! [17:40:46] <delewis> that's not DHCP [17:40:47] <astinus> PosixC: If you're installing Solaris onto a T2000, I can guarantee the tutorial works, because I used it last month :P [17:40:50] <PosixC> But On the server (which runs "snoop -v dhcp") there are no [17:40:50] <PosixC> traffic from this MAC addres. [17:40:58] <delewis> do I relaly want to ask what OBP command you're using off of the network? [17:41:01] <delewis> really, rather. [17:41:08] <PosixC> boot net - install [17:41:14] * delewis beats head against desk [17:41:20] <astinus> delewis: Yeah, I think we ended up TFTP booting our T2000 to install it, heh [17:41:29] <delewis> PosixC: so [17:41:31] <delewis> basically [17:41:35] <delewis> you've wasted the last hour or so [17:41:42] <delewis> trying to use rarp/bootp to boot it [17:41:47] <delewis> and configuring DHCP [17:41:58] <delewis> congratulations. [17:42:04] <PosixC> delewis, it is more than hour. [17:42:22] <PosixC> delewis, I don't understand: [17:42:25] *** estibi has quit IRC [17:42:26] <delewis> PosixC: and a simple "help boot" would've told you there's a DHCP option [17:42:33] <delewis> boot net:dhcp - install [17:42:48] <PosixC> ooooffffff [17:42:55] <PosixC> I saw it somewhere [17:42:59] * astinus hides from the wrath of delewis [17:43:32] <PosixC> delewis, but the strange thing here is that in the [17:43:41] <PosixC> Solaris 10 6/06 Installation Guide: Network-Based Installations [17:44:00] <PosixC> they talk about DHCP a lot and they don't mention it [17:44:11] <PosixC> they talk on boot net - install [17:44:29] <delewis> IIRC, there's a separate manual or chapter for DHCP-based installations [17:44:31] <PosixC> i am trying it now [17:44:48] <PosixC> yes , ip services (separat manaul) [17:44:49] <quasi> delewis: I think you're right [17:45:10] <PosixC> but they also talk a lot about DHCP in Network-Based Installations [17:45:37] <delewis> PosixC: well, regardless, you need to use boot net:dhcp [17:45:42] <delewis> if you want to use DHCP [17:45:49] <delewis> boot net does rarp/bootp [17:45:58] <delewis> which is the old way of doing things (hence, the extra dhcp paramter) [17:46:13] *** deather_ has joined #opensolaris [17:47:38] <astinus> Hm, what packages are required to operate a NFS server? [17:48:46] <delewis> SUNWnfsskr, SUNWnfssr, SUNWnfssu [17:49:15] <astinus> danke :) [17:51:19] *** boro has joined #opensolaris [17:53:01] *** deather has quit IRC [18:00:28] *** piwi has quit IRC [18:01:14] *** movement has joined #opensolaris [18:10:09] <PosixC> ooff ! [18:10:25] <PosixC> I had another DHCP server on the same LAN [18:10:32] <PosixC> I had to stop it [18:10:37] <PosixC> Now I am getting [18:11:02] <PosixC> Timed out waiting for BOOTP/DHCP reply [18:11:09] <PosixC> TFTP server not specified [18:11:18] <PosixC> this is on the T2000 seide [18:11:24] *** darkcmd has quit IRC [18:11:27] <PosixC> s/seide/side [18:11:35] <PosixC> BUT on the install server side [18:11:38] <PosixC> I do see : [18:11:55] <PosixC> DHCPDISCOVER [18:12:13] <PosixC> than: DHCPOFFER [18:12:18] <PosixC> than: [18:12:39] <PosixC> DHCP request : class identifeid : SUNW.Sun-Fire-T2000 [18:12:44] <PosixC> with the correct MAC address [18:13:11] <PosixC> what is this tftp server error mean? should a tftp server be running on the install server ? [18:14:21] *** estibi has joined #opensolaris [18:17:59] <lasseoe> yes [18:18:27] *** boro has quit IRC [18:19:34] *** rmorse has joined #opensolaris [18:22:09] <astinus> Anyone familiar with NFS? I've had it implemented on Linux for a while now, but its always a pain; you need the same UID/GID everywhere, or you run into permissions/authentication issues. It was my understanding there are "other options" for handling this - anyone got some links I could peruse for ideas? [18:22:26] <astinus> (I'm looking to get the fileserver (which is now Solaris) serving files over NFS) [18:23:32] <edp> i think what you're looking for is somethingg like NIS that will give you the same UID/GIDs across systems [18:23:38] * astinus nods [18:23:40] <Auralis> the keeping the UIDs in sync is a fundamental prinziples. you are accessing a other machine aftewr all [18:23:49] <astinus> edp: It was my understanding that NIS is EOL'd ? [18:23:52] *** DataStre1m has joined #opensolaris [18:24:04] <Auralis> however there are several methods how to make that as painfull as possible [18:24:10] <jamesd> hi DataStre1m [18:24:17] <Auralis> cpoy the authfiles, nis, nis+, ldap [18:24:27] * astinus nods [18:27:10] *** DataStream has quit IRC [18:35:15] *** LeftWing has quit IRC [18:35:18] *** LeftWing has joined #OpenSolaris [18:36:09] *** karrotx has quit IRC [18:59:15] <PosixC> Is there a way to remove a network from dhcpmgr GUI ? It does not have a delete button nor can I delete a single address from a range [19:08:55] *** PosixC has quit IRC [19:15:33] *** rmorse has left #opensolaris [19:21:35] *** kill9 has quit IRC [19:24:34] *** reflect has joined #opensolaris [19:24:57] *** McMurlock has quit IRC [19:25:46] <reflect> hi.. I'm writing a bit of a log checker, and I'm looking for "betatesters".. if you're interested, let me know and I'll give you access to it [19:27:17] <lasseoe> hey dude :) [19:27:57] <reflect> dude? [19:28:06] <reflect> hm.. [19:28:29] <reflect> where you in #solaris on #ircnet.. around.. well, 3-5 years ago? :) [19:28:38] <lasseoe> indeed [19:28:41] <reflect> haha [19:28:56] <lasseoe> 'tis a small world :) [19:29:01] <reflect> hm.. kirma.. [19:29:36] <reflect> that channel is still huge [19:29:40] <reflect> 50+ people there [19:30:08] <lasseoe> yes although quiet these days [19:34:36] <Somethingelse> naw, it gets loud in there [19:34:55] *** estibi has quit IRC [19:36:36] *** estibi has joined #opensolaris [19:40:48] *** ircminer03 has quit IRC [19:42:19] *** ircminer03 has joined #opensolaris [19:49:10] *** alanc_away has joined #opensolaris [19:49:10] *** alanc has joined #opensolaris [19:50:51] *** reflect has quit IRC [19:56:23] *** victori has joined #opensolaris [19:56:39] *** victori has left #opensolaris [20:00:22] <astinus> is there a kinda metapackage that covers all the graphical stuff? [20:00:30] <astinus> ie: Xorg, fonts, JDS [20:01:50] <Auralis> no [20:02:59] *** reflect has joined #opensolaris [20:05:29] <delewis> astinus: this is why you should do full installs. [20:05:49] <astinus> delewis: Actually I wanted to remove Xorg :p [20:06:03] <astinus> delewis: Its a headless box, it'll never have graphical users/clients [20:06:45] <delewis> you don't want to remove Xorg [20:06:51] <delewis> doing so would remove various X11 libraries [20:07:01] <delewis> which means you couldn't run any graphical application, whatsoever. [20:08:42] <astinus> delewis: Any why would I want to run graphical applications on something that is purely an NFS server? :/ [20:09:09] <delewis> you never know. You may find yourself forwarding an X11 session from it someday. [20:09:33] <astinus> fair enough :) [20:09:41] <delewis> the system being headless has little to do with it. [20:09:48] <delewis> you're thinking like a Windows luser :-) [20:11:23] *** stevel has joined #opensolaris [20:11:23] *** ChanServ sets mode: +o stevel [20:11:28] <Error_404> I have X turned off on my fileserver, but I never removed it [20:11:40] <reflect> I've written a little script that tells you about anomalies in logfiles.. [20:11:44] *** jacotton__ has joined #opensolaris [20:11:46] *** alanc_work has joined #opensolaris [20:11:49] *** dduvall_ has joined #opensolaris [20:11:56] *** alanc_ has joined #opensolaris [20:12:01] <delewis> if you should ever choose to install anything with a graphical installer, you'll need X11 libraries, if only to forward the X11 display [20:12:17] <reflect> I'm looking for testers.. people with visions, people who are used to breaking scripts.. people that aren't afraid of saying what they think.. [20:12:25] <astinus> reflect: I think you already asked once... [20:12:28] <delewis> like *cough* Oracle or DB2 *cough* [20:12:37] <astinus> delewis: Fair point, and yeah, I hadn't considered that :) [20:12:42] <reflect> astinus: perhaps I did.. however, my client died. [20:13:06] <astinus> delewis: done any centralized authentication using LDAP? [20:13:14] *** lin`` has joined #opensolaris [20:13:21] <astinus> delewis: if so, got any pointers for a Solaris-10 oriented setup/guide? [20:13:26] <reflect> astinus: yes, sure.. our entire infrastructure is based upon openldap [20:14:29] <delewis> reflect: I pity you :-) [20:14:49] <astinus> lol :x [20:15:05] <delewis> astinus: http://docs.sun.com/app/docs/doc/816-4556/6maort2rh#hic [20:15:16] <reflect> delewis: are you kidding me or something? [20:15:17] <astinus> delewis: thanks muchly :) [20:15:23] <delewis> it's really a straight-forward process, presume you're using a real LDAP server, like Sun ONE [20:15:35] <delewis> create a profile to automatically setup each client on the server [20:15:50] <delewis> use ldapaddent to add entries for various schemas, like passwd, group, etc. [20:15:59] <delewis> reflect: no [20:16:05] <delewis> I'm saying you're using a shitting LDAP server [20:16:09] <delewis> shitty, rather. [20:16:31] <delewis> OpenLDAP is a heaping pile of garbage in most cases. [20:17:44] <reflect> well, our 'shitty' ldap server was an option long before "sun one" ldap server was in existance. and I'm sorry, but please keep your opinions to yourself, especially if you express them without any facts to back them up, otherwise you're just inciting flamewars. [20:17:49] *** alanc has quit IRC [20:18:00] <delewis> reflect: you are wrong [20:18:09] <delewis> Sun ONE existed *long* before OpenLDAP did [20:18:12] <delewis> it originated from Netscape [20:18:24] <delewis> which has always been the pre-dominate LDAP server [20:18:29] <delewis> please, find a clue quickly. [20:18:31] <reflect> ok, perhaps I was [20:18:43] <reflect> now, please.. please, look at the second part of my message. [20:18:44] *** alanc_away has quit IRC [20:18:53] <reflect> I'm a big enough man to admit I was wrong, you see. [20:18:55] *** dduvall has quit IRC [20:19:22] <delewis> reflect: Sun ONE does replication, load balancing, thus it's meant to handle thousands and thousands of users. [20:19:44] <delewis> and Sun ONE LDAP (before Sun made the entire JES suite free) was free for implementations that had under 150,000 users. [20:19:46] *** jacotton_ has quit IRC [20:19:52] <reflect> delewis: wow.. you really need to check up on current events. [20:20:10] <delewis> now for Sun to make it free for under 150,000 users implies that good deal of customers using Sun ONE LDAP have installations that support over 150,000 clients. [20:20:19] *** Cybernd has joined #opensolaris [20:20:31] <delewis> and Sun ONE LDAP has excellent security and integration with a variety of operating systems, heterogeneous or not. [20:20:36] <reflect> now, I won't look down on you for picking .. well, whatever software you choose [20:20:44] <Error_404> ugh, it'd be nice if i could download stuff from dlc via command line [20:20:55] *** linma` has quit IRC [20:21:04] <reflect> now please, give me the courtesy of picking my own software, without you sitting there spitting at it. [20:26:56] *** stevel- has quit IRC [20:31:57] *** boro has joined #opensolaris [20:40:52] *** Burana has joined #opensolaris [20:41:40] *** deather__ has joined #opensolaris [20:42:44] *** deedaw has quit IRC [20:43:33] *** deather_ has quit IRC [20:43:56] *** srirama has joined #opensolaris [20:44:31] *** srirama has left #opensolaris [20:48:16] <Error_404> hmm... IBM really has it out for sun [20:48:27] <Burana> why do you think? [20:48:39] <Error_404> haven't a clue, tbh [20:49:02] *** sunnyDay has joined #opensolaris [20:49:22] <Burana> Isn't IBM the largest customer of sun? [20:49:22] <Error_404> not why they'd have more against sun than say, hp, anyways [20:49:47] <delewis> because HP isn't a big UNIX vendor, anymore. [20:49:47] <jbk> Burana: at one point global servers was a big customer i believe [20:49:55] <galt> Error_404, because Sun actually has useful Big Iron [20:50:18] <jbk> a friend who long ago worked for them joked about the fact they sold customers solaris solutions insead of aix [20:50:25] <Burana> I worked for IGS. In my location we had approx. 300 RS/600 Boxes and 200 Sun Servers. [20:50:55] <Burana> Customers always preferred Solaris :-) [20:52:30] *** vmhobbes has quit IRC [20:52:42] <Error_404> & soon, even IBM big iron'll be able to boot it [20:52:47] <Error_404> (referring to polaris) [20:53:02] <Burana> The problem for IBM in my location was, the HQ in Paris who decided the prizing, couldn't match the same discount on RS/6000, as Sun could give [20:53:20] <jbk> ibm has never been the low cost solution [20:53:49] <delewis> IBM is pushing AIX, nowadays. [20:54:00] <delewis> mostly because it's easy for vendors to push it [20:54:01] <Error_404> linux, you mean [20:54:05] <delewis> Error_404: no [20:54:12] <delewis> they push AIX to meaningful customers [20:54:15] <Burana> in one case I heard, Sun gave 51% discount instead of normal 35% 8-) [20:54:28] <delewis> anyway, it's easy to push AIX, because vendors don't have to be UNIX administrators to implement it. [20:54:38] <delewis> the quality of IBM consultants that I've met, is well, questionable. [20:54:52] <icon> afternoon all [20:54:54] <Burana> I was a great IBM consultant for Solaris ... [20:54:55] <jbk> well.. there is 'AIX = Ain't unIX' [20:55:01] <jbk> which i've sometimes found appropriate :) [20:55:17] <delewis> most IBM consultants like to bad-mouth Solaris, from my experience [20:55:22] <delewis> calling it a "dinosaur" [20:55:27] <icon> jbk: ive always been partual to Linux Is Not UniX ;) [20:55:30] <delewis> (when in reality they haven't touched Solaris since 2.4 on their SPARCstation" [20:55:33] <icon> s/partual/partial/ [20:55:38] <delewis> s/"/\)/ [20:56:04] <icon> meh, i work in an ibm shop here [20:56:16] <icon> ive been pushing pretty hard to get some sun boxen in here, looks like thats finally going to come true [20:56:19] <delewis> I'm sure that isn't the case everywhere, but the local Memphis IBM VAR is quite vocal [20:56:26] <delewis> on "AIX >>>>>> Solaris" [20:56:41] <Burana> My opinion is, that IBM has an advantage in selling the whole stack including services. otherwise no-one would by aix [20:56:45] <icon> delewis: youre in memphis? [20:56:52] <delewis> icon: about an hour away [20:56:58] <icon> delewis: interesting. st louis here [20:57:20] <delewis> Burana: not really. [20:57:21] <Triskelios> some kernel hackers from Sun are going to be recruiting at my uni [20:57:27] <delewis> AIX virtualization solutions are pretty easy to implement [20:57:30] <jbk> well sun's biggest problem (as observed from the customer's pov) is that their sales force sucks [20:57:32] <Triskelios> I'm going to make them fix my laptop wireless driver [20:57:33] <delewis> which is why vendors like it [20:57:40] <delewis> get yourself a few pSeries [20:57:43] <delewis> an HMC [20:57:47] <delewis> HACMP license/media [20:57:53] <delewis> an IBM storage solution [20:57:58] <delewis> and you've got yourself a full HA solution [20:58:15] <icon> well, in order to actually work with sun, you generally have to know what you are doing. ibm tends to sugar coat everything, which disturbs me hightly [20:58:18] <icon> highly [20:58:22] <delewis> and the local Memphis guys love to implement that sort of thing [20:58:23] <Burana> maybe the other advantage of aix is is the not too bad performance of the POWER architecture... but it seems they can't push new CPUs out as fast as they planed. [20:58:26] <delewis> because it's easy [20:58:37] <Error_404> btw, AIX requires actual hardware from IBM, yes? [20:58:42] <delewis> Error_404: of course.. [20:58:48] <Error_404> as in, i can't install it on a mac or a genesi machine [20:58:55] <delewis> you wouldn't want to run AIX on anything *but* IBM hardware [20:59:02] <quasi> and you can't [20:59:03] <delewis> the great thing about AIX is that the hardware integration is so tight. [20:59:13] <Error_404> well, I just meant to play w/ it [20:59:15] *** mercykiller has quit IRC [20:59:23] <delewis> there's a ridiculous amount of communication going on between AIX and the underlying hardware [20:59:25] <galt> which really sucks if you want to go IBM bastard shop and have full interoperability with the x86 desktops [20:59:25] <Error_404> I don't feel like dropping 5 grand on a pseries to play with [20:59:32] <delewis> Error_404: too bad [20:59:37] <Burana> delewis: the bad thing about AIX is that the hardware integration is so tight. namely it runs only on POWER :-) [20:59:44] <delewis> regardless, most of the "playing around with AIX" tends to be playing around with the underlying hardware. [20:59:55] <delewis> Burana: again, you wouldn't want to run it anywhere else [20:59:59] <quasi> Error_404: you can find a box to run aix on for a lot less [20:59:59] <icon> power5 isnt a terrible architecture, but yech... [21:00:04] <delewis> the hardware is great, period. [21:00:09] <Error_404> mostly because i don't have 5 grand, and if i did i have a list of bills that come first [21:00:12] <delewis> whatever you think of the operating system, AIX. [21:00:24] <icon> delewis: our data centers here are loaded with power5's and power5+'s [21:00:27] <Burana> delewis: no, I don't want it really near me :-) but we have some boxes :-( [21:00:29] <quasi> Error_404: last I looked at was a couple of hundred US [21:00:47] <Error_404> quasi: what was? [21:00:47] <galt> delewis, show me a POWER mangler's desktop though [21:00:56] <quasi> Error_404: a box to run aix on [21:01:01] <delewis> quasi: nonsense [21:01:08] <Burana> delewis: and they look ugly... I really like the new design of the recent Sun boxes [21:01:11] <delewis> you can buy a low-end pSeries (that's capable of doing LPARs) for $3k [21:01:24] <delewis> I purchased a second-hand p640 (7026-B80) for $1.1k [21:01:36] <quasi> delewis: you don't need pseries for aix5.e [21:01:38] <Error_404> Burana: when the ultra40 first came out, i saw it & thought "oh, a powermac g5" [21:01:39] <quasi> 5.3 [21:01:42] <icon> Burana: id have to agree, im a bit partial to them too [21:01:47] <delewis> quasi: what? [21:01:48] <icon> I cant wait to get my hands on a niagra [21:01:52] <delewis> of course, you do. [21:02:01] <quasi> delewis: nothing new [21:02:08] <delewis> no [21:02:14] <delewis> my p640 is circa 2000 [21:02:24] <delewis> it's from when IBM was converting the RS/6000 to pSeries [21:02:46] <galt> I want a niagra that gdamore's people have gotten their hands on ;) [21:02:47] <quasi> the newer RS/6000s runs 5.3 as well [21:03:04] <delewis> quasi: there are no newer RS/6000s [21:03:09] <delewis> RS/6000 died in 2000-2001 [21:03:20] <delewis> the p640 was one of the last systems to be part of both lines [21:03:22] <quasi> delewis: newest of the RS/6000 [21:03:31] <Burana> viagra 1 is really nice, but I can't wait to get my hands on viagra 2 and "the rock" [21:03:38] <boro> :)) [21:03:47] *** Cybernd has left #opensolaris [21:03:54] <quasi> still old junk, but runs 5.3 and practically being given away [21:04:08] <delewis> http://riemann.solnetworks.net/~dlewis/inventory/7026-B80.html [21:04:13] <Error_404> for $15000 ? [21:04:20] <delewis> Error_404: stop being silly [21:04:20] <Error_404> err, $1500 [21:04:21] <boro> http://www.googlefight.com/index.php?lang=en_GB&word1=niagara&word2=viagra [21:04:31] <delewis> you can get an RS/6000 capable of running 5.3 for $150-$500 [21:04:37] <delewis> 43p-150 or 44p-170 [21:04:48] <delewis> or 44p-270 or p640 if you want to splurge :-) [21:05:09] <Error_404> ahh, so you can [21:05:14] <quasi> 21:00 < quasi> Error_404: last I looked at was a couple of hundred US [21:05:29] <quasi> ;) [21:05:34] <Error_404> couple 43p's for ~$200 [21:05:43] <Error_404> slow, but only as slow as my U2 [21:05:58] <quasi> who needs fast with aix anyway? ;) [21:05:58] <delewis> AIX media kits run around $100, though IBM actually has set consistent pricing on 5.3 media kits lately [21:06:01] <delewis> $120 [21:06:04] *** karrotx has joined #opensolaris [21:06:11] <delewis> you'd be surprised by how fast the RS/6000s and pSeries are [21:06:27] <delewis> I've seen B80s do amazing things as far as IO is concerned (pushing 300-400MB/s) [21:07:07] <delewis> almost 24/7, that is :-) [21:07:30] <delewis> and stayed up since they were installed in 2000 [21:07:47] *** kimc has joined #opensolaris [21:07:52] <Error_404> next time i have a couple hundred bucks i'll pick one up I think [21:08:31] <delewis> I'm really a Sun person at heart, but I hate to see such injustices done to pSeries/AIX. [21:08:36] <delewis> it's really a fairly solid platform. [21:08:37] <quasi> Error_404: but be careful that you get one that can run 5.3 - getting one of the ones that is too old for that is no fun [21:08:52] <delewis> and I can say that I've seen AIX crash *less* than I have Solaris. [21:08:59] <Error_404> any way to tell (eg, black vs. beige) ? [21:09:03] <jbk> *shrug* [21:09:23] <Burana> What I dislike in AIX, is the lack of new features (useful ones like RBAC, Zones, ZFS) [21:09:26] <delewis> Error_404: generally, black systems run AIX [21:09:27] <jbk> we had one of the world's sp/2 clusters at one point [21:09:28] <quasi> delewis: beige is very likely to be too old [21:09:29] <delewis> er 5.3 [21:09:31] <jbk> and even with that [21:09:34] <jbk> we had a node die [21:09:37] <delewis> look for 64-bit systems [21:09:45] <delewis> particularly anything with a POWER3 in it [21:09:47] <jbk> and ibm spent *hours* pointing fingers between hw & software [21:09:54] <delewis> that would include the 43p-170 and 44p-270 [21:10:08] <jbk> it took us getting our VP involved before IBM shut up and actually tried to fix the problem [21:10:15] <delewis> Burana: you're forgetting that AIX was ahead of Solaris for a long time when it came to volume management [21:10:18] <reflect> was wondering if anyone had played around with the T1 cpu.. wether it be with solaris/opensolaris/linux.. ? [21:10:23] <delewis> even with VxVM [21:10:29] <jbk> and they eventually had to replace the entire node [21:10:48] <delewis> and AIX had a tracing implementation (very primitive, though) *way* before DTrace was around [21:10:48] <quasi> reflect: yes [21:10:53] <jbk> i've never had that sort of crap happen with sun service [21:11:08] <jbk> delewis: solaris had tracing prior to dtrace as well [21:11:13] <boro> http://cgi.ebay.com/IBM-RS6000-7044-170-400MHZ-512MB-18GB-AIX-RS-6000_W0QQitemZ190049688146QQihZ009QQcategoryZ11222QQrdZ1QQcmdZViewItem [21:11:19] <jbk> reflect: yes [21:11:26] <boro> this one looks useful, related toyour debate [21:11:28] <jbk> did some testing w/ ldap [21:11:34] <reflect> jbk/quasi: what can you say about it? [21:11:45] <jbk> for the right workloads, works great [21:11:48] <Burana> delewis: This is true... but realities are, that at the moment, there is not really much life in AIX [21:11:49] <jbk> in the ldap case [21:11:52] <reflect> what's the one thing that impresses you the most.. and, what's the one thing that puts you off the most [21:11:57] <jbk> we created a replica on it [21:12:09] <delewis> Burana: hardware virtualization is still superior in most cases [21:12:20] <lasseoe> reflect: biggest con is the lack of FPU [21:12:23] <jbk> then started pointing traffic to it [21:12:26] <delewis> DLPARs > LDOMs, that is until Sun starts doing something with the T1s hypervisor [21:12:29] <delewis> then that may change [21:12:46] <Burana> delewis: but can you remove cpu/memory boards online on an rs/6000? [21:12:48] <Auralis> delewis: sun is doing something with the t1 hypervisor [21:12:55] <reflect> lasseoe: and biggest pro? [21:12:57] <richlowe> ...LDOMs *are* using the T1s hypervisor [21:13:02] <jbk> utilization on the 280s dropped by 20%, while the t1 held steady by around 2% utilizaiton [21:13:03] <quasi> reflect: works fine for ww. but I had to drop it in one case because bea's jrockit doesn't run well on sparc [21:14:06] <reflect> we're dropping sparc.. we're dropping mips.. we're dropping everythng in favor of x86/linux [21:14:16] <delewis> DLPARs still can distribute CPU resources in very small quantums, and there's virtual SCSI, and virtual ethernet [21:14:23] <quasi> reflect: too bad for you [21:14:43] <quasi> ;) [21:14:47] <reflect> quasi: nah, generally it's quite good.. less to know, you know [21:15:02] <reflect> however, I really like the concept of the T1 cpu [21:15:07] <Burana> while technically DLPARs are nice, you still have n instances of the OS, I like Containers in Solaris much more... and for the bigger boxes Domain granularity should be enough [21:15:21] <quasi> reflect: but most of the sparcs absolutely blast away x86 on io [21:15:22] <reflect> if there was a blade with say, 2-4 T1 cpus on it.. we'd be all over it [21:15:37] <boro> i believe the operating life of most of x86 hardware lies somewhere up to 2 yesrs now [21:15:38] <lasseoe> reflect, that's coming :) [21:15:42] <Auralis> reflect: a T1 blade module is coming [21:16:03] <jbk> you also shouldn't run linux if you value your data :) [21:16:16] <reflect> quasi: perhaps. I highly doubt it, though, especially since we use opterons [21:16:44] <quasi> reflect: they have nothing on a box with usIV+ [21:16:47] <reflect> Auralis: would love it.. I fear it might be too late, though.. with the CELL [21:16:57] <quasi> reflect: (except floating point speed) [21:17:10] <lasseoe> clear [21:17:12] <lasseoe> woops [21:17:17] <delewis> boro: "up to" [21:17:26] <delewis> my Ultra 2 is still completely usable after 10 years, almost. [21:17:45] <reflect> quasi: let's see.. the sparc has never ever.. EVER had any advantage in speed.. over ANY other processor.. [21:17:53] <delewis> reflect: *wrong* [21:17:56] <quasi> reflect: IO [21:17:57] <Auralis> reflect: that is wrong [21:18:06] <delewis> would you please stop making further uninformed statements [21:18:09] * reflect smiles [21:18:10] <delewis> that display your lack of experience [21:18:12] <Auralis> the sparc2 was actually the fastest workstation of its days [21:18:13] <delewis> or knowledge of history [21:18:24] <boro> yes, nowadays we are experiencing maxtor disk badblock failures after some months of service [21:18:28] <astinus> reflect: I point you to earlier statements about who has the best LDAP server, and claim you are *wrong* :) [21:18:38] <jbk> and even *if* the sparc is the slower processor [21:18:40] <delewis> SPARC was almost always faster up until the PIII was released [21:18:45] <delewis> and it's still faster when doing IO [21:18:46] <jbk> there's still memory, io, etc. to consider [21:18:53] <Burana> astinus: who has the best ldap server? missed the answer [21:19:01] <galt> the SPARC architecture only really fell behind the speed curve in 2003ish, and a "slow" SPARC will still out IO a "blazingly fast" wintel any day [21:19:04] <reflect> the sparc has always been quite slow in.. well, most areas. it has always been very good in latency, especially network latency.. [21:19:11] <delewis> reflect: no [21:19:13] <delewis> you are clueless [21:19:17] <delewis> and hopeless [21:19:27] <delewis> you should look at benchmarks from pre-PIII days [21:19:27] <icon> ? [21:19:32] <astinus> Burana: delewis was pointing out why Sun ONE was/is superior to OpenLDAP [21:19:33] <reflect> it has always been crap in handling single tasks, but, weigh it down with thousands of jobs and it will trudge on [21:19:36] <icon> you know, i seem to recall the opposite [21:19:49] <Burana> astinus: is it? [21:19:53] <jbk> and the US-IV+'s are pretty speedy, but even then.. [21:20:06] <jbk> you need to ask yourself 'how much performance can i get for $$' [21:20:11] <reflect> most other processors, with the same amount of job, would do well in a single task environment, but would be CRAP when you start to weigh it down [21:20:20] <delewis> reflect: how could SPARC be slower then x86 had to have a math-co for how many years... [21:20:42] <delewis> and even then it still sucked [21:20:51] <jbk> even if it takex 2x the number of sparcs to get equivalent performance (not that it does), if it's half the cost of the equivalent solution from IBM, HP, Dell [21:20:56] <jbk> does it matter? [21:20:58] <galt> delewis, that was because intel can't get floating point down ;P [21:21:07] *** Yamazaki-kun has joined #opensolaris [21:21:37] <reflect> delewis: math coprocessor? what, did you just back up in time some 20 years? [21:21:52] <delewis> reflect: 486sx was not 20 years ago [21:22:00] <reflect> yeah.. just 15 [21:22:06] <boro> :) [21:22:08] <Burana> floating point performance is overratet. my banking software does only do about 4% float calcs [21:22:12] * quasi would also like to see the x86 box that could keep up with an e25k [21:22:15] <galt> reflect, uhm, the on-die math coprocesssor is still a separate unit to this day [21:22:32] <reflect> galt: yes, and? [21:22:37] <Burana> my core banking software would almost fly on the Niagara... [21:22:42] <reflect> quasi: that's not my argument, however. [21:22:52] <delewis> reflect: and the microSPARC was blowing even the 486dx or Pentium away in terms of performance [21:22:53] <galt> reflect, it's unique in chip design. most other processors don't need one [21:22:58] *** deather__ is now known as deather [21:23:19] <reflect> galt: "don't need one", or "doesn't have one, which is bad"? [21:23:22] <quasi> reflect: all the same, x86 is slow on io compared to usIV+ platforms [21:23:26] <delewis> UltraSPARC-I, -II, and -III kept up with raw integer and floating point performance until the PIII [21:23:26] <boro> btw...speedy in float was supposed to be new ibm cell [21:23:26] <Auralis> Burana: it should fly on the T2, which has 8 fpu units [21:23:27] *** ircminer03 has quit IRC [21:23:48] <galt> reflect, "don't need one, because they can do floating point math properly to begin with" [21:23:49] <reflect> quasi: ok, I'll give you that. I yield. [21:23:55] <delewis> which is when Intel went RISCish [21:24:04] <Burana> auralis: Yeah, I'm waiting for it. But the problem is, it needs RAS features.... [21:24:12] <reflect> when I was using sparc.. [21:24:18] *** ircminer03 has joined #opensolaris [21:24:31] <reflect> we're talking SPARC, ultraSPARC I and II mostly [21:24:42] <Error_404> jesus hell, i do a simple svn co of polaris about 20 hours ago, it's still running [21:24:50] <delewis> which should've been out-performing the equivalent x86 hardware [21:25:02] <delewis> unless you were comparing it to the latest and greatest (while the SPARC hardware was 2-3 years old) [21:25:04] <Yamazaki-kun> Anyone else having issues with the Sun download servers? 12kb/s if it works at all. :( [21:25:07] <reflect> they were cool, sure, but most other processors kicked their asses in most tasks. especially in crunching, where it counted [21:25:23] <Error_404> Yamazaki-kun, i'm pulling solaris10 down at about 400k/s [21:25:29] <galt> reflect, tehn compare apples to apples, math coprocessors were still in widespread use and intels capability for graphics processing was nascent at best [21:25:32] <Yamazaki-kun> hm. [21:25:47] <Yamazaki-kun> I suppose I can blame Verizon, then. [21:25:52] <galt> the US1 is circa 1995 [21:25:53] <reflect> galt: it's funny that everyone is assuming I'm touting intel here. [21:25:53] <Burana> the sun web sites suck in general. docs.sun.com, sunsolve.sun.com are f***ing slow... they should buy a couple of t2000 servers... [21:26:02] <delewis> and the USII is 1997 [21:26:05] <jbk> i would argue most places need i/o over number crunching [21:26:06] <delewis> late 1996, IIRC. [21:26:23] <reflect> the USII was used up until 2002 or some such [21:26:26] <icon> Burana: uh huh. and think about the massive amounts of traffic that sun pays for to serve free documentation to anyone who needs it... [21:26:27] <galt> reflect, okay, let's compare PPC. PPC what starting to fractionate between mac-PPC and RS/6000 PPC [21:26:28] <reflect> Sun had nothing better [21:26:34] <delewis> reflect: no [21:26:39] <delewis> the UltraSPARC-IIIs were out by then [21:26:45] <delewis> you were just too cheap or too uninformed to buy them [21:26:50] <delewis> you clueless dimwit. [21:26:51] <Burana> icon: i'm a paying customer... so give me premium access. [21:26:58] <icon> Burana: buy the books :) [21:27:01] <reflect> delewis: ok, you just stop this right now [21:27:15] <galt> reflect, want to compare MIPS? MIPS had even worse clock sppeds than US2 [21:27:20] <Burana> icon: outdated before they reach me [21:27:20] <delewis> reflect: the UltraSPARC-III had been out for 2 years by 2002. [21:27:24] <icon> whoa STOP [21:27:31] <icon> clock speed is NOT an indication of performance [21:27:37] <reflect> delewis: just cause I don't agree with you doesn't make me a clueless dimwit. [21:27:43] <galt> icon, according to reflect it is [21:27:50] <reflect> galt: bullshit [21:27:58] <reflect> galt: don't put words in my mouth, please. [21:28:15] <delewis> reflect: you don't agree with me about a date that can be easily determined? [21:28:16] <icon> the only indicator of performance is the clock on the wall... thats it [21:28:27] *** Fish- has joined #opensolaris [21:28:36] <reflect> delewis: don't make this about a date.. [21:28:37] <galt> 18MST] reflect quasi: let's see.. the sparc has never ever.. EVER had any advantage in speed.. over ANY other processor.. <- I don't need to, you do so yourself [21:28:38] *** sunnyDay has quit IRC [21:28:45] <jamesd> is reminded of the 80mhz chip that had greater performance than x86 chips that had 10 times the mhz. [21:28:45] <icon> MIPS (m. instructions per sec) isnt valid, clock speed isnt valid, flops, iops, none of it means a damned thing [21:28:55] <delewis> my SB1000 entered the market on 2000, and they *only* shipped with USIIIs. [21:29:11] <icon> think of the alphas... all they ever did was pump up clock speed when the arch couldnt sustain any more changes to enhance performance [21:29:18] <reflect> galt: yes, that was in reference to number crunching (that's what I mean by speed). [21:29:20] <boro> what do you say about this... [21:29:21] <boro> scientists at the University of Tokyo have built a 500MHz 512-core math co-processor chip that can perform up to 512 billion floating-point operations per second [21:29:32] <reflect> if you ever wanted a super computer, you rarely turned to sun [21:29:35] <galt> reflect, tehn it's your fault for not being specific [21:29:48] <galt> reflect, hmm, tell that to E10k users [21:29:51] <icon> think of the p4's once they hit the 4ghz mark they realized that netburst was a piece of shit and reworked the cores into something decent (core 2 duo) which is now clocked at around 2.3ghz [21:29:58] <reflect> galt: that may be. it could also be others fault for "assuming what I mean" [21:30:19] <icon> clockspeed usually indicates how much power you feel like throwing away to make users believe they are using something fast [21:30:22] <Error_404> the HPC virtual lab in ontario runs a cluster of a half-dozen e20k's [21:30:35] * delewis remembers when E4500s were marketed as HPC systems [21:30:36] <Error_404> so, *shrug* [21:30:41] <delewis> and those were UltraSPARC-II systems [21:30:42] <galt> reflect, well, if you're going to be vague, others have to assume because you never get specific [21:31:04] <icon> 'supercomputer' is incredibly vague [21:31:05] <reflect> galt: well, please just ask. I'd much rather a question than an insult. [21:31:14] <delewis> reflect: you don't question [21:31:17] <icon> there are 'supercomputer' made of of amd64 clusters now [21:31:18] <delewis> you make uninformed statements [21:31:24] <delewis> and a fairly large quantity of them [21:31:33] <delewis> "SPARC suzk0rz" is not a question. [21:31:38] <icon> the age of the supercomputer is dead. long live the network! [21:31:40] <galt> how about you don't make vague descriptions and claim that you never said that? is that so hard? [21:31:42] <reflect> I never said that [21:31:51] <delewis> reflect: I paraphrased, albeit accurately. [21:31:56] <Burana> sad is, that those supercomputers sold by Sun don't run on Solaris but crappy linux [21:31:59] *** sickness has quit IRC [21:32:05] <galt> reflect, again, you made a vague statement, what do you expect? [21:32:24] <reflect> I said that when it comes to number crunching, the sparc has never really been a contestant.. however, when it comes to network, the sparc has been incredibly good, especially when it comes to latency etc [21:32:24] <boro> what's crappy about linux.. ;) [21:32:38] <delewis> and you were incredibly wrong. [21:32:55] <delewis> Sun marketed their mid-range servers as HPC systems for years. [21:33:06] <delewis> and a lot of educational facitilities bought into it [21:33:16] <reflect> hm [21:33:25] <reflect> we seem to be talking about two different things [21:33:26] <Auralis> the sparc2 was actually the fastest workstation of its days [21:33:52] <delewis> http://www.ap.stmarys.ca/hpc/hardware.html [21:33:55] <delewis> need I say more? :-) [21:33:56] <galt> reflect, so you have numbers to back up your number-crunching theory? [21:34:15] <reflect> galt: yes, I'm a little bit drunk as it is now, which may explain my vagueness [21:34:15] <Error_404> i think my school picked stereotypes for their computer labs, tbh [21:34:17] <boro> so what's this little cpu fight all about ? [21:34:26] <boro> past is a past [21:34:27] <Error_404> sun workstations, ibm servers, and an SGI in the HPC lab [21:35:14] <delewis> http://www.ap.stmarys.ca/hpc/e4500blarge.jpg [21:35:19] <delewis> see the poster? :-) [21:35:19] *** LeftWing has quit IRC [21:35:20] <reflect> galt: you honestly believe that the sparc.. or ultrasparc.. was ever classed as "top of the line, this processor beats all other processors in numbercrunching"? [21:35:25] *** LeftWing has joined #OpenSolaris [21:35:32] <delewis> reflect: yes [21:35:38] <delewis> the larger systems scale amazingly well [21:35:39] <galt> reflect, no, but I don [21:35:44] <delewis> and if you have a properly written application [21:35:47] <delewis> that scales properly [21:35:50] <galt> 't believe YOUR statement without backing [21:35:57] <Auralis> reflect: since its a fact that it did, so yes [21:36:04] <reflect> delewis: no, see, that's where we differ [21:36:11] <reflect> I'm talking about the processor [21:36:11] *** mjf has joined #opensolaris [21:36:18] <reflect> not about an entire system [21:36:29] <delewis> why not? [21:36:37] <delewis> SPARC is the scalable processor architecture, btw. [21:36:38] <boro> btw did instruction sef between uultra sparc 2 and 4 change significantly [21:36:43] <galt> we aren't marketroids here. put up or shut up [21:36:46] <delewis> so it only seems natural to take into account all of its abilities. [21:36:53] <reflect> because specialized system to do a special task can always be built [21:37:14] <delewis> reflect: and SPARC was built to be scalable. [21:37:20] <reflect> indeed it was [21:37:23] *** Fish has quit IRC [21:37:28] <reflect> which was one of its very strenghts [21:37:48] <reflect> but let's break it down into a single processor [21:37:49] <boyd> Morning, all [21:37:51] *** sickness has joined #opensolaris [21:38:05] <delewis> so even if a single x86 processor does out-perform it (which wasn't the case in pre-2000), you can always scale higher than x86 to out-perform. [21:38:21] <reflect> yes, quite right [21:38:24] <Burana> delewis: nicely said [21:38:30] <reflect> very good [21:39:25] <reflect> except today it might be a bit different with clusters (if you have a job that is "embarrassingly parrallell") the cost/performance ratio sort of swings other ways [21:39:31] <galt> reflect, BTW yoiu were vague again. WHICH processor are you stating does number crunching better? [21:39:39] <boro> i guess anyway if sun will not shop their cpus in prices comparable to x86, it will disappear sooner and later .. [21:40:20] <delewis> reflect: the interconnects between the processors/memory on scalable, SPARC-based systems are amazingly fasted. [21:40:26] <delewis> fast* [21:40:29] <boyd> boro: Yes, that is why there are no products anywhere in any category that are more expensive than others [21:41:09] <delewis> the latest GigaPlanes do something like 20GB/s, IIRC. [21:41:24] <delewis> and even that figure might be way out-dated. [21:41:47] <reflect> galt: good question, and it depends on what the operation is, really. the sparc is good at some things, and not so good. it also depends on when in time we speak. if it's pre 90's, then I'm not so sure. the alpha was quite freaking fast in its day, and I'm pretty sure it beat the sparc on most accounts. [21:42:29] <Burana> The question is. How many Supercomputers vs. how many general purpose systems are needed? how much money can you make with supercomputers vs. general purpose systems? Sun's strategy includes the second kind of systems. [21:42:36] <reflect> in the later years, the power5 and opteron has been strong contestants to the top cruncher roll [21:42:51] <reflect> Burana: very good question [21:42:55] <reflect> Burana: it all comes down to need [21:43:06] <reflect> do you need a single image system? [21:43:15] <Auralis> Burana: given the state of Cray, not that many :) [21:43:23] <galt> reflect, until you're going to be specific about what you're talking about, don't tell others that their comparisons are irrelevant [21:43:24] <reflect> is your applications.. best suited, in a single image system? [21:44:00] <reflect> galt: floating point and integer calculations of the sparc, compared to other architectures. [21:44:37] <galt> reflect, you won't even be specific as to what time frame, much less what specific processor you're comparing [21:44:56] *** Yamazaki-kun has quit IRC [21:44:59] <reflect> 1995-2005, sparc against everyone else [21:45:20] <reflect> all tests uses top of the line processors from each respective company. [21:45:36] <delewis> the only two contenders would be Alpha and MIPS (and past 2003, Itanium, AMD64, or POWER) [21:45:59] <reflect> 1995, you'd have to include VAX aswell [21:46:52] <galt> okay, 1995, no AMDs at all, compare sun4u against R4400, first gen alphas, and intel p1/2 [21:47:10] <boro> http://www.top500.org/static/lists/2006/06/TOP500_200606.xls [21:47:13] <delewis> reflect: VAX was long dead in 1995. [21:47:22] <delewis> DEC had been doing Alpha stuff since 1992 (EV4) [21:47:41] <galt> delewis, I thought EV5 was 1996 [21:48:15] <reflect> Burana: for some applications, single image systems is "teh shit". there are no comparisons, no matter if a single processor beats a single processor of your system - however, your system has.. 10-1000 processors, so what do you care? :) [21:48:17] <delewis> 1995, according to wikipedia [21:49:04] <galt> okay, EV5, it still sucked at floating point. think multia [21:49:08] <reflect> considering we just retired our last VAX machines, I sort of figured they were still on sale around 1995-1999 [21:49:18] <delewis> reflect: nope [21:49:29] <delewis> that's the lifetime of a VAX :-) [21:49:34] <reflect> hehe :) [21:49:42] <galt> nothing sucks like a vax ;P [21:49:49] <Burana> R.I.P. VAX, now move on... :-) [21:49:50] <Error_404> delewis, which is, 30 years? [21:50:27] <delewis> the last VAXen were sold in the early, early 90s [21:50:28] <reflect> so, did I mention that syslog script I was writing? [21:50:29] <delewis> 90-91 [21:50:34] <delewis> and those were high-end systems [21:50:39] <jbk> what was the mantra of maintaing the vax -- always protect the fuse? something like that [21:50:57] <galt> wasn't PPC still a year or two off, so the US1 compared to POWER2 at taht point? [21:51:02] <reflect> I'm sort of looking for critiques [21:51:15] <delewis> galt: PowerPC made in-roads about 1992 [21:51:24] <delewis> and got picked up by Apple and friends ~ 1993-1994 [21:51:25] <Burana> reflect: what are you writing? [21:51:26] <galt> okay, so it was the 60X [21:51:35] <delewis> galt: definitely [21:51:38] <delewis> 601, 602, etc. [21:51:40] <reflect> beta-testers, enhancements.. and seeing since you guys really hate me now, I guess you'd be good betatesters [21:51:45] <delewis> 604 was 1995 [21:51:52] <delewis> and that was high-end for the time [21:52:03] <reflect> Burana: think syslog... 99% of the text is.. well, informational, not important [21:52:06] <galt> those really sucked at floating point. which is why they made good macs [21:52:10] <Burana> reflect: maybe we hate you even more then? :) [21:52:15] <delewis> I have a 7025-F40 (that I don't even want to know how much it costs) from 1995 that had 2x166MHz 604es and 640MB of memory [21:52:22] <Burana> reflect: have you tried Splunk? [21:52:37] <reflect> Burana: my script helps you weed out the "good to know" from the really, really important issues [21:52:45] <reflect> Burana: no [21:52:53] <Burana> www.splunk.com [21:53:06] <Burana> reflect: their slogan: be an IT super hero :-) [21:53:08] <galt> yo9u have to compare the PA-RISC, because in '95 it was a real contender in the cobra jrs [21:53:42] <reflect> life was more interesting 10 years ago when you had a plethora of varioous architectures to pick from [21:53:43] <delewis> heh, PA-RISC would've been hot in 1995. [21:53:48] <deather> hi, does anyone here already used a PC graphics card on a Sun workstation running Linux/Xorg? I'd like to use an old S3 ViRGE on my Ultra 30 [21:53:49] <delewis> first architecture to have an SIMD instruction set [21:54:14] <galt> deather, it won't boot with it, that's for sure [21:54:16] <Burana> reflect: splunk is a really great piece of software. I just ordered a license... benr did blog about it once [21:54:17] <delewis> (which Sun, HP, and Intel quickly picked up) [21:54:22] <deather> galt: sure, that's not de problem :) [21:54:23] <reflect> deather: I've read that's no problem, atleast with openbsd [21:54:28] <delewis> along with DEC [21:54:37] <deather> reflect: OpenBSD really? i'll install it right now, thanks! [21:54:59] <reflect> deather: woah, man.. first check out openbsd.org and their hw platforms [21:55:16] <deather> mhh, they doesn't run on sparc64 if I recall? [21:55:23] <delewis> deather: yes, it does [21:55:26] <boro> they do [21:55:30] <delewis> no SMP support, though [21:55:31] <reflect> deather: perhaps it works just as well with linux, it was a long, long time ago since I ran linux on sparc [21:55:39] <boro> us 1,2 and now 3 probably [21:55:40] <icon> delewis: is there even MP support? [21:55:44] <delewis> icon: nope [21:55:49] <icon> didnt think so [21:55:56] <reflect> that's the one drawback [21:55:59] <delewis> USIII is still slow as fuck, I think. [21:56:00] <deather> okay. My Ultra 30 runs with an UltraSPARC-II alone, so it's not a problem [21:56:09] <deather> delewis: quite, yes... [21:56:12] <reflect> delewis: wasn't that netbsd? [21:56:17] <Burana> reflect: linux on sparc? in some countries you get the death row for that! blasphemy! [21:56:20] <delewis> reflect: ? [21:56:26] <delewis> NetBSD has partial SMP support for SPARCv9 [21:56:29] <icon> reflect: netbsd and openbsd share damned near the same codebase on sparc/usparc [21:56:32] <delewis> but it is broken at the moment [21:56:36] <delewis> OpenBSD lacks any, period. [21:56:45] <reflect> delewis: netbsd which was slow.. or was it obsd which was slow on certain usiii cpu's.. I don't recall. [21:56:52] <galt> icon, and no SMP is relevant to a U30 how? [21:56:53] <icon> considering open was a fork from net (iirc) [21:56:55] <deather> Burana: i've got enough SPARC computers to run Linux on some of them ;-) [21:56:58] <delewis> reflect: both would've been slow as fuck on the USIII [21:57:03] <delewis> because of the MMU [21:57:07] <icon> galt: its not, i was just speaking to delewis :P [21:57:16] <delewis> it was far from optimal to say the least :-) [21:57:27] <galt> okay, delewis no SMP is relevant to a U30 how? ;) [21:57:37] <reflect> still? the usiii isn't exactly "new in the hood" [21:57:40] <Burana> deather: as a bad example? [21:57:51] <delewis> reflect: Sun never released the MMU specification [21:57:52] <deather> Burana: just to give it a try :) [21:58:07] <delewis> the only other operating system that has a semi-optimal memory allocator for it is Linux [21:58:12] <delewis> and they signed an NDA with Sun to get the specifications [21:58:24] <reflect> delewis: not even if you signed an NDA? I remember someone in the opensource area signed NDA's.. perhaps it was the linux guys.. [21:58:29] <galt> now if I utilized the power of this not-0quite fully armed and operational U60, then SMP might be an issue [21:58:35] <icon> delewis: heavy on the semi [21:58:38] <delewis> reflect: the OpenBSD project refused to sign an NDA [21:58:51] <icon> linux on sparc sucks. period. [21:59:06] <reflect> icon: on the T1 too? [21:59:14] <delewis> reflect: yes [21:59:16] <delewis> on any SPARC [21:59:16] <Error_404> icon, it really does... [21:59:21] <icon> linux has shit for a vm [21:59:24] <delewis> performance is horrific in most cases. [21:59:28] <icon> not to mention the kernel is filled with garbage [21:59:29] <boro> the reason would be to use linux because of great networking [21:59:38] <delewis> anything but Solaris is compiled with a compiler that generates shitty code for SPARC [21:59:41] <icon> netbsd was the only serious contender back in the day with solaris [21:59:42] <Burana> can linux make use of the cryptoaccelerator on niagara? [21:59:49] <Error_404> heh, 5 months ago when i was looking, a cpu module for my U2 would've cost me shy of $100 on ebay [21:59:55] <icon> netbsd is so far behind now, its not even funny, but it *still* kicks linux's ass [22:00:04] <Error_404> now i can pick up a lot of 4 for 15$ [22:00:33] <reflect> btw, anyone here living close to scandinavia that might be interested in a 6-cpu Enterprise 3000? [22:00:44] <lasseoe> heh [22:00:53] <reflect> sorta heavy, I don't want to ship it ;) [22:00:59] <Burana> where in scandinavia? [22:00:59] <boro> what does netbsd kick ass to linux in ? [22:01:01] * delewis awaits his 8-way E4500 [22:01:04] <reflect> Gothenburg [22:01:05] <delewis> should be here Monday.. [22:01:05] <delewis> :-( [22:01:06] <icon> Burana: BFE :D [22:01:07] <delewis> too far away [22:01:17] <deather> (what a troll I started...!) [22:01:21] <Burana> for me too... [22:01:37] <Burana> icon: BFE? [22:01:48] <icon> Burana: Bum F*ck Egypt :D [22:01:49] <reflect> Burana: towards christmas, I'll be going south [22:01:55] <lasseoe> hm E3000, worth what.. a case of beer? :) [22:01:56] <reflect> Burana: would that be interesting? [22:02:03] <reflect> lasseoe: *slap* [22:02:32] <Burana> I know somebody in finland who would like one.... and no, its not Linus Thorvalds. [22:02:34] <lasseoe> I just need another US-III 900MHz [22:02:35] <jamesd> lasseoe, its worth a couple hundred, but the power will cost $100 a month... is it worth running. [22:02:45] <icon> Burana: torvalds lives in the US now [22:02:48] <icon> iirc [22:02:54] <lasseoe> jamesd, it's the power consumption that's the issue. [22:02:56] <reflect> dude, I've almost killed myself with that machine.. twice.. 4 stories, no elevator when I moved in/out [22:03:04] <lasseoe> and the fact that the e3k doesn't support DR, afair [22:03:12] *** incogito has joined #opensolaris [22:03:21] <jamesd> lasseoe, it does support DR. [22:03:39] <lasseoe> jamesd, thought that only started with the 3500 [22:03:40] <reflect> jamesd: you sure? [22:03:47] <jamesd> yes i'm sure. [22:03:51] <reflect> I think the 3500 was the first ones [22:04:04] <Auralis> it does with the right i/o boards [22:04:04] *** Kernel86_ has joined #OpenSolaris [22:04:17] <Auralis> soc i/o boards do not support dr [22:04:18] <boro> what the hell is DR ? ;) [22:04:18] <reflect> right [22:04:21] <jamesd> and new io boards are only $15 each.. [22:04:21] <reflect> it DIDN'T [22:04:28] <reflect> boro: dynamic reconfiguration [22:04:32] <boro> thx [22:04:46] <lasseoe> ok well in that case.. two crates of beer? :) [22:04:47] <reflect> boro: ie, rip a cpu board while it's running and it'll still work [22:04:57] <reflect> lasseoe: 12! [22:05:00] <Auralis> on my e4000 i can pull out and insert a cpu board just fine [22:05:03] <boro> aha...i know, i do it often with x86 ;) [22:05:09] <lasseoe> reflect: 12-bottle case? :) [22:05:24] <reflect> lasseoe: you gotta have specific numbers of your cpu/io/boards [22:05:52] <reflect> Auralis: yes, but that's a bit "newer" compare to E3000 :) [22:06:01] <reflect> iirc [22:06:01] <lasseoe> I've no Exx00 hardware, except for about 2g of memory [22:06:07] <Auralis> no, same age [22:06:27] <boro> what's the consumption of e4000 or e4500 fully loaded by cpus' ? [22:06:38] <Auralis> boro: to much [22:06:41] <icon> boro: scary [22:06:42] <reflect> Auralis: well, for the E3000, you have to have certain part numbers on your boards to be able to swap out things on the fly [22:06:43] <boro> i've measured e450 with 4x300 and it was near 1kw [22:06:51] <jamesd> boro, about 1400 watts. [22:07:04] <Auralis> reflect: umm, thats what i said [22:07:05] <Burana> in winter, you can shut down heating. [22:07:06] <reflect> you alsay have to have certain part numbers to go above a certain amount of ram, or a certain amount of mhz on your cpu's [22:07:14] <icon> meh ive got my main opteron box i use at home just under 400W now (finally) [22:07:17] *** Kernel86|Laptop has quit IRC [22:07:18] <delewis> heh [22:07:22] <icon> still too much juice with the rest of the rack running [22:07:27] <delewis> I can't wait to see the power consumption of my 8-way E4500 [22:07:28] <delewis> :-D [22:07:33] <boro> :) [22:07:34] <jbk> heh [22:07:37] <icon> delewis: disturbing ;) [22:07:42] <jbk> ever seen the power specs for a 25k? :) [22:07:43] <boro> why 400W, does it have multi cpus ? [22:07:51] <boro> 1MW ? [22:08:02] <icon> boro: disk array, dual core proc (just a single socket) [22:08:06] <jamesd> the e4500 is a win if you replace 4-5 x86 boxes with it. [22:08:10] <delewis> E10K alone puts about 50,000BTUs/hr [22:08:17] <Auralis> boro: multiple cpus? a e4000 can have 14 cpus [22:08:18] <Burana> the POWER CPU's are also using a lot of .... ahhh power [22:08:19] <boro> i guess it's quite a lot, those 400W anyway [22:08:24] <delewis> and about 10KW of power consumption [22:08:54] *** LeftWing has quit IRC [22:08:59] <icon> when i first built a box, i threw a 350 in there with the array barely used, and it was doing double time as it was [22:09:24] <reflect> when I left the scene, an E4000 could have a top speed of a single cpu of 480MHz (iirc).. did that ever change? [22:09:27] <icon> the 450 handles much better. im probably goign to throw in a new server board to cut down on the extra power being pulled on the pci-e bus for video [22:09:31] <icon> (it runs headless anyway) [22:09:32] <delewis> reflect: no [22:09:45] <delewis> 480MHz procs are extremely rare, too [22:09:48] <lasseoe> the latest E25k PSU's are rated at 4.7KW, and you need at least 6 :) [22:09:51] <delewis> almost as rare as the 2GB DIMMs :-) [22:09:57] <delewis> er 1GB DIMMs [22:09:58] <boro> 480 were fastest US2 i guess [22:10:28] *** LeftWing has joined #OpenSolaris [22:10:29] <reflect> .. I remember we sold quite alot of them.. but in.. ultra10/5 [22:10:42] <delewis> reflect: E-class systems use different ones [22:10:43] <reflect> but those were.. US2i or US2e [22:10:45] <delewis> they're socket-based [22:10:50] <delewis> and 8MB of cache/processor [22:10:52] <reflect> delewis: yeah, I know :) [22:11:08] <delewis> U5 and U10 used 2MB cache variants, IIRC. [22:11:12] <astinus> delewis: So what's a E4500 with 8 * 400MHz US2s, 8 * 1GB memory, 2 I/O boards, 3 PSUs worth? [22:11:23] <boro> there were really much us2 cpu models, surprisingly [22:11:26] <delewis> I paid $550 for mine, astinus [22:11:31] <delewis> and it's a similar configuration [22:11:36] <reflect> astinus: 2 stones, a couple of beers and a sandwich. [22:11:36] <astinus> hm, cool [22:11:39] <jbk> just watch out for alpha particles :) [22:11:50] *** bengtf has joined #opensolaris [22:11:57] <reflect> astinus: seriously, it depends quite a bit on how much it is worth to you :) [22:12:05] <boro> and how about the weight [22:12:06] <Burana> astiuns: the same E4500 is now called T2000 [22:12:29] <Burana> astiuns: and comes in a small form factor :-) [22:12:30] <boro> when i ordered 450, it layed down several days until i could stand it up on wheels with friend's help [22:12:40] <delewis> it's amazing how a system depreciates from $250k to $550 [22:12:45] <delewis> in the period of 6 years [22:13:06] *** alanc_ is now known as alanc_away [22:13:08] <boro> and what in next 6 years [22:13:09] <reflect> most other computers wouldn't even be worth that, in 6 years. [22:13:21] <boro> 1 server for 1 liter of crude oil ;) [22:13:48] <incogito> speaking of old systems, can anyone help me with a question about solaris/sparc binary compatibility? [22:13:50] <delewis> reflect: it's the rate of deprecation that's scary with the E-class systems [22:13:57] <delewis> a peecee going from $3k to $100 isn't that bad in the period of 6 years [22:14:12] <delewis> $250k to $550 is a *huge* rate, though. [22:14:40] * reflect shudders [22:14:43] <boro> why are they sold, anyway, they can serve well for many purposes so far [22:15:00] <delewis> boro: shops being pushed by vendors that the systems have been EOL'd [22:15:01] <boro> people don't play opengl games on them... [22:15:03] <delewis> I've seen it [22:15:13] <reflect> I just remembered that.. 4 years ago, I used to know the part numbers of all the common parts on sale about then [22:15:24] *** polpol has quit IRC [22:15:36] <reflect> and the sun.. hm.. X7002A <-- RAM [22:15:41] <jbk> incogito: ask your question, and perhaps someone will know [22:15:42] <delewis> before I left my old job we were being pressured the IBM VAR to get rid of two RS/6000s that were doing a perfectly good job of what they were doing for two new ones that cost around $120k/each [22:15:46] <reflect> X7004A.. also ram [22:15:48] <Burana> boro: the power bill is too high. [22:15:57] <delewis> hey, power is cheap here. [22:15:58] <delewis> :-) [22:16:01] <boro> power power [22:16:02] <delewis> $0.06/kwhr [22:16:13] <Error_404> heh, unbelievable.... I'm getting faster transfer from sun download center than I am from another machine speaking to the fileserver w/ 80211-g [22:16:34] <boro> i suppose in a short time guys are going to have 1kw nvidia machines at their homes, anyway [22:16:52] <reflect> delewis, galt, whoever : I'd like to apologize for my abrupt behaviour earlier [22:17:26] <delewis> no problem [22:17:34] <reflect> that's really not how I am.. and I really do like the sparc, and know quite a bit about it [22:17:43] <icon> its irc ;) [22:17:49] <boro> hehe [22:17:51] <icon> efnet syndrome [22:17:54] <reflect> icon: that's not an excuse, really [22:18:51] <incogito> jbk: i have an ultra 10 workstation running sunos 5.5.1. what i'm wondering is if it's possible to determine whether an app currently running on the system will probably run on a current sparc workstation running solaris 10 without actually trying it out (e.g., by looking at libraries used by the app, or something) [22:19:13] <jbk> what app is it? [22:19:29] <reflect> incogito: run 'ldd <your app>' and it might tell you a little [22:19:29] <jbk> chances are it will [22:19:30] <incogito> it's a medical imaging program [22:19:54] <reflect> incogito: that command will tell you which libraries it tries to open [22:19:55] <delewis> incogito: binary compatibility is maintained if you're using stable interfaces [22:20:03] <reflect> well, which libs it is linked against [22:20:13] <incogito> reflect: i know, i'm wondering if i can look at those libs and know whether it'll be compatible [22:20:30] <reflect> didn't sun have a suite for just this kind of thing? [22:20:36] <jbk> there's two programs [22:20:38] <jbk> appcert [22:20:44] <reflect> there you go [22:20:46] <jbk> and something else whose name escapes me [22:20:56] <reflect> jbk: but appcert is new.. [22:21:05] <jbk> unfortunately, i think you have to stick the binary + libs on the newer version [22:21:08] <reflect> you move it to the new one and run it.. ? [22:21:14] <incogito> d'oh [22:21:23] <jbk> and i've never seen appcert give *any* program 100% ok :) [22:21:42] <jbk> like anything that links against libnsl will always fail [22:21:48] <incogito> i think it does [22:21:58] <icon> well [22:21:59] <incogito> oh, you mean it fails appcert [22:22:06] <jbk> but if you look at what it flags, you can usually go 'oh, that's fine' [22:22:07] <icon> there is always the option of statically linking [22:22:14] <jbk> not on 10 [22:22:19] <incogito> it's a proprietary program and i don't have the source code [22:22:31] <icon> jbk: ? [22:22:45] <Auralis> no static libc on 10 [22:22:48] <delewis> icon: static linking with libc is no longer supported in 10 or greater [22:22:51] <Error_404> incogito: i don't suppose you have the option to just try it? [22:22:53] <delewis> the unified threads model broke that [22:23:00] <icon> delewis: thats using sun studio though iirc [22:23:03] <icon> what about gcc? [22:23:04] <boro> aren't there programs like qemu / wmvare sparc emulators to emulate particular platform ? [22:23:07] <reflect> so, would anyone like to try out my syslog filter script.. ? [22:23:07] <delewis> icon: no, it's using *anything* [22:23:15] <icon> ahh interesting. didnt know that [22:23:20] <delewis> the libc static archives are not provided anymore [22:23:27] <delewis> so whatever compiler you use is irrelevant [22:23:28] <Error_404> boro: iirc, opensparc.net's got an emulator of sorts [22:23:34] <icon> delewis: thats a good change [22:23:36] <incogito> Error_404: the software requires a hardware key, which i'd probably have to remove from the current hardware to do a test [22:23:38] <delewis> icon: yes [22:23:45] <Error_404> ahh, so no then [22:23:53] <delewis> the unified threads model would've made static linking *very*, *very* ugly. [22:24:21] <jbk> incogito: if you can just copy everything over to a solaris 10 box, you don't have to actually run the program, but you can run appcert, which will analyize it and spit out a bunch of info [22:25:44] <incogito> jbk: guess i'll try that then, thanks [22:26:09] <icon> incogito: have you checked ldd output yet? [22:26:23] <icon> or file? [22:27:20] <incogito> icon: i saved the output of both of those commands, but it's on another machine [22:27:58] <icon> it should tell you immediately whether or not there is a chance you could run it [22:28:32] <icon> ldd will help with what libs it links against [22:28:37] <icon> so you can take a look at any version changes [22:28:53] <incogito> unfortunately i'm not very familiar with sunos/solaris [22:29:04] <icon> well its not a sun/solaris thing really [22:29:22] <icon> ldd will tell you what a bin is linked against dynamically [22:29:25] <jbk> i'm trying to remember if there's a way to get the symbol versions of anything it's referecing in any shared libs.. [22:29:26] <incogito> i realize that [22:29:28] <icon> file will give you information on the executable [22:29:39] <jbk> well to easily get it in a nice easily viewable format [22:29:51] <icon> jbk: hrmm [22:29:57] <jbk> as i suspect probably just want to make sure it's not touching anything marked SUNW_private [22:30:34] <icon> i didnt know that sol10 did away with static libc - that could be a killer for him right there [22:30:58] <incogito> is there a pastebot for this channel? [22:31:05] <icon> rafb.net/paste [22:31:06] <delewis> lots of platforms no longer support static libc [22:31:19] <delewis> AIX dropped support static libc long before Solaris 10 [22:31:30] <icon> delewis: i want to say bsd still supports it [22:31:31] <delewis> (for the same reason Solaris dropped it) [22:31:45] <icon> thats where i used to spend 90% of my time [22:32:40] <icon> yup [22:32:45] <dlg> icon: what do you mean by support? [22:32:48] <icon> freebsd 6 still has libc [22:32:53] <icon> libc.a rather [22:33:04] <icon> dlg: statically linking to libc [22:33:15] <dlg> freebsd dont statically link /sbin and /bin anymore [22:33:20] <icon> no they dont [22:33:25] <dlg> but they have a standalone env that does [22:33:26] <icon> but you can still static link to libc [22:33:29] <dlg> yes [22:33:36] <icon> sol10 completely did away with statically linking to libc [22:34:08] <delewis> which isn't really of consequence, if you're using stable interfaces to start with (as you should be, anyway) [22:34:08] <dlg> i guess you lose a lot in solaris if you cant dlopen [22:34:19] <dlg> thanks to things like pam and nss im guessing? [22:34:26] *** dclarke has joined #opensolaris [22:34:37] <dclarke> g'day [22:34:40] <icon> heya den [22:34:47] <delewis> and as I said, the unified threads model would've made static linking very ugly [22:34:59] <delewis> think re-entrant support [22:35:04] <dclarke> anyone have a clue how to login via SSH to a V40z Service Processor ? [22:35:19] <dclarke> I tried username sun and password sun [22:35:20] <incogito> file/ldd output: http://rafb.net/paste/results/ZYB3Gk91.html [22:35:29] <icon> delewis: not to mention the sheer waste of space and memory consumption [22:35:51] <icon> PIC is a good thing [22:36:27] <icon> wow [22:36:31] <icon> havent seen some of those in a while [22:36:39] <icon> what was that... solaris 5? [22:36:44] <incogito> sunos5.5.1 [22:36:48] <delewis> Solaris 2.5 [22:36:56] <Auralis> 2.5.1 [22:36:57] <icon> delewis: spoilsport :D [22:36:59] <delewis> 2.5.1, rather [22:37:47] <delewis> incogito: I don't see anything in there that looks suspect [22:37:54] <icon> me either [22:38:19] <jbk> what's the SUNWits package? [22:38:20] <incogito> thanks for looking [22:38:29] *** fik has joined #opensolaris [22:39:21] <delewis> jbk: I have /opt/SUNWits present on my Solaris Express install [22:39:52] <boyd> jbk: pkginfo SUNWits [22:39:54] <delewis> it's apparenly installed by SUNWxil* [22:40:01] <delewis> boyd: that package does not exist [22:40:13] <delewis> SUNWxil* just creates symlinks to /opt/SUNWits [22:40:18] <delewis> (for compatility, of course) [22:40:23] <boyd> Ah, in that case pkgchk -l -P SUNWits [22:40:50] <jbk> ahh SUNWxilrl [22:40:54] <delewis> $ pkginfo | grep SUNWxil [22:40:54] <delewis> application SUNWxildh XIL Loadable Pipeline Libraries [22:40:57] <delewis> application SUNWxilh XIL API Header Files [22:41:00] <delewis> application SUNWxilow XIL Deskset Loadable Pipeline Libraries [22:41:03] <delewis> application SUNWxilrl XIL Runtime Environment [22:41:05] <delewis> application SUNWxilvl VIS/XIL Support [22:42:19] <boyd> Frankly I'm none the wiser [22:53:13] <Burana> anybody running solaris on a recent HP blade server? [22:53:48] *** ircminer03 has quit IRC [22:53:53] <Auralis> boyd: XIL is hardware accel X stuff, its EOLed [22:54:16] <boyd> m;kay [22:54:33] <Auralis> all the newer cards don't support it anymore [22:54:48] <jbk> incogito: what sort of box is this running on? [22:54:54] <boro> what's instead, opengl ? [22:55:01] <Auralis> boro: nothing [22:58:06] <incogito> jbk: current workstation is an ultra 10 (sun4u sparc SUNW,Ultra-5_10) [22:58:06] *** goedel has joined #opensolaris [22:58:33] *** pikapika has joined #opensolaris [23:00:10] *** Somethingelse has quit IRC [23:00:28] <pikapika> hello [23:00:36] <icon> evening [23:02:32] <Error_404> dclarke: you broke your v20z already? [23:02:44] *** ircminer03 has joined #opensolaris [23:05:01] *** joed has joined #opensolaris [23:05:32] *** Tpenta has joined #opensolaris [23:05:52] *** ChanServ sets mode: +o Tpenta [23:06:01] <boyd> Morning, Tpenta [23:06:06] <Tpenta> morning boyd [23:08:12] *** dclarke has quit IRC [23:08:31] <Error_404> i suppose that's a yes then [23:12:40] <Error_404> hmm... I gotta get another drive or two... 20M/s sucks [23:12:45] <boro> good night [23:12:53] *** boro has left #opensolaris [23:13:15] <icon> take it easy guys [23:13:16] * icon & [23:16:28] *** DataStream has joined #opensolaris [23:17:08] *** _Hunger- has joined #opensolaris [23:19:20] *** smoco has joined #opensolaris [23:23:07] *** Burana has quit IRC [23:24:53] *** tigerssb has joined #opensolaris [23:25:23] *** vmhobbes has joined #OpenSolaris [23:26:46] <tigerssb> can anyone tell me whats the system requirements 4 instaling opensolaris? [23:27:53] <smoco> 256 Mb ram is minimum x86 or 64_86 or sparc architecture [23:28:07] <tigerssb> ty [23:28:16] <Error_404> I've had it running on some pretty crap machines [23:28:19] <hali> a pc with a supported ide controller.. 256megs ram .. probably want 512 [23:28:24] *** Hunger- has quit IRC [23:28:28] <hali> ram is key [23:28:33] <Error_404> like, 400 megs of ram and a single USPARCII 300mhz processor [23:28:44] <hali> it runs great on a p3-600 with 1gb [23:29:26] <tigerssb> installed solaris 7 on a packard bell - shitty system - but worked - the first cd np - but the hd 2 small 4 the docs and man pages [23:29:47] <Error_404> please don't use numbers in place of words [23:29:50] *** DataStre1m has quit IRC [23:31:43] * joed wants to ask a really perhaps dumb question, (not trolling) but - why should I look at OpenSolaris as comapred to Fbsd / Linux having used Fbsd since 2.2.6 and Slack since about the initiation, besides ZFS? (Which I thought was really cool) [23:32:19] <astinus> joed: dtrace, zfs, self healing [23:32:27] <hali> joed: dtrace is one good thing.. and performance... and possibly ease of use if you want to run enterprise applications like oracle, websphere etc [23:32:43] <Error_404> a VM that doesn't crap it's pants trying to copy a file via NFS [23:32:44] <jamesd> zones [23:32:49] <hali> joed: but i'd say you should use what gets the job done and you feel comfortable using [23:32:50] <Saltsa> astinus: how does self healing work in practise? [23:32:54] <joed> astinus, - truss?, Oracle I run on Slack, Had code commits from Oracle on that. [23:33:11] <joed> Error_404, - that I would buy :) [23:33:15] <hali> Oracle on slackware? sounds like you are asking for trouble [23:33:30] <astinus> Saltsa: http://www.sun.com/software/solaris/availability.jsp [23:33:34] <hali> and thye have the privilege to refuse you support :) [23:33:35] <joed> hali, - Since about 2000 [23:33:47] <joed> hali, they actually support us. [23:34:16] <joed> All the Oracle Linux is really is a set of stub libraries anyways on the linux side. [23:34:57] <hali> oracle really enjoys to bitch about libc versions and stuff like that though.. like the classic libcwait error [23:35:18] <joed> Yeah? and it's really easy to get around. :) [23:35:24] <goedel> does opensolaris contain /usr/bin/xpg4/sh? [23:36:20] *** bougie has quit IRC [23:40:14] <goedel> does opensolaris contain a SUSv3 compliant shell? [23:41:40] <Error_404> solaris express does in either case [23:41:58] <goedel> thanks a lot Error_404 [23:49:06] *** goedel has quit IRC [23:50:13] <boyd> We shoudl charge for homework questions :) [23:50:58] <tigerssb> can I install opebsd on a slice while running windows, or is it better to run a dedicated hd? [23:51:25] * astinus wonders when #opensolaris became #openbsd [23:51:53] <joed> tigerssb, Yes. [23:52:13] <tigerssb> yeah - solaris is stuck in cde - I think cde sux as a desktop [23:52:34] <delewis> tigerssb: you are clueless [23:52:36] <delewis> go elsewhere [23:52:55] <tigerssb> I just think cde is just outdated [23:53:06] <delewis> you also said "Solaris is stuck in CDE" [23:53:09] <delewis> which is far from the truth [23:53:12] <boyd> tigerssb: You are right about that. You are not right about solaris being stuck in it [23:53:15] <delewis> it will definitely be dropped at some point. [23:53:25] <delewis> in the very near future [23:53:33] <tigerssb> ty 4 the clarification [23:54:00] <boyd> tigerssb: Solaris ships with the Java Desktop System, which is basically re-branded gnome [23:54:08] <boyd> (despite the name) [23:54:13] <astinus> delewis: think Solaris 11 (when it finally ships) will have dropped CDE? [23:54:27] <Error_404> I doubt it [23:54:30] <delewis> astinus: doubtful. [23:54:31] <tigerssb> seen the jdk - nice stuff I admit [23:54:34] <Error_404> there are still a few people who rely on it [23:54:35] <delewis> more likely the release after that [23:54:50] <delewis> Error_404: the libs will still be present [23:54:53] <Error_404> & i think it would break SUS not to have it [23:55:00] <boyd> tigerssb: jdk is a different thing. You mean JDS [23:55:03] <delewis> if you rely on your desktop environment you should re-evalulate your head. [23:55:15] <delewis> Error_404: no [23:55:20] <boyd> Error_404: I think only the libs, not the desktop [23:55:21] <tigerssb> boyd - I got mixed up [23:55:21] <delewis> SUS does not specify a desktop environment [23:55:29] <Error_404> okie, my bad [23:55:31] <delewis> only UNIX98 does (and just the separate, workstation portion) [23:55:48] <delewis> and Sun hasn't bothered re-qualifying Solaris 10 for UNIX98 anyway [23:55:48] * boyd defers to delewis [23:55:51] <delewis> only UNIX2003 [23:56:05] <delewis> it's likely they want re-qualify it for future releases [23:56:09] <delewis> s/want/won't/ [23:56:20] <boyd> Who would bother toady [23:56:23] <boyd> err today [23:56:26] <boyd> :) [23:56:46] <tigerssb> well, I all but gave up on windows - seems microsift takes it up the ass just to shovel out a crappy os [23:57:15] <boyd> That's not quit how I would have put it, but... [23:57:20] <boyd> argh quite [23:57:27] <Error_404> I actually haven't used windows in about 7 years [23:57:32] <delewis> tigerssb: who cares [23:57:37] <tigerssb> lol [23:57:38] <delewis> only lusers actually care what Windows is up to [23:57:47] <Error_404> it works for some people i guess [23:57:55] <boyd> I have a dual boot box that I use for Wintendo sometimes [23:58:22] <boyd> For me Windows is basically a console platform [23:58:23] <delewis> boyd: but I bet you wouldn't make it your lifetime goal to develop an operating system > Windows, like lusers do :-) [23:58:36] <delewis> they have Windows on the mind, you know. [23:58:43] <boyd> I see your point [23:58:50] <tigerssb> I need windows 4 a few apps, but leaning to solaris - avoiding sco tho - they are the ms of unix [23:59:01] * astinus sighs [23:59:09] <Error_404> stop using numbers for words [23:59:13] <delewis> tigerssb: actually, you'd be surprised how much they've given UNIX. [23:59:14] *** sickness has quit IRC [23:59:17] <incogito> clueless troll [23:59:21] <delewis> and you'd be surprised how much Microsoft has given to UNIX. [23:59:35] <delewis> do you have any clue where virtual consoles came from? [23:59:39] <delewis> I'll give you a hint [23:59:46] <delewis> it starts and ends with an 'x' [23:59:55] <tigerssb> so in other words, I switch to sco and pay them 800 bux? they can get lost