October 8, 2007 
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 | 31
October 8, 2007 [00:04:46] *** mega has joined #opensolaris [00:12:45] <l1s> huah, ext2 working great... [00:14:08] <l1s> hollenjf did the logs show up anything? [00:14:54] <hollenjf> logs? [00:15:56] <l1s> /var/log [00:23:06] *** dpn` has quit IRC [00:24:04] <hollenjf> nothing [00:24:13] *** hollenjf has quit IRC [00:24:16] <Triskelios> l1s: why would it lop to /var/log? [00:24:20] <Triskelios> *log, rather [00:25:22] <l1s> hm, thought its the same like any other unix system... [00:25:24] <l1s> or not?!? [00:25:45] <l1s> if something bad happens it should log the failures to syslog [00:25:50] <l1s> or not? [00:26:47] <Triskelios> this is a graphical wizard, not a daemon [00:29:07] <iron_angel> ext2 in Solaris? Nice. [00:30:05] <Triskelios> iron_angel: yes, it's available read-only, as is ntfs [00:30:47] <iron_angel> hmmm... Perhaps I should try poking at ext2/3 write support eventually. Alas for GPL license-incompatbility issues or the code could be taken straight from Linux. [00:31:39] <Triskelios> the ext2 and ntfs support actually runs in userspace to avoid the licencing issues [00:32:01] <l1s> hm, is the sun update manager used to update solaris from dvd? [00:32:38] <Triskelios> no. it downloads patches from sunsolve [00:32:40] *** mega has quit IRC [00:33:15] <l1s> how am i possible to update solaris within a running system? [00:33:22] <l1s> s/possible/able/ [00:33:48] <iron_angel> Triskelios: oh, heck, so like FUSE... In that case who cares if we have a GPLed driver... [00:33:48] <Triskelios> patches are installed on-line [00:34:11] <iron_angel> fully working ext2/3 would be useful for data interchange (thumbdrives, etc) [00:34:13] <gisburn> jmcp: ping! [00:34:15] <Triskelios> iron_angel: basically, yes [00:34:18] <gisburn> jmcp: ---> /msg [00:34:39] <gisburn> jmcp: /registeragainstnickserv, please [00:40:29] *** vmlemon has quit IRC [00:42:33] <jmcp> gisburn: done [00:44:14] *** ferret_0567 has quit IRC [00:45:28] <l1s> mhhhwhahahaha [00:45:33] <l1s> patching works from dvd [00:45:41] <l1s> you simply click on the install icon... [00:45:46] <l1s> and it installs any new package [00:46:00] <l1s> now i try to upgrade the system somehow... [00:46:03] <l1s> damn i love this system, [00:46:15] *** karrotx has joined #opensolaris [00:51:08] *** sior|away is now known as sioraiocht [00:54:51] *** Gman has quit IRC [00:56:25] *** jcea has quit IRC [01:09:25] *** deather_ is now known as deather [01:12:14] *** sfire||mouse has joined #opensolaris [01:19:34] <SYS64738> is there a way to clone a zone from a server to another one ? [01:21:54] <gisburn> SYS64738: cloning is not legal in the US... [01:22:16] <gisburn> (at leas not the cloning of humans) [01:22:23] <SYS64738> winners don't use drugs [01:22:36] * gisburn hides [01:23:09] <gisburn> SYS64738: AFAIK zones are bound to a specific server [01:23:29] <SYS64738> I have two cloned servers [01:23:36] <SYS64738> both with sx 69 [01:25:03] *** Gman has joined #opensolaris [01:27:11] <gisburn> SYS64738: try to ask in zones-discuss at opensolaris dot org [01:28:40] [01:29:04] <jamesd_> SYS64738, you can move/send zones from one machine to another... but it does require that the zone be halted i beleve. [01:29:31] <SYS64738> jamesd_, not with zfs snapshots [01:30:13] <jamesd_> SYS64738, not sure if zone importatation works if the zone was running when the snapshot was created. [01:36:41] <eryc> SYS64738: zoneadm detach [01:37:21] <SYS64738> ok [01:42:40] *** Pietro_S has quit IRC [01:45:10] <gisburn> Did anyone see comay recently ? [01:45:48] *** Somethingelse has quit IRC [01:49:52] *** sleepcat has joined #opensolaris [01:50:31] <l1s> bye all [01:50:32] *** l1s has quit IRC [01:50:52] *** sfire||mouse has quit IRC [01:53:56] <SYS64738> good night [01:55:34] <gisburn> !summon comay [01:55:38] *** cypromis has quit IRC [01:58:34] *** libkeiser has joined #opensolaris [02:07:16] *** sfire||mouse has joined #opensolaris [02:07:27] *** Gman has quit IRC [02:10:37] *** libkeise1 has quit IRC [02:11:00] *** NikolaVeber has quit IRC [02:11:48] *** yongsun has joined #opensolaris [02:13:31] *** chrisso has joined #opensolaris [02:13:42] <Triskelios> gisburn: what's the deal with the shared server for the summit? [02:15:26] *** TpentaNB has joined #opensolaris [02:17:26] <gisburn> Triskelios: it somehow got messed up. [02:17:37] <TpentaNB> 'ello from vegas roland [02:17:41] *** TpentaNB is now known as Tpenta [02:17:46] <gisburn> Tpenta: Hi! :-) [02:18:01] *** ChanServ sets mode: +o Tpenta [02:18:02] <gisburn> Tpenta: You're not a hostage of the local mafia there, right ? [02:18:08] <Tpenta> :) [02:19:32] *** cypromis has joined #opensolaris [02:20:02] <gisburn> At least here in germany many people think Las Vagas is infested by mafia, mobsters and Oceans 11/12/13/14.5-style thiefs [02:20:17] <gisburn> and each day an assasination [02:20:27] <delewis> gisburn: that hasn't been the case since the 1970s. [02:20:49] <gisburn> delewis: weeell... we can see that in the movies each day! [02:20:54] <gisburn> :-) [02:21:04] <delewis> sure, and those movies are based in the 1960s and 1970s, usually. :-) [02:21:36] * gisburn wonders whether it would be good for the world to send Mr. Bush to Dallas to a tour in a cabio [02:21:37] <gisburn> er [02:21:40] <gisburn> cabrio [02:21:47] <gisburn> Another bad joke ruined by typos [02:23:10] <cypromis> why take a detour on the road to hell ? [02:26:04] * Tempt grumbles about not being on a junket in vegas... [02:30:00] *** xinkeT has quit IRC [02:30:28] *** noobuntu has joined #opensolaris [02:32:28] *** yongsun has quit IRC [02:37:55] *** xinkeT has joined #opensolaris [02:43:53] *** deedaw has quit IRC [02:46:07] *** Disreali has joined #opensolaris [02:46:19] <sleepcat> how hard is it to install oracle on solaris 10? [02:46:41] <Tempt> sleepcat: Not very. [02:46:46] * delewis isn't quite sure how to answer that question [02:46:51] <delewis> it's easy if you've done it before. :-) [02:46:56] <sleepcat> good, cause I'm downloading 10g right now [02:46:57] <Tempt> sleepcat: vi /etc/system ; vi /etc/project ; run the oracle installer [02:47:04] <Tempt> everything is easier with 10gr2 [02:47:17] <sleepcat> it looks like you can download 10g and play with it at no cost [02:47:19] <paulf> I'm only lacking one thing to tell you the answer: any knowledge of installing Oracle :) [02:47:25] <Tempt> sleepcat: You can. [02:47:34] <sleepcat> that's what i'm doing [02:47:52] <sleepcat> now I can catalog my pr0n [02:48:26] <sleepcat> actually, i'd probably need a license for that... [02:49:07] <sleepcat> so how's zfs compared to ufs on oracle? [02:49:15] <delewis> acceptable. [02:49:19] <sleepcat> not too slow i hope [02:49:33] <sleepcat> cause every filesystem except for / is zfs on my machine [02:49:36] <delewis> it won't match the performance of UFS after tuning, but it'll get fairly close. [02:49:43] <delewis> I use ZFS on my Oracle development systems. [02:49:48] <delewis> snapshots are *extremely* useful. [02:49:59] <sleepcat> its not like i'm doing anything production wise, just goofing off with oracle [02:50:07] <delewis> ZFS will be fine then. [02:50:13] <delewis> throw it in a zone that's backed by ZFS. [02:50:18] <sleepcat> prentending I'm a big shot admin running a huge corporation [02:50:25] <delewis> that way you can easily rollback the snapshot should you fudge the Oracle install or database creation. [02:50:35] <sleepcat> delewis: good tip [02:50:59] <sleepcat> i've created snapshots, have yet to rollback to one though [02:51:28] <sleepcat> i'm at 11% of the download right now [02:51:40] <delewis> just create a snapshot after you create the zone and another one after you install Oracle, but before you create a database with 'dbca' [02:51:49] <delewis> (make sure the zone is halted when you create the snapshots) [02:52:07] <sleepcat> can oracle be installed on a zone? [02:52:18] <delewis> if you fudge the install, you just do zfs rollback export/zones/<name of your Oracle zone>@pre-oracle (or whatever) [02:52:25] <delewis> sleepcat: absolutely. [02:53:20] <sleepcat> i have a lot of postgresql experience, and I tried oracle express on windows. [02:53:21] <delewis> DB2 on the other hand.. [02:53:40] <sleepcat> yeah I avoid DB2 because I'm more familair with MVCC [02:53:54] <Triskelios> does it make sense to use zvols for Oracle DB files? [02:54:01] <delewis> no. [02:54:36] <delewis> with Oracle the only acceptable ways to back your tablespaces, nowadays, are via files and ASM. [02:55:05] <delewis> managing raw tablespace (even if they are zvols) is a PITA, by comparison. [02:55:25] *** Tpenta has quit IRC [02:55:26] <sleepcat> Triskelios: you have an oracle install also? [02:55:39] <sleepcat> does it oracle work on SXCE? [02:55:44] <delewis> yes. [02:55:49] <Triskelios> sleepcat: I don't have an Oracle install, I was just hypothesizing [02:55:52] <delewis> you'll get an error or two during the install process, but just ignore it. [02:56:07] <delewis> it'll be during the part where Oracle is linking. [02:56:08] <sleepcat> delewis: i usually ignore errors anyhow... [02:56:21] <delewis> in almost 99% of all installations, those linking errors are irrelevant. [02:56:22] <sleepcat> delewis: I have solaris 10 u4 right now [02:56:31] <delewis> (a couple of binaries fail to link that most people won't be using) [02:56:49] <sleepcat> i'm probably going to create a data warehouse [02:57:04] <sleepcat> i've got ralph kimball's book [02:57:12] <Tempt> Oracle 10 on SXCE 32bit will not work properly, FYI. [02:57:23] <sleepcat> Tempt: good to know [02:57:34] <Tempt> Too many link problems, it'll choke and die. [02:57:41] <sleepcat> because that's what i'd be using if i were to use sxce [02:57:52] <Tempt> Well, think about getting a 64bit CPU ;-) [02:58:01] <delewis> Tempt: those same linking problems occur on SPARC. They're safe to ignore, unless these are different linking errors. [02:58:08] <Tempt> Nope [02:58:11] <Tempt> Different linking problems [02:58:18] <sleepcat> i have 2 Netra T 105s under my couch [02:58:29] <Tempt> On SXCE x86 32bit, critical parts won't link. [02:58:39] <delewis> ah. [02:58:40] <Tempt> even sqlplus will fail to link [02:58:48] <sleepcat> with SPARC IIi [02:58:49] <dlg> mmmt1105 [02:58:58] <delewis> sleepcat: you probably won't want to run Oracle on those. [02:59:12] <sleepcat> delewis: I can't get into them anyway [02:59:15] <delewis> I'd say you need at least 1GB of memory to run 10gR2 with Enterprise Manager. [02:59:17] <sleepcat> NVRAM password [02:59:22] <delewis> T1 105s tend to be pretty thin on memory. [02:59:26] <Tempt> drop EM and you drop a lot of requirements. [02:59:48] <sleepcat> I can use one for parts and double up on the memory [03:00:17] <sleepcat> i was thinking of pulling one of the SCSI disks so I have 2 10k disks in one netra [03:00:33] <sleepcat> plus yanking the memory out of one also [03:00:44] <delewis> makes sense. [03:00:47] <sleepcat> man oracle's pipes are fast [03:00:58] * delewis just uses his E4500 and Tadpole for Oracle [03:01:21] * sleepcat jealous of delewis's tadpole [03:01:35] <kjetilho> please guys, get a room [03:01:40] <iron_angel> Incidentally, how expensive do 1.2GHz CPU modules for Blade 1k/2k tend to be? [03:01:44] <delewis> makes for a decent portable Oracle server -- 4GB of memory. [03:01:49] <delewis> iron_angel: not cheap. [03:01:54] <delewis> > $500 in any case. [03:01:58] <iron_angel> ouch. [03:02:48] <delewis> the mininum you're going to see a 2x1.2GHz 280R or Blade 2000 config going for is $2,000. [03:02:55] * Tempt just uses a v880 for Oracle ... [03:03:16] <delewis> and given 2x750MHz 280R and Blade 2000 configs typically go for $500, you can see where most of the hurt is coming from. [03:03:57] <delewis> (well, 2x900, given Blade 2000s didn't ship with 750s) [03:04:34] <delewis> the 1.2GHz modules are fast, though. [03:05:14] <delewis> I've got a 280R with 2 of them, and there's a huge difference between the other 280R with 2x900MHz modules. [03:13:02] *** comay has joined #opensolaris [03:13:04] <iron_angel> delewis: Mine's a 2x900, that's why I was hoping to find the 1.2GHz modules. [03:13:21] <Tempt> The 900s are pretty good [03:13:36] <Tempt> I wouldn't worry about upgrading to the 1.2s; it won't be that helpful unless you're really CPU starved. [03:14:20] <iron_angel> true. [03:14:29] <sleepcat> is sparc even competitive in the lower end market? [03:14:34] <iron_angel> I might put my second XVR-1000 in it, though, instead of in the Ultra 80. Still deciding. [03:14:39] <iron_angel> sleepcat: define 'lower end'. [03:14:52] <sleepcat> the x86 space [03:15:11] <iron_angel> The V210/V215 compete pretty well with low-end/lower-midrange 1U servers, but the Ultra 20 is no match for a $300 Dell, for example. [03:15:18] <iron_angel> Except in I/O and reliability. [03:15:31] <sleepcat> the amd chips trounce SPARC from what I've heard [03:15:40] <delewis> for single-threaded workloads, yes. [03:16:13] <iron_angel> well, twice the clock speed does help, even if the SPARC has slightly better IPC. [03:16:25] <delewis> and given the lower-end SPARC systems (with the exception of the T1000/T2000) are usually at most 2-way systems, an x86-64 system will own them. [03:16:36] <Tempt> "Ultra 20 is no match for a $300 Dell" [03:16:40] <Tempt> please explain? [03:16:47] <delewis> iron_angel: IPC is irrelevant on a 2-way system. [03:17:32] <sleepcat> are the SPARC IIi chips really slow? [03:17:45] <delewis> the only place SPARC has any market whatsoever is in the high-end market (where SPARC's scalability pays off) and the UltraSPARC-T1/-T2 systems. [03:17:52] <delewis> sleepcat: yes. [03:18:05] <Tempt> and for test and development platforms that go with the higher end stuff [03:18:22] <Tempt> I'm working with V245s a lot this week, building model clusters to go with the real deployment on 890s and up [03:18:53] <sleepcat> how does a quad core dual CPU system from intel compare to a 890? [03:19:05] <sleepcat> with lets say 4 gigs of ram [03:19:06] <delewis> sleepcat: IO-wise the v890 *might* win. [03:19:12] <delewis> for single-threaded workloads, the x86 system will win. [03:19:17] <Tempt> And the 890 will definately have more than 4Gb of RAM [03:19:21] <delewis> and maybe for multi-threaded workloads, the v890 might win. [03:19:23] <Tempt> I don't think they ship them that small anymore [03:19:31] <sleepcat> well, 8 gigs of ram [03:19:42] <Tempt> That's an entry level config then [03:19:48] <Tempt> so two CPUs and 8Gb of RAM [03:19:53] <Tempt> The x86 will win. [03:20:02] <Tempt> But a stacked 890 will eat it for breakfast [03:20:18] <sleepcat> but cost wise, the x86 will win [03:20:19] <sleepcat> :-( [03:20:20] <delewis> SPARC is pretty much irrelevant nowadays in the low-end. [03:20:34] <delewis> in the mid-range and high-end, SPARC still has IO and reliability going for it. [03:20:38] <delewis> but that's about it, really. [03:20:41] *** wnorrix_ has joined #opensolaris [03:20:46] <Tempt> sleepcat: Cost-wise, the nastiest whitebox hardware from the corner chinese computer shop will always win [03:20:53] <Tempt> sleepcat: But do you want to use it? [03:21:03] <sleepcat> sad [03:21:19] <delewis> the only reason I'm still buying SPARC gear for my personal use is that I can load it up with memory fairly cheaply. [03:21:20] <g4lt-mordant> delewis, until niagra2 makes you a liar ;P [03:21:24] <Tempt> white box yum-cha is always going to be cheaper than any branded hardware [03:21:33] <delewis> g4lt-mordant: as always, I'll believe it when I see it. [03:21:34] <iron_angel> mips, pa-risc and alpha have died alyogether, it's a bit sad [03:21:45] <iron_angel> at least power is still cooking [03:21:52] <delewis> given the actual, real-world performance that Niagara1 turned out to produce, I'm not too enthuasistic about Niagara2. [03:21:53] <Tempt> Yep, POWER is still cooking [03:21:55] <sleepcat> i think the us government should prop sparc up [03:21:58] <Tempt> (literally, they run toasty-warm) [03:22:13] <sleepcat> give sun a bunch of government contracts [03:22:20] <Tempt> sleepcat: Nah, telcos prop SPARC up [03:22:21] <iron_angel> they do. [03:22:33] <sleepcat> maybe they can use sun gear to monitor our phone calls and irc [03:22:38] <delewis> and the financial sector. [03:22:39] <iron_angel> lots of it has gone amd64 lately [03:22:58] <delewis> I'm curious to see how much SPARC gear my new employer has. [03:23:12] <delewis> (we're a large HP and Sun customer) [03:23:13] <g4lt-mordant> delewis, part of that is I'm sorry, but the T1K is NOT where I'd put a screamingly fast low power consumption dualcore. I'd be looking at pizzaboxes/lunchboxes again [03:23:14] <sleepcat> i saw a comcast commercial on TV and they were talking about VOIP and they displayed their datacenter. It was all sun gear [03:23:38] <iron_angel> seriously, the pizza box needs to come back. [03:23:56] <sleepcat> i have 2 pizza boxes under my couch [03:24:13] <sleepcat> i just need to get my rj45-db25 cable [03:24:21] <sleepcat> so I can load sol 10 on them [03:24:40] <iron_angel> SPARC pizza boxes? [03:24:45] <delewis> sleepcat: they were using SPARC gear to monitor our phone calls. [03:24:54] <sleepcat> netra T 105s [03:24:55] <delewis> Narus was using a v890 for the Oracle back-end. [03:25:00] <iron_angel> ahh, cool. [03:25:32] <sleepcat> well one good think about their tapping our phone lines is that it gives SUN money to develop solaris [03:25:39] <iron_angel> I was thinking of pizza-box workstations like the SS20. [03:25:42] <kjetilho> SPARC is really slow on branch happy code == compilers :-/ [03:26:12] <sleepcat> and then I can download solaris 10 for free courtesy of uncle sam propping up sun. [03:26:20] <g4lt-mordant> I was thinking more the SS1/2/4/5/10/20/ultra 1/2/5 [03:26:31] <Tempt> Well, the Telco market is still a huge SPARC market [03:26:40] <Tempt> and some telco specific apps are only developed for SPARC/Solaris [03:26:49] <iron_angel> I have both an SS20 and U2 here, actually. [03:26:54] <iron_angel> But I need to fix the SS20. [03:27:46] <sleepcat> yeah, I was just on oracle's site. They had a pdf on the history of oracle , and oracle started with CIA cash [03:27:47] <Tempt> delewis: New job? [03:27:58] *** wnorrix has quit IRC [03:28:01] * delewis is trying to decide whether his next hardware purchase should be a second-hand v480 or v880 or a new x4100 or x4150. [03:28:34] <delewis> Tempt: yep @ FedEx Services (the branch of FedEx that does IT for the other branches, like FedEx Corp., Freight, etc.) [03:29:03] <sleepcat> delewis: a lot of djkstra's shortest path alogrithms.. [03:30:23] *** yongsun has joined #opensolaris [03:32:33] <Tempt> delewis: An 880, of course. [03:32:44] <delewis> you'd say that, of course. :-) [03:33:09] <delewis> I might do a try-and-buy on an x4100 or x4150 (whenever it hits the try-and-buy list) to see how it performs. [03:33:23] <sleepcat> delewis: try and send back you mean [03:34:12] <delewis> if I'm not satisfied with it, I'll go for the x480, v880, or maybe an E4800. [03:34:17] <Tempt> 6900 [03:34:20] <delewis> s/x480/v480/ [03:34:26] <delewis> too large and still too expensive. :-) [03:34:38] <delewis> I did see an E6800 for $5k, though. [03:36:24] <sleepcat> http://www.acclinet.com/sun_servers/midrange/sun_fire_e6800/index.asp [03:36:39] <sleepcat> that would impress the babes [03:36:50] <sleepcat> A HUGE computer in your living room [03:37:12] <sleepcat> does it run solaris 10? [03:37:19] <Tempt> Of course it does [03:37:19] <g4lt-mordant> sleepcat, depends if you had a sunvideo on it and used it as a faux-tivo ;P [03:37:53] <sleepcat> that'd be kind of cool to aquire as a hobbyiest [03:38:27] <sleepcat> What to do with it though...1. install oracle...2. ... [03:38:37] <delewis> Sun Ray Server [03:38:40] <sleepcat> 3. get laid [03:39:01] <sleepcat> hey honey look at my super computer [03:39:12] <delewis> I'd be looking at something more mid-range, like an v880 or E4800. [03:39:27] <delewis> the E6800 is too large even for me, and I certainly don't have the electrical or cooling requirements for it now. [03:39:37] <kjetilho> anyone here installed Solaris on Dell PE1950? [03:39:49] <sleepcat> delewis: would that run postgresql? [03:40:10] <iron_angel> kjetilho: I've installed it on a PE1850, but not a 1950. [03:40:13] <kjetilho> I'm trying to apply a 3rd party driver for the PERC5, but it has a conflict with mpt on the PCI-id. [03:40:13] * delewis laughs out loud. [03:40:19] <sleepcat> 24 SPARC CPUs... [03:40:22] <delewis> sleepcat: why wouldn't it? [03:40:41] <sleepcat> i think postgresql konks out at 8 cpus [03:40:42] <Tempt> I'd like to see any intel box keep up with a fully loaded 6900 [03:40:42] <delewis> could PostgreSQL scale over that many processors well? probably not, but it'd work. [03:41:07] <g4lt-mordant> delewis, just stick it out in the snow, it'll stay cool ;P [03:41:23] <kjetilho> iron_angel: I'm wondering if I should replace the mpt driver with this mfi driver, or simply bind only pci1028,50 to mfi, not the others. [03:41:24] <sleepcat> would linux even be able to handle a e6800 [03:41:24] <delewis> g4lt-mordant: we get almost zero snow here. [03:41:37] <delewis> Linux? on a 6800? what a waste. [03:41:46] <delewis> Linux can't even utilize an E4500, which is 10-year-old gear. [03:41:50] <sleepcat> delewis: I agree. [03:41:53] <delewis> it certainly couldn't utilize an E6800. [03:41:54] <Tempt> No hardware support and I don't see the scheduler handing 24 CPUs properly [03:41:57] <g4lt-mordant> delewis, then send it to me, I can provide the snow [03:42:06] <delewis> g4lt-mordant: can you provide the money, too? :-) [03:42:23] <sleepcat> the one i linked to was 36,000 dollars [03:42:26] <iron_angel> actually, the scheduler does OK with large numbers of CPUs, but drivers are the real issue. [03:42:33] <sleepcat> i wonder what it costs new [03:42:40] <Tempt> pfft [03:42:41] <delewis> sleepcat: I've seen a 24-way config on eBay for $5,000 recently. [03:42:48] <iron_angel> kjetilho: the latter is what I'd do, I think, but I'm not sure. [03:42:49] *** kFuQ has joined #opensolaris [03:42:52] <Tempt> From what I've seen the linux schedular fails horrible with more than 2 CPUs [03:42:57] <sleepcat> delewis: wow [03:42:59] <Tempt> unless you're running custom Altix style stuff [03:43:00] <sleepcat> that's cheap [03:43:12] <iron_angel> Tempt: have you looked at that since, oh, 2.6.0? [03:43:24] <Tempt> Oh, come on, version of the week bullshit [03:43:27] <sleepcat> why is it so cheap? That thing must cost much more just to manufacture [03:43:28] <g4lt-mordant> Tempt, ask brendang, he'll tell you precisely how the scheduler breaks in minute detail ;P [03:43:33] <Tempt> every other UNIX has been doing this since ... god knows when [03:43:39] <paulf> Tempt: actually, SGI ProPack uses a basically stock Suse enterprise kernel [03:43:57] <Tempt> It's not even a real operating system anyway [03:44:02] <paulf> I can't speak to how well it works, though [03:44:02] <kjetilho> iron_angel: ok thanks, it's booting now, we'll see :) [03:44:30] <Tempt> seems like scheduling is the linux argument of the month [03:44:35] <Tempt> never stop fighting over it [03:45:14] <iron_angel> Tempt: No, I'm not claiming that the latest version works while version -1 doesn't - I'm claiming that 2.6.x does a much better job of it that 2.4.x. (Which is not to say that it's better than Solaris, BSD, etc - just better than an earlier version of itself.) [03:45:15] <delewis> it's basically what all the UNIX vendors went through in the mid-90s, and the solution ended up being to provide multiple schedulers, so that users with whatever workload could pick the right scheduler. [03:45:49] <delewis> (without recompiling) [03:46:14] <sleepcat> what about a scheduler running multiple schedulers depending on the workload? [03:46:20] <delewis> does Linux still require you to recompile if you want to pick a different scheduler? [03:46:26] <delewis> sleepcat: welcome to Solaris. [03:46:39] <sleepcat> delewis: how do i play with that feature? [03:46:46] <g4lt-mordant> iron_angel, here's my question to you, what would you consider the maximum number of propcs linux can use effectively? [03:46:57] <Tempt> 20? [03:47:10] <Tempt> Oh, number of CPUs, not number of processes ;) [03:47:14] <Tempt> half, perhaps? [03:47:16] <Tempt> a quarter? [03:47:23] <sleepcat> 8 [03:47:38] <sleepcat> it seems systems desinged on x86 tend to max out on 8 cpus [03:47:47] <sleepcat> postgresql has the same issue [03:47:50] <g4lt-mordant> you realize taht the v880 that's been bandied about can hold that many? [03:47:53] <delewis> sleepcat: it's not "it seems", but design flaws. [03:47:58] <delewis> x86 just doesn't scale well. [03:48:11] <delewis> and x86-64 currently doesn't scale well (not compared to SPARC, anyway) [03:48:46] <delewis> Fujitsu makes monolithic SPARC systems with > 1024 processors. [03:48:48] <delewis> *monolithic* [03:48:52] <sleepcat> delewis: do you think we'll see a resurgence of SPARC once moore's law effectively ends within the next few years? [03:49:11] <iron_angel> g4lt-mordant: I'd peg it at 24-32, but that's not very well tested. The scheduler scales OK, but other stuff might not. [03:49:24] <Tempt> T1 already caused a SPARC resurgence [03:49:24] <delewis> sleepcat: no. Intel will beat Sun any day of the week when it comes to fabrication costs, and thus, Intel can produce a processor much cheaper than Sun can. [03:49:29] <Tempt> T2 will do that again [03:49:32] <delewis> and that cost effects what the customer pays. [03:49:44] <Tempt> and if "rock" ever happens it'll be another kicking [03:49:44] <g4lt-mordant> 32 is a typical 5-digit sun enterprise, and solaris runs on it just fine [03:49:46] <delewis> and now SPARC is being attacked by POWER. [03:49:56] <Tempt> But not in the desktop space, and not in the dell-customer space [03:50:13] <sleepcat> what is the future of SPARC then? [03:50:33] <iron_angel> Frankly, that depends a lot on POWER. [03:50:42] <delewis> as of right now, limited to customers that need to run SPARC-only applications for the forseeable future, unless Niagara2 and Rock actually do something. [03:50:50] <Tempt> POWER is quite expensive though. [03:51:03] <iron_angel> SPARC and POWER are fighting it out for the high-end stuff, while x86/amd64 have a stranglehold on the low-end. [03:51:04] <Tempt> and IBM isn't as good at Sun at getting out there and marketting it (at least not in my experience) [03:51:04] <delewis> Tempt: no worse than SPARC. [03:51:22] <Tempt> delewis: Well, IBM mark up non-US markets by 200% [03:51:25] <iron_angel> They're fighting their way toward the high-end, but it's still largely saddled with lots of old DOS-era cruft. [03:51:29] <Tempt> delewis: a "not in America" tax [03:51:33] <delewis> Tempt: non-US. :-) [03:51:51] <iron_angel> And also, SPARC and POWER don't have to worry about breaking Windows. [03:51:53] <Tempt> delewis: Well, Sun sells a lot of hardware outside of the US; you see a lot of Sun in DCs [03:52:00] <sleepcat> if for example, intel is unable to increase clock speeds, it has to then add more cores to the chips. Since SPARC is better at scaling across CPUS this may be a chance. [03:52:03] <Tempt> delewis: But you very rarely see AIX around here. [03:52:10] <delewis> Tempt: interesting. [03:52:17] <Tempt> Too expensive. [03:52:19] <g4lt-mordant> iron_angel, given itanic, neither does X86 [03:52:25] <delewis> sleepcat: where scalability becomes a concern is for large, large systems. [03:52:30] <Tempt> The cheapest box you can get (p505) sells for around $AU10k [03:52:42] <delewis> Tempt: what's that in USD? [03:52:43] <Tempt> (yes, that's a lot more than IBM US webprice) [03:52:47] <delewis> base p505 config goes for $3k list over here. [03:53:08] <sleepcat> 3 k is pretty cheap, does that include AIX license? [03:53:11] <delewis> no. [03:53:16] <delewis> that's hardware-only. [03:53:17] *** Disreali has quit IRC [03:53:18] <iron_angel> g4lt-mordant: heh, itanic is, err, weird. Remember UltraJAVA/MAJC? Well... Intel tried to muscle it. [03:53:25] <sleepcat> AIX is expensive [03:53:31] <delewis> it's not that bad. [03:53:39] <sleepcat> compared to solaris 10 [03:53:41] *** theRealBallchalk has joined #opensolaris [03:53:54] <sleepcat> plus you don't even have dtrace nor zfs [03:53:55] <delewis> sleepcat: talk to some actual Sun customers. [03:54:12] <delewis> right now, the stability of Solaris 10 is questionable in some circles. [03:54:21] <sleepcat> delewis: how so? [03:54:26] <sleepcat> the number of patches? [03:54:35] <sleepcat> i seem to get security patches every week [03:54:39] <Tempt> $AU10k is $US9014.20 [03:54:45] <delewis> that, panics with ZFS, and some other oddities. [03:54:52] <delewis> Tempt: ouch. [03:54:58] <Tempt> man, the AU dollar just keeps climbing against the greenback [03:55:04] <Tempt> I should order some crap online or make some ebay purchases [03:55:07] <delewis> everything is climbing against the USD. [03:55:12] <Tempt> celebrate the devalued USD [03:55:15] <delewis> the peso is worth $0.10 now. [03:55:22] <theRealBallchalk> hmmmmmm [03:55:22] <delewis> 10 years ago it was worth barely $0.01. [03:55:29] <Tempt> bloody hope the USD stays where it is [03:55:42] <theRealBallchalk> the US is at war so shouldn't that improve the economy? [03:55:42] <Tempt> at this rate OPEC will switch completely to euros [03:55:45] <delewis> Tempt: it's a conspiracy to create a global currency. [03:55:50] <theRealBallchalk> defense spending is good [03:55:53] <delewis> everything higher than the USD is falling [03:55:58] <delewis> and everything lower than the USD is gaining. [03:56:01] <Tempt> or flanian pobble beads [03:56:08] <Tempt> altairian dollars? [03:56:10] <e^ipi> theRealBallchalk: not if you've got a fiat currency and start printing money to pay for it [03:56:12] <sleepcat> ah my oracle download finished [03:56:18] <theRealBallchalk> Tempt: the canadian dollar is higher than the US now [03:56:19] <cypromis> hmm delewis, why is the pound growing than ? [03:56:20] <cypromis> lol [03:56:33] *** mcnamarabrian has joined #opensolaris [03:56:37] <delewis> cypromis: you're failing to look at the overall trend. [03:56:46] <theRealBallchalk> e^ipi: true then in that case then i'd need a really good printer to print greens [03:56:53] <delewis> the *overall* trend of the of the pound is falling. [03:56:56] <theRealBallchalk> shieah i'd do it [03:56:57] <theRealBallchalk> jk [03:57:00] <delewis> and the same goes for the euro. [03:57:14] <delewis> the pound used to be roughly twice as much as the USD. [03:57:21] <delewis> and now it's what? 1.5? 1.25? [03:57:27] <sleepcat> the us administration is devaluing the dollar so its war costs are not as expensive [03:57:30] <e^ipi> theRealBallchalk: that's what the US is doing... between borrowing money from the chinese & printing off more money, anyways [03:57:31] <cypromis> ??? [03:57:36] <cypromis> you kidding me ? [03:57:48] <cypromis> 1 ound = 2. something USD [03:57:49] <sleepcat> the dollar drop is alway expected in times of war [03:57:51] <kjetilho> delewis: 2.03 [03:58:17] <kjetilho> but it's not long ago the pound was 14 NOK, now it's 11. [03:58:19] <cypromis> 2.04 today [03:58:20] <cypromis> :) [03:58:22] <sleepcat> at least people in the EU can afford to buy SUN chips [03:58:28] <delewis> what's the euro at then? [03:58:35] <theRealBallchalk> i dunno i think we got a monkey in the White House man someone throw that guy in the Washington zoo or something [03:58:42] <cypromis> 1.44 EUR = 1GBP [03:58:54] <cypromis> so 1 EUR = 1.44 USD [03:59:00] <sleepcat> man oracle uses a lot of java [03:59:02] <cypromis> sorry, 1.41 [03:59:19] <sleepcat> SUN probably has oracle and IBM by the balls with java [03:59:28] <kjetilho> cypromis: don't be a square, root [03:59:50] <gisburn> can anyone check comay's cube whether he died there ? [04:00:21] <sleepcat> who is cornay? [04:03:04] *** cypromis has quit IRC [04:03:28] *** gm152 has joined #opensolaris [04:07:13] *** jcsmith has quit IRC [04:14:47] *** jcsmith has joined #opensolaris [04:20:20] <g4lt-mordant> gisburn, he heard you were looking for him and went to antarctica [04:20:25] *** palowoda has quit IRC [04:20:56] *** mcnamarabrian has quit IRC [04:21:08] *** troff has joined #opensolaris [04:21:49] * gisburn waits for troff meeting groff and tbl [04:22:21] <gisburn> eqn can't make it today [04:28:21] *** Fullmoon has quit IRC [04:30:22] <kjetilho> hooray, my PE1950 is finally being jumpstarted :-) [04:30:49] <dlg> which controller? [04:33:58] <kjetilho> PERC5 something. your driver, I reckon :) [04:34:04] *** AtomicPunk has quit IRC [04:34:17] *** AtomicPunk has joined #opensolaris [04:34:17] <sfire||mouse> pe1950 tends to have perc5 [04:34:28] <sfire||mouse> at least the ones at work [04:34:53] <kjetilho> yes, except perc5 != perc5 :-/ [04:35:09] <sfire||mouse> ?? [04:35:22] *** iron_angel has quit IRC [04:35:29] <kjetilho> there are several possible PCI-ids, it seems. [04:36:31] *** comay has quit IRC [04:36:36] <sfire||mouse> ah [04:36:41] <kjetilho> but I could be confused. [04:36:46] * dlg look annoyed [04:36:57] <kjetilho> ? [04:37:10] *** sleepcat has quit IRC [04:40:26] <paulf> g'night [04:40:26] *** paulf has quit IRC [04:46:37] *** gisburn has quit IRC [04:47:53] *** lloy0076 has joined #opensolaris [04:49:07] <lloy0076> http://opensolaris.pastebin.ca/729558 [04:49:36] <lloy0076> Why does "less" suddenly think any text file is a binary file, and when I open something up I often see "^[a-zA-Z]" escapes...?? [04:50:33] *** chrisso_ has joined #opensolaris [04:58:48] *** cewx has joined #opensolaris [04:59:53] <kjetilho> dlg: why were you annoyed? [05:02:17] *** cewx has left #opensolaris [05:23:01] *** kaiwai has joined #opensolaris [05:24:38] <dlg> kjetilho: support for the perc5s should be in opensolaris by default [05:25:40] <kjetilho> ah, but I must admit I'm currently installing 10u3 [05:29:56] *** iron_angel has joined #opensolaris [05:33:46] *** chrisso__ has joined #opensolaris [05:34:18] *** sleepcat has joined #opensolaris [05:34:39] <sleepcat> ORA-27102 [05:34:57] <sleepcat> Getting an out of error message on my oracle 10 G install [05:35:03] <sleepcat> have solaris 10 [05:38:36] <Tempt> Did you edit /etc/project? [05:40:04] <sleepcat> no [05:40:11] <sleepcat> am I supposed to? [05:41:30] <Tempt> Did you read the docs? [05:41:35] <sleepcat> http://opensolaris.pastebin.ca/729594 [05:41:54] <Tempt> [10:46] <Tempt> sleepcat: vi /etc/system ; vi /etc/project ; run the oracle installer [05:42:20] <Tempt> It's all in the install guide [05:42:29] <Teltariat> Greetings and salutations to you all [05:42:35] <sleepcat> ok I guess I'll have to read. Thanks Tempt [05:46:56] <kaiwai> hmm, gcc 4.2.2 has been released (in the cvs any how) [05:48:30] <sleepcat> is it hard to increase the swap size in solaris 10? [05:48:35] <jmcp> no [05:48:47] <jmcp> swap -a /dev/dsk/$swapdev [05:49:05] <sleepcat> jmcp: what's that do? [05:49:14] <jmcp> adds another swap partition or swap file [05:49:23] <sleepcat> and its safe? [05:49:27] <jmcp> safe? [05:49:35] <Bartman007> sleepcat: safe if you don't add the wrong device ;-P [05:49:36] <sleepcat> it won't turn my system into a brick [05:49:42] <jmcp> why would it do that? [05:50:04] <sleepcat> so I just type that into my shell and I should be good to go? [05:50:10] <sleepcat> here it goes... [05:50:26] <Teltariat> lol, swap your root device [05:50:43] <Bartman007> sleepcat: you do understand that $swapdev should be replaced with the device you want to use, right? [05:50:50] <jmcp> sleepcat: libdiskmgmt should tell you whether a partition is already in use [05:50:59] <g4lt-mordant> swap on a zpool for ultimate psin [05:51:09] <sleepcat> can i swap on zfs [05:51:18] *** lloy0076 has left #opensolaris [05:51:19] <kaiwai> IIRC I don't think you can [05:51:20] <jmcp> you'd need to use the zvol, iirc [05:51:20] <sleepcat> I have a zpool... [05:51:36] <g4lt-mordant> I was JOKING FFS [05:51:48] <sleepcat> can I steal some space from my zpool and claim it for swap? [05:51:49] <jamesd_> sleepcat, you should avoid it like the plague.... even if it is possible. [05:51:54] <Teltariat> So I spent 2 days trying to setup ZFS boot with Solaris 8/07, which a certain someone here told me did include the ZFS boot bits. Only to find out that I'd been completely wasting my time because no, sol_10-8/07 does _not_ include ZFS boot bits [05:52:07] <jmcp> sleepcat: if you want to do that, rtfm for "zfs create" first [05:52:08] <Teltariat> yay [05:52:20] <jmcp> g4lt-mordant: it's perfectly possible ... but as always, you should have a clue about what you're doing first :) [05:52:31] *** Drone has quit IRC [05:53:05] <g4lt-mordant> jmcp, well, it had best be, or zfs / is going to be a major pain eveen mor than usual [05:53:13] <delewis> Swapping to a file seems to be gaining popularity in Linux, nowadays, instead of creating a seperate swap partition. Little do these morons know what overhead the filesystem places on swap performance. [05:54:04] <iron_angel> some of it is idiocy, some of it is figuring that on a workstation/desktop, you're unlikely to need to swap anyway, and if ever you do, interactive performance is apt to suck. [05:54:18] <iron_angel> Of course, some of *that* comes from Windows. [05:54:24] <iron_angel> (the mindset, that is) [05:54:49] <delewis> iron_angel: that's what most people don't understand. On UNIX systems where not all the workloads are interactive, swap isn't necessarily bad. [05:55:11] <delewis> I get irritated at these so-called 'admins' that think any amount of swapping is bad, and that it automatically implies you need more physical memory in the system. [05:55:20] <Bartman007> swap is a necessary evil IMO. [05:55:25] <jmcp> sleepcat: read http://www.solarisinternals.com/wiki/index.php/ZFS_Best_Practices_Guide [05:55:46] <jamesd_> iron_angel, with the latest enhancements to the desktop you need swap more than ever... if you are switching users do you really want both enviroments sitting in ram? do you really need all of kde start up code left in ram, i would perfer to use it as disk cache. [05:55:47] <delewis> Bartman007: not if the workload isn't continuously IO-intensive, and if the pages can be brought into memory quickly enough. [05:56:06] <Bartman007> delewis: machines should be designed to not use it, but it should still be available. [05:56:10] <delewis> in that case paging is absolutely acceptable. [05:57:22] <iron_angel> jamesd_: I agree with that, actually. [05:57:25] <delewis> Bartman007: so you'd rather have processes that are sleeping most of the time draining physical memory? [05:57:27] <Bartman007> delewis: exactly, if you have a heavy load, the machine should be spec'd accordingly, but swap should still be available to sue. [05:57:48] <Bartman007> delewis: no, poor wording on my part. [05:59:17] <delewis> where swapping becomes bad is when it effects the iowait for a particular disk (especially, if this disk is shared with other filesystems) and the working set of pages is too large to be brought into physical memory (this will in turn effect the iowait for that disk). Each of these can be measured and quanitified. [06:01:12] <delewis> now, if you've got a process that's sleeping most of the time, and it's working set of pages can be easily brought into memory without having the other processes' working sets paged out is swapping necessarily evil? [06:02:39] *** mritr has joined #opensolaris [06:02:52] <Teltariat> jmcp: Thanks for the ZFS best practices link; thats really useful [06:02:59] <jmcp> you're welcome [06:03:08] <delewis> I'm just saying that seeing swap usage doesn't necessarily imply you need to increase the amount of physical memory in the system immediately. The impact of swapping for a workload should be weighed and if service requirements aren't being met, increase physical memory is probably needed. [06:07:05] * steleman yawns [06:07:18] <Bartman007> delewis: fair enough. I guess I've just dealt with too many poor swapping algorithms [06:08:24] *** rubymonk has quit IRC [06:10:17] <jteo> such as the one in opensolaris? ;) [06:11:50] <Bartman007> jteo: no comment, I'm a new convert. Last time I played with Solaris in the mid 90's and I wasn't even a teen yet :-P [06:12:05] <Bartman007> haven't run into swapping issues yet. [06:12:26] <jteo> ;) [06:14:36] *** gm152 has quit IRC [06:15:36] *** halton has joined #opensolaris [06:20:42] <Triskelios> are single directories in zfs implemented as trees? [06:22:05] <sleepcat> can I over commit prctl memory for oracle? [06:25:06] *** sleepcat has quit IRC [06:25:43] <iron_angel> I'm looking to MacGyver a home storage server out of a Blade 2000, with a combo of external USB, 1394a, 1394b and SCSI disks, with ZFS. Should this work? Most ZFS stuff I've seen is geared toward enterprise-type setups. [06:25:59] <iron_angel> I'll be sharing over the network with nfsv3/4 and smb, but that I've done before. [06:26:28] *** mritr has quit IRC [06:27:00] <jamesd_> should work, but you would be better off getting a supported 8x sata controller and plugging in 500GB drives since they are the sweet spot in storage. [06:28:52] <Triskelios> iron_angel: I use USB enclosures and an Ultra-5 myself... [06:29:36] <jamesd_> Triskelios, not exactly setting IO speed records... [06:29:51] <Triskelios> jamesd_: it's over 100mbit, so who cares? [06:30:32] <jamesd_> Triskelios, yeah i guess that is the limiting part... [06:31:04] <jamesd_> Triskelios, though QFE's and managed 100mbit switches are dirt cheap these days. [06:32:38] <iron_angel> actually it's over 1Gbit. [06:32:51] <iron_angel> found a Broadcom PCI64 NIC. [06:33:14] <iron_angel> also, the B2k has a bit of an issue with places to put said internal drives. [06:33:19] <jamesd_> iron_angel, i know the blade 1k are gigabit, the u5/10 on the other hand really can't handle gigabit plus disk io at the same time. [06:33:26] <iron_angel> ah, good point. [06:34:33] <Triskelios> the funny thing is that USB is an upgrade from DEC-era StorageWorks array that *did* manage to be an I/O bottleneck [06:37:03] <jamesd_> yeah modern drives have large caches that help alot compared to the old small scsi drives... [06:37:18] <iron_angel> hmm, actually, I suppose I could use an eSATA/external SAS controller. And I wonder if they make 146+GB FC-AL drives? [06:37:43] <jamesd_> yes they make 146GB fc-al drives but they cost an arm and a leg. [06:38:45] <jamesd_> better off getting a bunch of external boxes and using them for storage, 411's would do just cut a whole for the sata cables. [06:39:31] *** szt has quit IRC [06:42:16] <jamesd_> $200 for 2x 500GB drive + $150 for a 8x port sata controller, and $40 for a 411 drive, box gives you 500GB of mirrored storage for about what a single 146GB fc-al drive will cost you used on ebay. [06:43:18] <iron_angel> yowza. [06:43:37] <iron_angel> plus add in a few more USB and 1394 ones that I already have... [06:43:44] <delewis> 146GB disks aren't *that* expensive on eBay. [06:43:51] <delewis> you can get one nowadays for between $200-$300. [06:44:04] <jamesd_> there are still 6x sata ports free for more storage if you want. [06:44:08] <flyingparchment> fujitsu make 500GB 7,200rpm FC-AL drives, but i'm not sure they sell them separately [06:44:09] <dlg> sata is cheaper [06:44:27] <delewis> dlg: SATA is cheaper for a reason. It's a simpler technology that doesn't perform, as well. [06:44:40] <iron_angel> actually, for many things, it performs OK. [06:44:49] <iron_angel> For enterprise stuff, yeah, you want FC or SAS [06:44:52] <delewis> with SATA you don't get command queueing. [06:45:00] <iron_angel> NCQ anyone? [06:45:03] <dlg> yes you do [06:45:06] <dlg> not much of it [06:45:09] <delewis> NCQ isn't widespread. [06:45:14] <iron_angel> I'll grant, implementation is spotty, but it does exist. [06:45:16] <dlg> and you cant interleave io with other commands [06:45:36] <jamesd_> for a small fileserver sata will do fine, unless you are planning to put a TB database on it... [06:45:39] * dlg know very well what sata can and cant do [06:45:58] <iron_angel> but for what amounts to a large NFS holding-tank, it's probably OK. [06:46:09] <jamesd_> sata will do much better than usb and firewire external drives. [06:46:26] <delewis> the size of the data set has nothing to do with it. [06:46:39] <delewis> the number of iops you're throwing at the disk does. [06:46:44] <delewis> and that's independent of size. [06:46:54] <delewis> you could have a relatively small data set, but doing lots of random IO. [06:47:03] <delewis> such a workload would benefit hugely from command queueing. [06:48:08] <delewis> if you're doing largely linear IO, then you won't see much of a performance boost from command queueing. [06:48:56] <dlg> uhuh [06:49:36] <jamesd_> so too sum it up, if you are doing database or sendmail serving... get scsi/fc-al storage, for cheap home based fileserver that is mostly streaming mp3 and porn get sata.... and buy a large tv/monitor with the savings. [06:51:32] <delewis> jamesd_: provided you're using a relatively decent filesystem (like ZFS), which preserves data locality to some degree, such that what should be linear IO doesn't become random IO, then yes. [06:52:47] <jamesd_> yes ZFS was a given... [06:53:29] <iron_angel> I suppose UFS+old-style software RAID would work too, but why? [06:54:01] <delewis> with UFS, you risk having data splattered all over the disk, especially with directories that have many, many files. [06:54:15] <delewis> so what might be a linear IO, could be random IO. [06:55:31] <delewis> (UFS tries to allocate an inode in the same cylinder group as it's directory) [06:55:56] <iron_angel> ohhhh... that could be troublesome. [06:56:15] <delewis> yes, especially if you have a directory with lots and lots of files, like I said. [07:02:38] *** wnorrix_ has quit IRC [07:05:38] <kjetilho> dlg: looks like your postinstall script doesn't handle being relocated? [07:06:24] <dlg> i'll take patches [07:13:02] <kjetilho> something like http://kjetilho.at.ifi.uio.no/tmp/ITIGmfi.patch I think (but it's untested) [07:17:16] <dlg> jmcp: does that look right? [07:17:39] <jmcp> m/e checks [07:18:05] <jmcp> yeah, it does [07:18:05] <dlg> i dont think the : is needed [07:18:21] <jmcp> just below the case $BASEDIR? [07:18:25] <dlg> yes [07:19:07] <jmcp> um [07:19:08] * jmcp cogitates [07:19:13] <kjetilho> doh, to patch it in place failed, since the chksum got wrong [07:22:59] <jmcp> dlg: you're correct - don't need the ":" [07:23:13] * dlg smrt [07:23:50] * kjetilho still writes shell scripts which work on Ultrix :-) [07:27:06] <iron_angel> hehehe [07:27:52] *** sfire||mouse has quit IRC [07:28:46] <dlg> why does this have to be relocatable anyway? [07:29:07] <flyingparchment> dlg: probably to install into the miniroot [07:29:11] <flyingparchment> (assuming this is an hba driver) [07:29:40] *** Kush- has quit IRC [07:29:40] *** jmcp has quit IRC [07:29:46] <dlg> does the uninstall side need tweaks too then? [07:29:51] *** Kush- has joined #opensolaris [07:30:00] *** jmcp has joined #opensolaris [07:30:13] <kjetilho> dlg: so I can install it using pkgadd during postinstall [07:30:24] * dlg nod [07:30:27] * dlg a solaris newb [07:30:35] <jmcp> you reckon? [07:30:36] <kjetilho> I installed it into the miniroot manually [07:30:37] * jmcp doesn't agree [07:31:18] <kjetilho> dlg: ok, now on my server/OS release it bombs with ("pci1000,50") already in use as a driver or alias. [07:31:32] <jmcp> kjetilho: no doubt it's there for mpt already [07:31:45] <jmcp> kjetilho: are you sure that's the correct driver alias you should be using? [07:31:56] <jmcp> gah [07:31:56] <dlg> mpt and mfi should sit on completely different chips [07:31:58] * jmcp fails to read, again [07:32:11] <flyingparchment> pci1000,50 is an mpt controller [07:32:19] <jmcp> $ grep mpt /etc/driver_aliases [07:32:20] <jmcp> mpt "pci1000,30" [07:32:20] <jmcp> mpt "pci1000,50" [07:32:20] <jmcp> mpt "pci1000,54" [07:32:21] <jmcp> mpt "pci1000,56" [07:32:22] <jmcp> mpt "pci1000,58" [07:32:24] <jmcp> mpt "pciex1000,56" [07:32:27] <jmcp> mpt "pciex1000,58" [07:32:28] <jmcp> that's from snv_62, even [07:33:01] <dlg> hrm [07:33:08] <dlg> add_drv $ADD_DRV_BASE -m '* 0666 root sys' \ [07:33:09] <dlg> -i '"pci1000,50" "pci1028,15" "pci1000,54"' \ [07:33:09] <dlg> -c scsi mfi || exit $err_code [07:33:32] <jmcp> I'd just have the pci1028,15 [07:33:37] <flyingparchment> dlg: 1000,54 is mpt too [07:33:46] <jmcp> yeah, it's the LSISAS1064 [07:34:09] <jmcp> pci-x version [07:34:15] <dlg> grumble grumble [07:34:20] <flyingparchment> hm, it's listed as LSI Logic / Symbios Logic SAS1068 PCI-X Fusion-MPT SAS here [07:34:39] <dlg> people should give me mfis so i can fix these thiungs [07:34:43] <jmcp> dlg: I think you really want the subvendor/subdevice ids [07:34:46] <flyingparchment> but this is dell's weird modified version, who knows ;) [07:34:50] <Teltariat> I have a dumb question: how tied in is /usr to a particular nevada build? If I have /zfsroot/snv_69, and then a /zfsroot/snv_73, could they both use the same data from a /zfsroot/usr that is based on the /usr form snv_69? [07:34:51] * jmcp agrees with dlg [07:35:02] <jmcp> flyingparchment: depends on how much churn there has been [07:35:15] <jmcp> generally *not* recommended to mix+match [07:35:30] <Teltariat> Oh, ok [07:35:37] <Teltariat> I figured [07:35:43] <dlg> jmcp: should be enough to match on the actual product/vendor id [07:35:55] <jmcp> dlg: normally, yes [07:36:08] <jmcp> but if that conflicts with an existing driver alias, you need to match on the sub.... instead [07:36:47] * jmcp re-caffeinates [07:37:59] <flyingparchment> i wish x86 had more named aliases [07:38:10] <flyingparchment> like LSI,SAS1068 [07:38:21] <dlg> openbsd differentiates between mfi and mpt based on product ids just fine [07:38:52] <dlg> and man, caffeine sounds great [07:38:59] <delewis> speed. [07:39:07] <Triskelios> flyingparchment: those are reported by the devices themselves via OBP/Open Firmware... [07:39:08] <delewis> caffeine doesn't compare. :-) [07:40:31] <dlg> its enough [07:40:49] <delewis> for you. [07:41:07] * delewis is now immune to the effects of caffeine [07:41:28] <delewis> I can drink a pot of the strongest, black coffee and it does nothing for me hardly. [07:48:42] <bda> delewis: Ever had a Mr. Browns? [07:48:55] <delewis> bda: nope. That one of those super-caffeinated drinks? [07:49:17] <delewis> I've had Bawls before and that does something for me, slightly. [07:49:31] <bda> http://www.8bitjoystick.com/archives/jake_review_mrbrown_iced_coffee.php # You could say that. [07:50:17] <delewis> hmm, I'll have to order some. [07:50:19] *** theRealBallchalk has quit IRC [07:50:20] <bda> A few years ago the local crew was addicted to the stuff. [07:50:42] * delewis will be pulling 8-hour shifts and going to school full-time starting today. [07:50:45] <bda> My roommate drank so much it exacerbated a problem with his heart, it started misfiring and he was coughing up blood [07:50:48] <delewis> for the next 2 months, anyway. [07:51:07] <delewis> so any bit of caffeination I can get my hands on to carry me through the next 2 months or so is welcomed. :-) [07:51:24] <bda> I claim no responsibility if that shit kills you. :P [07:51:40] <jmcp> bda: s/claim/admit/ [07:52:18] <bda> jmcp: Yo, havin' a couple teardrops ain't no thing in this 'hood. [07:52:19] <bda> Or.. something. [07:52:25] <jmcp> heh :) [07:52:42] <jmcp> bda: yoh! yo' tok the tok, but can you wok the wok? [07:52:45] <jmcp> etc etc [07:52:54] * jmcp tries modload gangsta [07:53:02] * bda panics jmcp. [07:53:03] <jmcp> driver failed to attach [07:53:04] <e^ipi> bda: my boss isn't allowed to drink coffee anymore on account of heart murmurs [07:53:13] <e^ipi> not much, anyways [07:53:14] <bda> e^ipi: Yeah, I yell at my roommate whenever I see him sneaking some. [07:53:40] <delewis> for anybody that's worked in a corporation before -- you typically sign an agreement stating any works you create during your employement there (and perhaps a year afterwards) is owned by the employer, right? [07:53:47] <jmcp> my father-in-law used to have a problem like that, then he underwent experimental surgery to ablate a nerve on his heart muscle, he's been fine for 10+ years now [07:53:52] <jmcp> delewis: yes [07:53:56] <bda> delewis: Yea. [07:54:05] <e^ipi> bda: your roommate doesn't manage a chain coffeeshop [07:54:07] <delewis> OK, so that's fairly typical, right? [07:54:08] <kjetilho> yes, success! [07:54:18] * delewis has just signed one with his employer [07:54:19] <bda> e^ipi: No, I got the irony there. :) [07:54:25] <jmcp> ... with the caveat that if it's done at the direction of $employer, or on $employer's systems, or using knowledge gained while working for $employer, then it's theirs [07:54:26] <kjetilho> dlg: thank you very much :-) [07:54:31] <delewis> so I take it any code I wish to contribute now is off-limits. [07:54:42] <jmcp> delewis: be very careful about when you write it, and what knowledge you use to do so [07:55:18] <delewis> jmcp: OK, I haven't been hired as a programmer or developer, but I still had to sign one. I very much doubt I'll be working on projects where I'll be exposed to internal code. [07:55:28] *** neoxed has quit IRC [07:55:31] <e^ipi> delewis: my brother works @ EA, his agreement is noncompete ( not allowed to work for a competing games company ) and nondisclosure, but he's allowed to contribute to non-games related opensource [07:55:39] <bda> What dinks me up more than that are non-competes. Just evil. [07:55:44] <e^ipi> he doesn't [07:55:49] <e^ipi> but he could [07:55:52] <delewis> my agreement seems to encompass all works. [07:56:00] * delewis shrugs [07:56:01] <jmcp> delewis: you should make sure of that [07:56:20] <delewis> jmcp: should I contribute code at some point, I'll definitely ask about it. [07:56:30] <jmcp> ok [07:57:54] *** chrisso__ has quit IRC [07:59:14] *** Murmuria has joined #opensolaris [07:59:26] *** Murmuria has left #opensolaris [08:00:29] <delewis> Coy would probably be another person to ask about this, given he's in a similar position -- working for a corporation as an admin/operator, but not a developer, and exposed to very little internal code which won't be relevant towards anything that's being contributed. [08:00:57] <delewis> unfortunately, he's not online at the moment. :-( [08:01:09] *** sahafeez has joined #opensolaris [08:02:00] *** sahafeez has quit IRC [08:02:16] *** sahafeez has joined #opensolaris [08:02:24] <delewis> it seems the case, though, if you want to make decent money, nowadays, and play with cool gear, you have to sign your life over in the process. :-) [08:05:46] <bda> http://www.funnyordie.com/videos/be4da693f0 # sigh, Mitch. [08:06:17] *** Drone has joined #opensolaris [08:08:30] *** estibi has quit IRC [08:15:53] *** agony__ has joined #opensolaris [08:18:18] *** Teltariat has quit IRC [08:22:12] *** palowoda has joined #opensolaris [08:26:35] *** sahafeez has quit IRC [08:44:11] *** zassyu has quit IRC [08:44:24] *** Triskelios has quit IRC [08:47:16] *** kUdtiHaEX has quit IRC [08:47:17] *** kUdtiHaEX_ has joined #opensolaris [08:51:18] *** pablomh has joined #OpenSolaris [08:52:22] <Bartman007> yay, http://www.macrumors.com/2007/10/04/apple-seeds-zfs-read-write-developer-preview-1-1-for-leopard/ [08:52:45] <Bartman007> so my dreams of ZFS on OS X haven't gone down the tubes yet. [08:55:01] <delewis> nope, Apple knows they're going to have to get rid of HFS+ soon. [08:55:10] <delewis> especially, if they're wanting to pushing their storage products. [08:56:07] <delewis> if Apple were to bundle ZFS with some of their storage products, Thumper might see some competition. [08:56:20] * dlg look unconvinced [08:56:29] <Bartman007> had to monkey around with rsync patches to get resource forks backed up right now [08:56:30] <delewis> I certainly wouldn't be interested in it, but I'm sure existing Apple customers would. [08:56:47] <iron_angel> resource forks are a bad implementation of a good idea. [08:56:58] <dlg> apple customers are happy if they get stickers in the box [08:57:05] <Bartman007> iron_angel: yes, but a fact of life for mac users. [08:57:31] <iron_angel> this is true. [08:57:37] <iron_angel> less so in OS X, but yes. [08:57:57] <delewis> dlg: one of the primary markets for Thumper is video, though. [08:58:15] *** jonkri has joined #opensolaris [08:59:26] <delewis> I don't know what iSCSI target/initiator support is like in OS X now, but I could imagine some bundled solution that included Final Cut Pro and a Thumper-like device. [08:59:46] <dlg> i dont think apple turns over their product line in that space fast enough to be seen as viable all the time [09:00:06] *** jonkri has quit IRC [09:00:49] *** phimic has joined #opensolaris [09:01:35] <delewis> dlg: dunno. If Apple put out something like that, though, they could present an alternate, inexpensive solution to fabric for video storage. [09:02:03] <dlg> id be interested to see if they sell a sas hba and soemthing like the stk 2500 or dell md1000 stuff [09:02:22] <Bartman007> delewis: Apple has their x-raid product, doubt they would consider thumper. [09:03:13] <delewis> Bartman007: Xraid is more expensive, though, and doesn't necessarily present the benefits that a ZFS+iSCSI solution would have. [09:03:28] <bda> tank/backups 24.5K 457G 24.5K /export/backups [09:03:29] <bda> USB1. [09:03:30] * bda cries. [09:03:38] <delewis> bda: ouch. [09:03:41] <Bartman007> delewis: I know, but Apple is quite predictable in regards to market seperation. [09:03:46] <flyingparchment> bda: neat! you'll never run out of disk space, because the universe will end before you can fill it [09:03:46] <dlg> id avoid iscsi if i really cared io [09:03:53] <bda> flyingparchment: hahaha [09:04:07] <delewis> dlg: iSCSI done correctly isn't so bad. [09:04:10] <delewis> it's matured a lot. [09:04:22] <dlg> its not so great either [09:04:33] <flyingparchment> i don't see any benefit to 'iscsi done correctly' over fc [09:04:37] <delewis> it's still cheaper than fibre, which is the only alternative for shared storage. [09:05:19] <delewis> flyingparchment: compare the price of a Thumper to something EMC- or Eugenio-based. [09:05:32] <dlg> engenio? [09:05:38] <flyingparchment> delewis: you can get cheap fc arrays too [09:05:46] <delewis> the price you're saving on storage, alone, can be put towards increasing the quality of the interconnect and HBAs. [09:05:49] <flyingparchment> delewis: but you still need a proper infrastructure to do iscsi properly [09:05:54] <flyingparchment> (unless it's just for an edge system..) [09:06:11] *** trede has joined #opensolaris [09:06:17] <delewis> flyingparchment: it'll still be much, much cheaper. Fabric isn't cheap. [09:06:35] <delewis> dlg: that's the vendor StorageTek and IBM use for most of their low-end and mid-range storage products. [09:06:38] <LeftWing> HBAs, GBICs, switches, etc, aren't cheap either -- it's not just the arrays. [09:06:51] <delewis> LeftWing: exactly. [09:07:01] <flyingparchment> LeftWing: you need hbas, gbics and switches for iscsi too. other than hbas, cheap fc and ethernet stuff is about the same [09:07:04] <dlg> delewis: i know, its part of lsi [09:07:04] <Tempt> Laying and terminating fibre isn't free either [09:07:05] <delewis> with iSCSI you can use you're existing GigE switches, and purchase some decent iSCSI HBAs. [09:07:14] <dlg> everyone who does a sas enclosure uses it [09:07:15] <delewis> and you'll still save a ton of money. [09:07:28] <flyingparchment> delewis: running iscsi over your network infrastructure doesn't sound too great to me [09:07:51] <delewis> flyingparchment: you can seperate it out, but you can still use you're existing switch hardware. [09:08:01] <delewis> with fabric, you're either looking at Qlogic, Cisco, or Brocade. [09:08:21] <delewis> a 16-port fabric switch is $3-$4k minimum. [09:08:24] <LeftWing> Whereas a lot more people can do Ethernet properly. [09:08:36] *** Cyrille has joined #opensolaris [09:09:51] <delewis> and with proper iSCSI you still get all the benefits of fabric. [09:10:03] <delewis> multi-pathing, open standards, etc. [09:10:11] <delewis> and it's certainly a lot easier to debug. [09:11:09] *** tsoome has quit IRC [09:11:25] <delewis> anyone that's tried to diagnose fabric before knows it isn't easy, and usually it's completely dependent on the diagnostics your switch and HBAs provide. [09:11:40] <iron_angel> plus, what about iSCSI over 10GBE? [09:11:43] *** cydork has joined #opensolaris [09:13:09] <bda> This USB chassis is like a goddamn Myst puzzle. [09:13:16] <delewis> iron_angel: Thumper could certainly benefit from that. [09:13:21] <Bartman007> bda: written in Hypercard? [09:13:28] <bda> Bartman007: Oh, burn. [09:13:29] <delewis> and right now, fabric tops out at 4Gb/s. [09:13:39] <delewis> and 4Gb/s HBAs are *not* cheap. [09:13:45] <delewis> 10GbE is pretty reasonable. [09:14:15] <Bartman007> bda: I couldn't resist :-) [09:14:28] <bda> :P [09:16:03] <trochej> 'lo [09:16:05] <trochej> Coffee? [09:18:39] <e^ipi> hmm? [09:18:56] * bda was waiting for that. [09:23:22] *** sgnut has joined #opensolaris [09:25:52] * quasi is drinking herbal tea instead [09:28:11] <sgnut> the probe scsi_all detects my asic fastor2 but the inquire -cl (legato's command) doesn't find it. How can I know where is the problem? [09:28:12] *** nostoi has joined #opensolaris [09:29:02] <sgnut> is there a command to see all the connected scsi devices? [09:30:55] <e^ipi> every once in a while I wonder if I should try to break my caffeine addiction before coming to the conclusion that that's a stupid idea and I'll have no part in it [09:31:39] <sgnut> e^ipi: computer science engineers are machines that transform caffeine into lines of code :P [09:33:09] <e^ipi> yes, but it's lost it's effectiveness [09:33:28] <palowoda> Move to the hard stuff. [09:33:35] <e^ipi> I can drink massive amounts of coffee ( I work at a place where I press a button and espresso comes out ) and still feel nothing [09:33:41] <e^ipi> palowoda: meth? [09:33:53] <sgnut> e^ipi: and is it free? [09:33:54] <palowoda> Well it's a start. [09:33:59] <e^ipi> sgnut: for me anyways [09:34:06] <e^ipi> customers gotta pay [09:34:26] [09:34:31] <quasi> e^ipi: starbucks? [09:34:40] <e^ipi> quasi: yes [09:35:16] <sgnut> anyone of you knows a command to see all the scsi devices attached to my system? [09:35:23] <flyingparchment> sgnut: cfgadm -al [09:35:35] <sgnut> thanx flyingparchment [09:35:38] <quasi> sgnut: coffee should be free at work for all people in IT - it doesn't seem fair to charge people for going to work ;) [09:36:08] <e^ipi> IT places should invest in a commercial super-auto espresso machine [09:36:29] <e^ipi> relabel the "double" button to "make codemonkey go now" [09:36:57] <sgnut> quasi: at university one of my teachers said one time that the coffee machine place is one of the most importat place of an entreprise, since there all the people discuse about the problems of it [09:37:29] <e^ipi> like the campus pub for students [09:37:32] <quasi> sgnut: veru true [09:37:35] <quasi> very [09:38:02] <sgnut> it's a way to detect problems related to the projects [09:39:44] <g4lt-mordant> no , just a good old fashioneed Bunn 50 cup. it should be considered creul and inhuman punishment for a person to be expected to work without coffee [09:40:52] <sgnut> flyingparchment: cfgadm -al doesn't show me the asic fastor 2, but probe_scsi all at boot prompt detects it... any idea where is the problem? [09:41:01] <flyingparchment> sgnut: try `devfsadm` [09:42:30] *** tsoome has joined #opensolaris [09:43:39] *** peteh has joined #opensolaris [09:43:40] <sgnut> flyingparchment: yeeeeeeeeeeeeeeees it works now!!!! thanx a lot!!! do you know why it failed? [09:44:08] <flyingparchment> sgnut: solaris doesn't scan for new devices unless you boot -r, or run devfsadm [09:44:32] <sgnut> flyingparchment: oh, I didn't know that... [09:44:48] <sgnut> flyingparchment: thank you very much!!!!!!!! [09:52:16] *** Dar_HOME is now known as Dar [09:57:55] *** TpentaNB has joined #opensolaris [10:00:04] *** dmarker has quit IRC [10:00:15] *** dmarker has joined #opensolaris [10:01:43] *** Gman has joined #opensolaris [10:01:49] *** TpentaNB has quit IRC [10:05:14] *** KermitTheFragger has joined #opensolaris [10:13:17] *** bengtf has quit IRC [10:15:24] *** Erwann has joined #opensolaris [10:17:49] *** agony_ has joined #opensolaris [10:18:05] *** Gekz[sleep] is now known as Gekz [10:18:21] *** andyshack has joined #opensolaris [10:23:47] <andyshack> Could someone suggest a good guide to setting up a mirrored root drive ? there seem to be a few out there that are a tad different and im unsure which way to go. [10:25:15] <flyingparchment> andyshack: the one in the svm manual at docs.sun.com [10:25:40] <andyshack> Yeah, that sounds like a good idea. [10:28:01] *** agony__ has quit IRC [10:28:34] *** mega has joined #opensolaris [10:29:13] *** Murmuria has joined #opensolaris [10:29:47] *** Murmuria has left #opensolaris [10:36:02] *** bengtf has joined #opensolaris [10:37:15] *** razrX_ is now known as razrX [10:37:20] <kjetilho> I'm having a weird problem with SSH in a zone. if I ssh into the global zone, then ssh to the zone, I can no longer ssh back into the global zone. I should get in with my key, but am prompted for my password, and it fails. [10:37:50] <kjetilho> the global zone and the local zone have proper IP addresses [10:39:29] <WickedWicky> good morning [10:40:26] <trochej> Mornin' [10:40:28] <trochej> Coffee? [10:40:43] <WickedWicky> yes pleeas! [10:41:05] <WickedWicky> wow, what a kick ass weekend, I saw some great bands :D [10:44:18] <razrX> where did ya go WickedWicky ? [10:44:33] <razrX> my weekend sucked but let's not focus on that [10:55:11] <WickedWicky> I went to ProgPower EU [10:57:56] *** MattMan has joined #opensolaris [10:58:37] *** dnilsson has joined #opensolaris [10:59:20] *** alobbs has joined #opensolaris [11:01:39] *** Mdx4 has joined #opensolaris [11:04:00] *** mega has quit IRC [11:05:46] *** GiacoX has joined #opensolaris [11:06:05] <GiacoX> hi [11:06:26] <GiacoX> where can I find the svc and manifest for squid from coolstack ? [11:10:26] <WickedWicky> I'd presume where all manifests go, /var/svc [11:11:48] *** bunker has joined #opensolaris [11:11:58] <phips> GiacoX: look at the package contents, or grep in /var/sadm/install/contents? [11:12:31] *** Snake007uk has joined #opensolaris [11:12:53] <GiacoX> I used the src pkg [11:13:14] <GiacoX> I think that I have to installe the binary pkg and after recompile the srcs over it [11:14:12] <WickedWicky> can't you do a 'find /var/svc *squid*' ? [11:15:08] <GiacoX> Proxy1 # find /var/svc/manifest/ -name *squid* [11:15:08] <GiacoX> Proxy1 # [11:15:18] <GiacoX> I think that only the binary pkg install it [11:15:55] <GiacoX> I installed CSKsquidSrc [11:16:35] <WickedWicky> maybe there is some directory in the sourcedir that has the manifests [11:16:39] *** timsf has joined #opensolaris [11:16:48] <WickedWicky> Inever used coolstack or source package [11:17:12] <timsf> morning all [11:17:21] <WickedWicky> good morning [11:19:14] *** noyb has joined #opensolaris [11:23:05] *** doof has left #opensolaris [11:23:27] *** seanmcg has joined #opensolaris [11:27:15] <quasi> hmm, http://www.gekkou.co.uk/fsrewrite/ - sounds kind of interesting for building packages [11:34:24] *** halton has left #opensolaris [11:34:34] *** noobuntu has quit IRC [11:37:59] *** jonkri has joined #opensolaris [11:40:49] <trygvis> quasi: sounds like something similiar I did with lofs and chroot [11:40:53] <trygvis> (for building packages) [11:48:24] <FastJack> hmm, I created a raidz2 with a spare. now I tool one of the disks offline. zpool status says it is degraded now but the spare did not jump in. is that a bug or a feature? :) [11:51:09] <timsf> I think they should only jump in when a device breaks FastJack [11:51:50] <timsf> You can simulate it with zpool replace AFAIR [11:53:04] <FastJack> I wanted to simulate a broken hard disk to see how that would look like [11:54:04] <timsf> Dropping one from a great height might do the trick ;-) [11:54:27] <timsf> but yeah, hang on - I'll mess about with a test system here... [11:54:35] <FastJack> I said _simulate_ :) [11:55:05] <FastJack> I can manually replace them. but that's not my defintion of a hot spare [12:00:36] <timsf> :-/ you don't have zinject handy, do you - that should simulate io errors [12:01:18] <timsf> http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/zinject/ [12:03:08] <FastJack> ok, I'll have to check that out when I get back home. I don't have a compiler installed on that machine yet [12:03:49] *** Mdx4 has quit IRC [12:03:50] *** sgnut has quit IRC [12:07:53] *** NikolaVeber has joined #opensolaris [12:07:54] <noyb> timsf: are you at cec? [12:08:38] <timsf> Nope [12:09:44] <timsf> - I get to fly to the US on Wednesday, but not to Vegas. Probably a good thing :-) [12:10:56] <seanmcg> timsf: can you get to San Fran then ? Oktober fest this weekend :) [12:11:13] <timsf> dammit. [12:11:27] <timsf> Nope - will be down in Santa Cruz for the weekend [12:12:38] <timsf> Then Santa Clara on Monday & Tuesday, back home on Wednesday. [12:13:17] <noyb> bummer. quasi told me that brendang didn't have time for cec this time. [12:14:43] <timsf> Yeah, never been to one of those conferences, they sound like a bit of a blast alright [12:15:18] <noyb> it would be cool to meet some actual live people instead of typing to my computer. :-) [12:16:00] <seanmcg> noyb: you going anyway ? [12:16:02] <timsf> Yeah, there's definitely something to be said for face to face - my boss is in the US and I only get to see her once a year or so.. [12:16:17] <noyb> yes. I'm here in las vegas. [12:16:41] <noyb> waiting for 7am to show up so I can wake up and grab some breakfast... :-) [12:17:33] <noyb> we've come a long way from my first computer... (timex sinclair 1000) [12:17:36] <timsf> Good grief - what time is it there ? (and isn't this Vegas, where life goes on 24x7, and you can get breakfast at any time of day ?) [12:18:02] *** andyshack has quit IRC [12:19:31] <noyb> I'm on my new macbook with ubuntu 7.10 running in a vm, while solaris snv73 is installing, and chatting with you via xchat, and streaming audio from the internet with rhythmbox... [12:19:38] <noyb> it's only 3:20 am [12:20:14] <timsf> jet lag ? [12:20:40] <noyb> I'm only a few hours away in Lancaster, CA. [12:20:56] *** Mdx4 has joined #opensolaris [12:20:57] <timsf> Ah, caffeine then :-) [12:22:18] <noyb> naw. I just don't sleep well out of my own bed. not to mention pure uninterrupted nerd-O-rama doing lots of things with the new mac. [12:22:54] <madhatter> Do I have to look for some special features in a laser printer to make it work with opensolaris' cups? [12:31:19] <delewis> no, but if it supports Postscript out of the box, that certainly makes things easier, usually. [12:31:32] <delewis> and that's not specific to CUPS, but any UNIX print daemon. [12:32:11] <noyb> why can't they give away high speed internet instead of giving away lame prime rib dinners? $12.99 for 24 hours ? sheesh. [12:33:08] *** estibi has joined #opensolaris [12:33:15] *** jcea has joined #opensolaris [12:33:22] <madhatter> delewis: Okay, I will try to find info on that then [12:33:25] <madhatter> delewis: Thanks [12:33:53] <delewis> madhatter: what kind of printer is it? [12:35:42] *** Mdx4 has quit IRC [12:36:25] *** yongsun has quit IRC [12:37:51] *** nostoi has quit IRC [12:39:09] <madhatter> delewis: I am just planning to get one. Right now I found a Samsung ML-2010R+, but it seems that it has no Postscript support [12:39:56] <flyingparchment> madhatter: PS isn't required, PCL (HP) is common and works well too [12:40:05] <flyingparchment> almost all laser printers will do either ps or pcl [12:40:43] *** calumb has joined #opensolaris [12:41:29] <madhatter> flyingparchment: Seems that Samsung has it's own language (SCL) [12:41:51] <madhatter> flyingparchment: SPL, sorry [12:42:04] <delewis> madhatter: if you've gotten this printer to work in Linux using CUPS before, then it should work in Solaris flawlessly. [12:42:52] <madhatter> delewis: I did not, because I did not buy it yet. But Samsung gives compatibility guarantees for Linux and OSX. [12:44:14] <delewis> that could either mean they've got support in CUPS or they just let you download some proprietary binary, which has little chance of working on Solaris. [12:48:43] *** pablomh has quit IRC [12:50:48] <kjetilho> not on Solaris/SPARC, but Brandz should handle it on x86 [12:53:19] <madhatter> kjetilho: Okay, that would be a workaround then... I will try to figure out what support they really have. [12:56:57] *** noobuntu has joined #opensolaris [13:00:14] *** _devzero_ has joined #opensolaris [13:06:16] *** mega has joined #opensolaris [13:06:36] *** jonkri has quit IRC [13:17:19] *** GiacoX has quit IRC [13:29:50] *** mega has quit IRC [13:29:53] *** calumb is now known as calLNCH [13:34:41] * paul wonders what hardware would be best for a cheap, SFF, low-power (W), OSol-based filer box [13:35:07] <trochej> A Toaster [13:35:19] <Cyrille> two for high availabilty. [13:35:23] <Cyrille> availability even [13:36:14] <quasi> paul: how SFF? [13:36:15] <paul> trochej: hmm, cheap alright, but I'm sure it'd fail the 'OSol-based' and 'filer box' requirements [13:36:26] <paul> plus i reckon, in operation, it'd not be particularly low-W [13:36:45] <paul> quasi: reasonably [13:36:56] <quasi> paul: something via based would probably be the cheapest [13:37:05] <paul> quasi: something neater than a midi-tower. [13:37:14] <paul> yeah, casper seems fond of those. [13:37:20] <paul> judging by his blog [13:37:55] <quasi> paul: asus does a pile of semi sff things for reasonable prices if you want a bit more cpu in it [13:38:48] <paul> quasi: pointers to any retailers? (EU based preferably, UK ideally) [13:39:18] <paul> (i know of 'mini-itx.com') [13:39:32] <quasi> for the asus? [13:39:35] <madhatter> Anybody tried to use SpliX drivers for cups in snv yet? (splix.sf.net) [13:39:50] <quasi> linitx.com for other via bits [13:40:19] <quasi> http://www.asus.com/products.aspx?l1=1&l2=3 is some of the asus boxes [13:43:48] <paul> interesting [13:44:08] * paul wonders which osol list my question would be ontopic on. [13:44:15] <paul> (-discuss i guesS) [13:46:05] <quasi> there's a list for building nas like thingies [13:46:42] *** FelipeBare has joined #opensolaris [13:46:59] <quasi> http://opensolaris.org/os/community/appliances/ [13:47:56] <paul> quasi: ah indeed, danke :) [13:48:22] *** _devzero_ has quit IRC [13:49:24] *** mcnamarabrian has joined #opensolaris [13:50:21] *** Kush- has quit IRC [13:56:51] *** mcnamarabrian has quit IRC [14:09:10] *** mcnamarabrian has joined #opensolaris [14:09:51] *** kaiwai has quit IRC [14:12:20] *** kUdtiHaEX__ has joined #opensolaris [14:15:13] *** kUdtiHaEX_ has quit IRC [14:17:03] *** mcnamarabrian has quit IRC [14:17:04] *** master_o1_master has joined #opensolaris [14:26:03] *** jeanBlack has joined #opensolaris [14:28:04] *** master_of_master has quit IRC [14:31:11] *** laca has joined #opensolaris [14:37:33] *** calLNCH has quit IRC [14:46:23] *** Murmuria has joined #opensolaris [14:46:37] *** Murmuria has left #opensolaris [14:51:04] *** masta has joined #opensolaris [14:51:05] *** jcea has quit IRC [15:00:39] *** alanc_away has quit IRC [15:02:53] *** migi has joined #opensolaris [15:04:25] *** jonkri has joined #opensolaris [15:08:35] *** migi_ has joined #opensolaris [15:10:05] *** mcnamarabrian has joined #opensolaris [15:11:03] *** mcnamarabrian has quit IRC [15:11:34] *** jcea has joined #opensolaris [15:13:41] *** nachox has joined #opensolaris [15:14:16] *** migi has quit IRC [15:18:49] *** Gekz is now known as Gekz[PDA] [15:19:28] *** dmesg has joined #opensolaris [15:19:30] *** JWheeler has quit IRC [15:30:14] *** baijiutong has quit IRC [15:30:24] *** baijiutong has joined #opensolaris [15:35:09] *** peteh_ has joined #opensolaris [15:35:10] *** jcea has quit IRC [15:39:16] *** peteh_ has quit IRC [15:39:38] *** peteh_ has joined #opensolaris [15:39:46] *** storycrafter has left #opensolaris [15:39:55] *** RaD|Tz has joined #opensolaris [15:41:46] *** Mdx4 has joined #opensolaris [15:45:43] *** migi_ has quit IRC [15:46:12] *** nicoAMG has joined #OpenSolaris [15:51:43] <nachox> morning [15:54:13] *** Mdx4 has quit IRC [15:54:41] <jteo> nachox: morn. [15:55:34] *** Yaksha has quit IRC [15:56:05] *** jonkri has quit IRC [15:57:51] *** peteh_ has quit IRC [16:01:04] *** millhouse has joined #opensolaris [16:01:29] <millhouse> morning [16:02:26] <quasi> afternoon [16:02:51] *** blindfish has joined #opensolaris [16:03:04] *** logic has quit IRC [16:03:07] *** logic_ has joined #opensolaris [16:03:12] *** zassyu has joined #opensolaris [16:04:51] *** RaD|Tz has quit IRC [16:06:57] <millhouse> is there a way i can mount an nfs share with bigger buffers (like 16MB)? [16:09:57] <kjetilho> what do you mean by buffers? [16:10:27] <Stric> nfs3+4 over tcp can do 1MB buffers at least [16:10:36] <Stric> kjetilho: rsize/wsize [16:10:43] <millhouse> ok [16:10:48] <millhouse> is there a way to increase that? [16:11:15] <Stric> I don't think there is much to gain over that [16:11:26] <millhouse> ok, i wasn't sure [16:11:31] <millhouse> i haven't played much with nfs [16:11:40] <Stric> unless you're filling a 40Gbit/s connection or so [16:12:00] <millhouse> sadly no [16:12:03] <Stric> the overhead is quite small with 1MB blocks [16:12:34] <kjetilho> max IP packet size is 64 KiB, and max Ethernet frame is 9000 (jump packets) [16:12:46] <kjetilho> s/jump/jumbo/ [16:16:12] <millhouse> ah ok [16:16:24] <jteo> true. [16:16:50] <millhouse> i've mostly used samba and i know there you can increase the size of the read/write buffers so i wondered if you could do similar with nfs [16:17:02] <millhouse> is there a good nfs client for windows i could use? [16:17:34] <trochej> millhouse: We did some test some time ago and one thing became clear: for Windows use cifs [16:17:48] <trochej> millhouse: But you can use unix services for windows (downloadable from microsoft.com) [16:18:02] *** pablomh has joined #OpenSolaris [16:19:47] *** snuff-home has quit IRC [16:21:16] <millhouse> trochej: I figured you'd say that.. Honestly for mainstream use, we'd have to use CIFS because the apps require it.. I've been curious about the Unix services from Microsoft too.. Maybe I'll try it.. I'll probably be the only person using NFS anyway [16:23:37] *** pschow has joined #opensolaris [16:24:10] *** virgee has joined #opensolaris [16:26:49] *** kszwed has joined #opensolaris [16:30:37] *** Triskelios has joined #opensolaris [16:31:44] <kszwed> 'lo Triskelios [16:32:25] <Triskelios> hey [16:34:31] *** Kush- has joined #opensolaris [16:35:46] *** dnilsson has quit IRC [16:38:05] <jteo> hmm. technically, everything is a resource issue. [16:39:02] *** pfa3rh has joined #opensolaris [16:39:05] <kszwed> i assume that includes lazyness [16:39:12] * Triskelios resists the temptation to respond with "your mom is..." [16:40:53] <jteo> laziness is a lack of a particular resource. [16:43:02] *** alanc_away has joined #opensolaris [16:43:32] <kszwed> jteo: so how come there's a trend that many obese people are lazy. there's plenty of resources there ;) [16:44:30] *** tsoome has quit IRC [16:44:57] *** Gekz[PDA] is now known as Gekz[sleep] [16:46:01] <jteo> i'll keep quiet now, else Godwin's Law will assert itself. [16:47:35] *** trede has quit IRC [16:49:55] <kszwed> jteo, heh. hail! ;0 [16:49:56] *** phimic has quit IRC [16:53:04] *** charlesnw has joined #opensolaris [16:57:58] *** nadir has quit IRC [17:00:14] *** Megaf has joined #OpenSolaris [17:00:16] *** kszwed has quit IRC [17:05:47] *** gm152_ has joined #opensolaris [17:06:02] *** wms has joined #opensolaris [17:07:07] *** ghatak has joined #opensolaris [17:10:21] <ghatak> Hi, i have a small problem, i am using a zfspool for my data, from that pool i create a snapshot which is then used to create another snapshot with some modification in data by mysql. I call this snapshot the working copy. I create clones off this for databases. every night i have to create a new working copy of snapshot, the problem is that, removing the working copy also removes any clones or snapshots created off it. I want to keep them and r [17:17:21] *** noyb has quit IRC [17:18:11] *** Teltariat has joined #opensolaris [17:19:52] <Triskelios> ghatak: your question was truncated at "I want to keep them and" [17:20:15] *** nicoAMG has quit IRC [17:20:27] *** nicoAMG has joined #OpenSolaris [17:20:27] <ghatak> Triskelios:remove the working copy every night. Is there anyway i can get around this ? [17:21:23] *** tsoome has joined #opensolaris [17:22:48] <Triskelios> ghatak: you can promote the clone to a full filesystem [17:23:18] <ghatak> Triskelios: yeps, when i do that, now the working copy is a dependency of the new full filesystem [17:23:43] <ghatak> and i can not destroy that, that is what needs to replaced each night [17:24:13] *** derchris has joined #opensolaris [17:25:24] *** baijiutong has quit IRC [17:26:26] <Triskelios> are you sure destroying the snapshot is the right decision if you're going to keep its clones? [17:27:23] <Triskelios> they're all copy-on-write so you're not using more storage than you would need to [17:31:05] *** dmesg has left #opensolaris [17:33:00] *** peteh has quit IRC [17:35:36] *** Megaf has quit IRC [17:36:32] *** sommerfeld has joined #opensolaris [17:36:32] *** ChanServ sets mode: +o sommerfeld [17:37:41] *** ghatak has left #opensolaris [17:42:02] *** noyb has joined #opensolaris [17:42:43] *** bunker has quit IRC [17:44:44] *** gm152_ has quit IRC [17:46:59] *** stevel has joined #opensolaris [17:46:59] *** ChanServ sets mode: +o stevel [17:48:02] <jteo> wb stevel. [17:48:29] <stevel> 'morning jteo [17:50:08] <jteo> new job all good? [17:50:17] <stevel> so far so good :) [17:50:17] *** slowhog has joined #opensolaris [17:50:35] *** bengtf has quit IRC [17:51:17] <jteo> smashing. [17:51:18] *** TpentaNB has joined #opensolaris [17:56:45] *** tsoome1 has joined #opensolaris [18:01:08] *** snuff-away has joined #opensolaris [18:02:38] *** Fish- has joined #opensolaris [18:04:03] *** tsoome has quit IRC [18:05:13] *** troff has quit IRC [18:06:12] <Fish-> hello [18:11:35] <Doc> hrmmm [18:13:00] <jteo> hrmmm? [18:13:29] *** bengtf has joined #opensolaris [18:14:20] *** jcea has joined #opensolaris [18:15:07] <nachox> hey stevel :) how's the new office? [18:16:24] *** tsoome has joined #opensolaris [18:17:10] *** vmlemon has joined #opensolaris [18:18:02] *** millhouse has quit IRC [18:18:04] <stevel> nachox: nice :) spent most of the week just meeting people and reading. it's tough going into a place being the idiot who knows nothing ;-) [18:18:18] *** Erwann has quit IRC [18:19:35] <jteo> it's called the honeymoon [18:19:42] *** Snake007uk has quit IRC [18:19:45] <nachox> stevel: i've been at my current job for over a year, the first few months i was "the idiot that knew nothing". after a while they got tired, it was too long to call me that, so now i'm just "the idiot" ;) [18:20:14] <stevel> lol [18:22:31] *** Cyrille has quit IRC [18:23:30] *** jmcp has quit IRC [18:23:40] *** noyb has quit IRC [18:23:51] *** jmcp has joined #opensolaris [18:24:41] *** TpentaNB has quit IRC [18:25:45] <Doc> a year and you're still the idiot? cool! [18:26:17] <Doc> stevel: how long were you at sun? [18:26:38] <stevel> doc: 4 years, 3 months [18:28:30] <Doc> so it's not too long since you were last the idiot that knew nothing on the first day :) [18:28:51] <stevel> indeed [18:28:57] <stevel> fortunately at Sun i was eased into it :) [18:29:36] <jteo> tis romantic. [18:30:01] *** laca has quit IRC [18:30:30] <Doc> bawha.. i just got an email from a microsoft recruiter asking if i'd be interested in a job [18:30:47] <nachox> you accepted of course [18:30:51] <the-decider> Doc: their new MS Solaris project [18:30:58] *** KermitTheFragger has quit IRC [18:31:03] <nachox> asked for 1m per year ;) [18:31:14] <jteo> maybe it's for email software that doesn't read your email. [18:31:37] <Doc> i've replied saying i might be interested (i'm not), and asking where he got my details [18:33:09] *** duri has quit IRC [18:33:23] *** duri has joined #opensolaris [18:33:59] *** tsoome1 has quit IRC [18:34:50] *** glagasse has joined #opensolaris [18:35:00] *** Dar is now known as Dar_HOME [18:37:33] *** agony_ has quit IRC [18:37:48] *** vmlemon has quit IRC [18:38:39] *** Jondice has quit IRC [18:38:43] *** vmlemon has joined #opensolaris [18:43:00] *** iron_angel has quit IRC [18:50:43] *** Jondice has joined #opensolaris [18:55:36] <jteo> reading internet forums might prove damaging to my mental state. [18:56:22] <Doc> just stay away from the ones with idiots on them [18:56:39] <Doc> ie, unplug your network connection [18:56:44] *** sartek has joined #opensolaris [18:57:06] <jteo> then i would miss out on your wonderful wisdom. [18:57:10] <jteo> tis a Catch-22! [18:57:11] *** slackjr has joined #opensolaris [18:57:35] <Doc> you can subscribe to my weekly snail-mail newsletter [18:57:58] *** slackjr has quit IRC [18:58:06] <jteo> true. [18:58:54] *** sartek has quit IRC [19:01:24] *** timsf has quit IRC [19:01:49] *** Mdx4 has joined #opensolaris [19:02:09] *** MattMan has quit IRC [19:06:17] *** alanc-away is now known as alanc [19:07:15] *** Mdx4 has quit IRC [19:14:49] *** Pietro_S has joined #opensolaris [19:15:38] *** gnut has joined #opensolaris [19:18:43] *** deather_ has joined #opensolaris [19:21:49] *** polk__ has quit IRC [19:28:48] *** pfa3rh has quit IRC [19:28:49] *** Murmuria has joined #opensolaris [19:30:42] *** cydork has quit IRC [19:32:40] *** Murmuria has left #opensolaris [19:33:43] *** sahafeez has joined #opensolaris [19:35:55] *** deather has quit IRC [19:36:05] *** Murmuria has joined #opensolaris [19:36:21] *** Murmuria has left #opensolaris [19:39:40] *** neoxed has joined #OpenSolaris [19:46:46] *** cga has joined #opensolaris [20:01:08] *** deather_ is now known as deather [20:04:15] *** baijiutong has joined #opensolaris [20:04:31] *** Murmuria has joined #opensolaris [20:04:39] *** Murmuria has left #opensolaris [20:07:30] <gdamore> hi * [20:09:48] *** sfire||mouse has joined #opensolaris [20:09:56] <the-decider> irc: hi: No match. [20:12:58] <moazamraja> gdamore: The M4000/M5000 is seriously needed. [20:13:09] <gdamore> what for? [20:13:20] <moazamraja> cuz like, santa cruz will be cold that weekend [20:13:25] <gdamore> heh. [20:13:27] <moazamraja> and like, i need warmth. [20:13:28] <moazamraja> :/ [20:13:49] <moazamraja> we have great 1u-2u hardware that does an excellent job, we should make use it. [20:13:52] <moazamraja> use of it...even [20:14:32] <moazamraja> mpk16 runs sun ray servers off T2000s, seems to be a lot better than the machines they used to use [20:15:34] <sommerfeld> they've gone through several generations of sunray servers. E4500 -> V880 -> T2000 [20:15:41] <gdamore> right. no point in not using that t2k. i think a couple of sun rays setup would still be useful, despite an earlier post by ben pooh-poohing the idea. [20:16:02] <gdamore> ISTRC they were using E10Ks at one point. [20:16:14] <gdamore> i know in SD they were, dunno about MPK [20:16:50] <jbk> define cold [20:16:55] <sommerfeld> that's probably because san diego had lots of e10k's [20:17:18] <alanc> we could pull out the "Sun Ray doesn't support running on OpenSolaris" card, though I think it mostly works on Nevada [20:17:35] <alanc> definitely not on Indiana though until they finish the Xorg port [20:17:58] <coffman> smb-tng smb-tng smb-tng20:13 moazamraja> :/ [20:17:58] <coffman> 20:13 moazamraja> :/ [20:19:50] *** bowdengl has joined #opensolaris [20:20:09] *** bowdengl has left #opensolaris [20:20:37] <moazamraja> ? [20:20:56] <coffman> moazamraja: eh, sorry, paste error [20:21:08] <moazamraja> ah [20:21:17] <moazamraja> and here I thought I was popular :( [20:21:57] <moazamraja> sommerfeld: the T2K sunray server felt a heck of faster than the v880 [20:22:10] <moazamraja> seat of the pants test tho [20:22:13] <the-decider> but not as warm ;) [20:24:18] <sommerfeld> based on what I saw the other day in the bur02 server room they have a lot more T2000's than they had v880's so each one is going to be less loaded. [20:25:04] <sommerfeld> after all you can fit ~8.5 T2000's in the space of a single V880 [20:25:29] <moazamraja> did v880s require 220v power? [20:25:37] <moazamraja> I know the 1280/2900s did [20:25:40] <moazamraja> what a pain that was [20:26:53] <jteo> ah yes the US doesn't use 220v. [20:27:04] <delewis> we do. [20:27:14] <delewis> it's just not standard for most consumer devices. [20:27:28] <delewis> only high-amperage devices, like appliances and so fourth. [20:27:39] <jteo> ah. [20:27:46] <jteo> the usual is 120v? [20:28:00] *** slowhog has left #opensolaris [20:28:06] <PerterB> well a v880 isn't exactly a consumer device [20:28:18] <delewis> jteo: 110v. [20:28:30] <jbk> i'm guessing the v880s can probably take either [20:28:37] <coffman> lul, yeah europe has it better on this [20:28:44] * delewis wonders since other countries are ahead of us circuit-wise, when will they move to three-phase? :-) [20:28:45] <jbk> since i seem to remember they could be used as a sort of 'desktop' server [20:28:45] <quasi> in 220V land, 380V is common to for things like electric stoves [20:28:59] <jteo> hmm. [20:29:15] <WickedWicky> yep [20:29:21] <WickedWicky> or kick ass deep friers in restaurants [20:29:25] <delewis> jbk: visualization workstation is the term I think you're looking for. :-) [20:29:34] <delewis> that is, if you purchased say, an XVR-4000. [20:29:55] <the-decider> WickedWicky: gah! now I'm hungry [20:29:56] <sommerfeld> V880 would run off of 110V. V890 needs 208/220/240V [20:30:01] * delewis remembers E4500s being marketed for similar purposes. [20:30:11] <jteo> Finger Lickin' Good! [20:30:19] *** xOmega has joined #opensolaris [20:30:30] * sommerfeld wonders how well a V890 would do as the heat source for a deep fryer [20:30:37] <jteo> not very well. [20:31:05] <quasi> WickedWicky: we had an electrician who messed up and sent 380V into our primary server (many moons ago) - the blue smoke definetely got out of that box fast [20:31:06] *** baijiutong has quit IRC [20:31:18] <the-decider> sommerfeld: I think that'd get messy. [20:32:30] <sommerfeld> yep. residential use of 220 is largely for clothes dryers, electric stoves, and high-capacity air conditioners. [20:32:41] * the-decider used it for an Onyx at home for awhile [20:32:43] <sommerfeld> (in the US, that is). [20:32:54] <the-decider> but that would fall under a high-capacity heater ;) [20:33:13] <WickedWicky> the hairstyler of your girlfriend [20:33:36] *** nachox has quit IRC [20:34:37] <jbk> delewis: E4500's or E450s? [20:34:44] <WickedWicky> quasi: I hope he was insured :p [20:34:54] <sommerfeld> former coworker of mine described a rather bizarre hack involving a 3-phase electric motor that home machine shop owners did to get 3-phase equipment running when you didn't get all 3 phases in your residential service. [20:35:09] <the-decider> sommerfeld: did we have the same friend? [20:35:38] <the-decider> I knew someone who had one of those in his basement to convert his juice to 3phase to power an Onyx rack [20:35:46] <quasi> WickedWicky: it was one of those frightfully expensive ibm junk boxes where they had to fly in a tech with the motherboard chained to his arm [20:36:05] <WickedWicky> *shrug* [20:36:06] *** mritun_ has joined #opensolaris [20:36:21] <mritun_> Hi folks [20:36:41] <sommerfeld> the-decider: he claimed it was a well-known hack in certain circles. [20:36:51] <the-decider> basically, you power a 3-phase motor off one phase, and you'll end up with 3 phase power pulling off 3 taps. [20:37:05] *** virgee has joined #opensolaris [20:37:12] <the-decider> noisy though -- having that motor chugging around doing nothing. [20:37:55] <sommerfeld> well, he was getting set up to power surplus machine shop equipment (lathes & drill press & such for metalworking) [20:37:58] <mritun_> hmm.. I was wondering if suspend/resume ever got integrated ? [20:38:09] <the-decider> so noise wasn't really an issue ;) [20:38:13] <sommerfeld> Right. [20:39:08] <sommerfeld> mritun_: sparc suspend/resume integrated a really long time ago. x86 suspend-to-memory isn't done yet [20:39:37] <mritun_> sommerfeld: aha... thanks [20:39:57] <mritun_> I guess not all drivers are compliant yet [20:41:20] *** xOmega has left #opensolaris [20:41:21] <sommerfeld> mritun_: and they won't be even after it integrates, but the problem becomes a lot more manageable once you have the needed framework working on both platforms. [20:41:52] <mritun_> oh... [20:42:09] <mritun_> so that means it won't be available anytime soon, right ? [20:42:21] <sommerfeld> fixing an individual driver is a much more manageable task than fixing the entire suite of drivers. [20:42:40] <mritun_> yeps [20:42:52] <jteo> lack of resources? [20:43:25] <mritun_> um, I'd say even with enough resources it'd take a long time... [20:43:54] <sommerfeld> IIRC the initial target is "devices found on an ultra 20" [20:44:00] <mritun_> some HW just doesn't behave right after waing up... so it'd be lot of trial & error too [20:44:40] <sommerfeld> it will shake out those drivers that don't fully initialize the chips they drive. [20:45:23] <mritun_> hmm [20:45:32] <mritun_> and is someone working on bluetooth ? [20:46:00] <mritun_> there was a post sometime back... some student had a working driver done as course project [20:49:30] <mritun_> I'm tempted to do one [20:49:50] <mritun_> but the streams concept takes time to wrap my head around :( [20:52:07] <jbk> seems like streams are becoming more and more of a legacy/compatability bit [20:53:16] <sommerfeld> jbk: only for networking. [20:53:39] <jbk> where else are they used? [20:54:49] <sommerfeld> they started off in the terminal subsystem (real and pseudo ttys) and are still used there. [20:54:56] <sommerfeld> USB makes heavy use of streams as well. [20:55:39] <jbk> oh yeah... dunno why i forgot about ttys [20:57:18] <sommerfeld> using streams between tcp and ip wasn't done right to begin with (too much code duplication); a big part of "fire engine" was welding the two modules together. [20:58:55] <masta> wasn't fire engine the name of express at some point? [20:59:59] <sommerfeld> no. [21:00:08] <Triskelios> FireEngine is the name of the GLDv3 stuff, no? [21:01:08] <flyingparchment> no, fireengine is the rewritten ip stack [21:01:21] <flyingparchment> (of which i think gldv3 is a part..) [21:06:18] *** Shiv_1 has joined #opensolaris [21:08:16] *** Jondice has quit IRC [21:11:53] *** Shiv_1 has quit IRC [21:14:16] *** kUdtiHaEX__ is now known as kUdtiHaEX [21:17:47] *** Jondice has joined #opensolaris [21:21:25] *** doof has joined #opensolaris [21:21:28] <doof> hi [21:25:54] *** iron_angel has joined #opensolaris [21:29:36] *** virgee has quit IRC [21:41:31] * iron_angel kicks netbooting. [21:45:35] <doof> how can i test mouse under console mode ? [21:48:07] <yofuh> cat /dev/mouse [21:48:54] *** cga has quit IRC [21:48:59] * iron_angel debates whether to use Solaris 10, or SXCE on ths box... [21:53:02] *** agony_ has joined #opensolaris [21:53:29] *** wms has quit IRC [21:58:34] *** mcnamarabrian has joined #opensolaris [21:59:50] *** micken has joined #opensolaris [22:00:10] *** cypromis has joined #opensolaris [22:01:08] *** masta has left #opensolaris [22:02:21] *** Gman has quit IRC [22:06:17] *** wnorrix has joined #opensolaris [22:10:52] *** blindfish has quit IRC [22:15:15] *** sarahj has joined #opensolaris [22:20:54] *** mritun_ has left #opensolaris [22:21:23] *** Fish- has quit IRC [22:22:59] *** neoxed has quit IRC [22:23:58] *** micken has quit IRC [22:28:13] *** agony_ has quit IRC [22:29:51] *** agony_ has joined #opensolaris [22:31:08] *** JWheeler has joined #opensolaris [22:34:16] *** mcnamarabrian has quit IRC [22:45:11] *** jeanBlack has quit IRC [22:54:14] *** nostoi has joined #opensolaris [22:56:24] *** mikefut has joined #opensolaris [22:57:25] *** FelipeBare has quit IRC [22:59:47] *** ottom has joined #opensolaris [23:00:29] *** sarahj has quit IRC [23:04:10] *** mikefut has quit IRC [23:05:32] *** pina has joined #opensolaris [23:12:10] <yofuh> hmmm, frank ludolphs mail crashes my evolution... [23:21:55] <BadKarma> have you EVER considered seeing a shrink to talk about it? [23:22:21] <iron_angel> M-x psychoanalyze-pinhead [23:25:11] <BadKarma> well here comes the prozac [23:26:30] <holcomb> zippy is my hero [23:26:54] *** jcsmith has quit IRC [23:27:02] *** jcsmith has joined #opensolaris [23:28:36] *** jcsmith has quit IRC [23:28:41] *** jcsmith has joined #opensolaris [23:31:04] <WickedWicky> skippy? [23:31:23] *** Gman has joined #opensolaris [23:34:17] <BadKarma> spooky [23:38:00] *** Mdx4 has joined #opensolaris [23:38:39] *** bra1 has joined #opensolaris [23:38:44] <bra1> Hello [23:39:02] *** baijiutong has joined #opensolaris [23:39:09] <bra1> Does anyone know why pkg-get refuses to work for me on solaris express [23:39:11] <bra1> ? [23:39:43] <iron_angel> hmmm, what sort of error do ye get? [23:39:53] <bra1> I followed their instructions and everything went smoothly, but when I try to use pkg-get it says pkg-get not found [23:39:55] <iron_angel> I was using Blastwave's pkg-get just fine under SXCE b63. [23:40:04] <iron_angel> oh, it's not in your PATH [23:40:09] <iron_angel> try /opt/csw/bin/pkg-get [23:41:11] <bra1> Thanks [23:42:51] <bra1> Do you also know off-hand the command to see what shell i'm using? [23:43:02] <iron_angel> you'll ultimately want to modify your PATH to include /opt/csw/bin, and perhaps /opt/csw/sbin, /opt/csw/gcc4/bin and /opt/csw/kde-gcc/bin [23:43:04] <bda> echo $SHELL [23:43:06] <iron_angel> echo $SHELL [23:43:18] <bra1> Thanks [23:44:20] <iron_angel> foo, blastwave still has XFCE 4.2. What's the holdup for 4.4? [23:44:37] <iron_angel> (I know I can build it easily on SXCE, but I'm not so sure about on Solaris 10) [23:44:54] *** mcnamarabrian has joined #opensolaris [23:50:43] *** pina has quit IRC [23:50:50] *** polk__ has joined #opensolaris [23:51:21] *** FelipeBare has joined #opensolaris [23:52:50] <Pietro_S> iron_angel: there is xfce project - holding 4.4 and also goodies [23:53:36] <iron_angel> Yeah, I saw that, for SXCE. Should that build OK on standard Solaris 10 too? [23:53:40] <iron_angel> (SPARC, if it matters) [23:55:56] <Pietro_S> well - not sure about it? is latest update out? That should have all dependencies - but cause I even didn't test xfce on sparcs with sxce - I wouldn't recommend it for production ... [23:56:59] *** glagasse has quit IRC [23:57:52] *** sommerfeld has quit IRC [23:58:11] *** neoxed has joined #OpenSolaris [23:58:20] *** sommerfeld has joined #opensolaris [23:58:21] *** ChanServ sets mode: +o sommerfeld [23:59:01] *** Tiger^ has quit IRC [23:59:58] <iron_angel> Pietro_S: I've built XFCE on SXCE on an Ultra 60, it worked fine. I'll have to experiment with it on Sol 10, then.