September 17, 2008 
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
September 17, 2008 [00:00:08] <TrX> LINK MEDIA STATE SPEED DUPLEX DEVICE [00:00:08] <TrX> e1000g1 Ethernet up 1000 full e1000g1 [00:00:54] <sstallion_work> are you running into a switch ? [00:01:19] *** FallenHi1okiri has quit IRC [00:01:36] <TrX> Yup, intel netstructure fiber switch [00:01:57] <TrX> a centos box on the switch using the same NIC can transfer 25MB/s to a windows desktop [00:02:26] <sstallion_work> a few quick things to check: disable flowcontrol, and enable jumbo frames [00:02:37] <TrX> flowcontrol is disabled [00:02:48] <sstallion_work> also, see if you can isolate the box and the switch as much as possible [00:02:54] *** postwait has quit IRC [00:03:03] <TrX> Would like to see the normal non jumboframe speeds (20MB/s instead of 10MB/s) before enableing jumbo frames [00:03:15] <TrX> sstallion_work, Thanks for the help, btw :) [00:03:19] <sstallion_work> what kind of a machine is this on ? [00:03:43] <TrX> A DL360 G4 [00:04:12] <sstallion_work> 4-way core 2 xeon isnt it ? [00:04:31] <TrX> 2x dual core 2.6ghz xeons yeah [00:04:32] <TrX> ddr ram [00:04:35] <TrX> u320 drives [00:04:36] <sstallion_work> ah okay [00:04:47] <sstallion_work> maybe it was the 580's that were 4-way [00:05:02] <sstallion_work> kstats look normal ? [00:05:04] <TrX> btw, flowcontrol is disabled on the switch, should I disable it on the NIC under solaris too? [00:05:10] <TrX> *how :P [00:05:32] *** spiki has quit IRC [00:05:35] <sstallion_work> one sec, let me see if e1000g supports brussels [00:05:39] <TrX> not touched solaris for a while i'm afraid, and my memory for commands is shoddy at best ;) [00:06:02] <sstallion_work> dladm show-linkprop e1000g1 will give you an idea [00:06:04] <sstallion_work> it should default to 'no [00:06:06] *** spiki has joined #opensolaris [00:06:35] <sstallion_work> (its default is bi, but if flowcontrol is disabled at the switch, it should be set to no) [00:06:52] <TrX> e1000g1 flowctrl bi bi no,tx,rx,bi [00:06:58] <sstallion_work> okay [00:07:00] <sstallion_work> shut that off [00:07:00] <TrX> Could be onto something here [00:07:15] *** danroberts has joined #opensolaris [00:07:25] <sstallion_work> # dladm set-linkprop -p flowctrl=no e1000g1 [00:07:42] <TrX> hehe, was just manning [00:07:43] <TrX> cheers [00:07:58] *** calAFK has left #opensolaris [00:08:13] <sstallion_work> dladm can be a little intimidating at first [00:08:19] <sstallion_work> but its handy as hell when you need it :) [00:08:38] <TrX> yeah, I need to get my head around ipf at somepoint as well [00:08:44] <TrX> i'm more an iptables person atm [00:08:48] *** h3sp4wn has quit IRC [00:09:00] <sstallion_work> yuck... I wouldnt say that very loud :) [00:09:14] <TrX> would i need to bounce that interface for the new property to take effect? [00:09:30] <sstallion_work> shouldnt [00:09:40] <sstallion_work> but it may not be a terrible idea to down the interface and then bring it back up [00:10:31] <sstallion_work> (ifconfig e1000g1 down; ifconfig e1000g1 up) [00:10:34] <TrX> Still getting 8MB/s unfortunately [00:10:40] <TrX> sstallion_work, I know that command ;) [00:10:55] <sstallion_work> what does your cpu load look like ? [00:11:11] <sstallion_work> also, issue a dlamd show-linkprop to verify that the property stuck [00:11:14] <TrX> Total: 86 processes, 299 lwps, load averages: 0.11, 0.05, 0.02 [00:11:15] <sstallion_work> dladm rather [00:11:18] <TrX> prstat [00:11:31] <sstallion_work> vmstat would be more useful here [00:11:49] <TrX> e1000g1 flowctrl no [00:12:01] <TrX> vmstat [00:12:01] <TrX> kthr memory page disk faults cpu [00:12:01] <TrX> r b w swap free re mf pi po fr de sr lf s0 s1 s2 in sy cs us sy id [00:12:01] <TrX> 0 0 0 1083608 827056 6 2 4 0 0 0 1 5 0 0 -1 450 100 322 0 1 99 [00:12:02] *** not-me-guv has joined #opensolaris [00:12:20] <sstallion_work> okay you need to run vmstat 1 and take the second sample... and don't /exec -o :P [00:12:26] *** not-me-guv is now known as niq [00:12:32] *** McBofh has left #opensolaris [00:12:36] *** McBofh has joined #opensolaris [00:12:42] <TrX> 0 0 0 377844 104288 35 62 0 0 0 0 0 1 0 0 0 380 420 355 0 0 100 [00:12:45] <danroberts> Hi All. Anyone know if there is an IRC channel for the townhall going on right now? [00:12:45] *** lisppaste3 has joined #opensolaris [00:12:48] *** spiki has quit IRC [00:12:55] <McBofh> danroberts: #opensolaris-meeting [00:13:07] <danroberts> tx! [00:13:41] *** spiki has joined #opensolaris [00:14:26] *** CIA-25 has quit IRC [00:14:32] <sstallion_work> TrX: was that done during the transfer? :) [00:15:12] *** Dominic has joined #opensolaris [00:15:39] <TrX> 0 0 0 378212 104656 11 59 0 0 0 0 0 0 28 55 0 7030 850 1711 0 11 88 [00:15:49] <TrX> ^ That is during transfer [00:16:01] <sstallion_work> looks fair [00:16:07] <sstallion_work> nothing anything off in your kstats ? [00:16:44] <TrX> say that again sorry? [00:16:45] <sstallion_work> (in particular the mac kstats) [00:16:56] *** lisppaste3 has quit IRC [00:17:14] *** lisppaste3 has joined #opensolaris [00:17:30] <TrX> You may have to be a little more verbose ;P [00:17:32] <TrX> *apologies [00:17:59] *** spiki has quit IRC [00:18:21] <sstallion_work> look at the output of 'kstat e1000g' and pay attention to the various error kstats in the mac module [00:19:05] <sstallion_work> i.e. kstat e1000g:1:mac [00:20:02] <sstallion_work> look at ierrors and oerrors (and the various collisions) [00:20:10] <TrX> sqe_errors 197270 [00:20:16] *** lisppaste3 has quit IRC [00:20:33] <TrX> link_autoneg 1 [00:20:33] <TrX> link_duplex 2 [00:20:35] *** lisppaste3 has joined #opensolaris [00:21:07] *** bahumbug has quit IRC [00:21:08] <sstallion_work> TrX: do those numbers increase while you are doing the transfer? [00:21:43] *** netj has quit IRC [00:21:45] *** spiki has joined #opensolaris [00:21:55] <TrX> nope [00:22:06] *** jbasse has quit IRC [00:22:08] <TrX> well packet counts do obviously, but nothing with error in the name [00:22:22] <sstallion_work> no collisions ? [00:22:34] *** _teo_ has joined #opensolaris [00:23:00] *** Vagrant has joined #opensolaris [00:23:29] <Vagrant> hi, which ver of sun studio should i use with sxce snv_97? [00:23:32] <TrX> tx_late_collisions 0 [00:23:39] <TrX> multi_collisions 0 [00:24:02] <TrX> ex_collisions 0 [00:24:03] <TrX> fcs_errors 0 [00:24:03] <TrX> first_collisions 0 [00:24:03] <TrX> ierrors 0 [00:24:10] <TrX> carrier_errors 0 [00:24:10] <TrX> collisions 0 [00:24:16] *** katohisa has joined #opensolaris [00:24:29] <sstallion_work> TrX: gotta cut that out, a simple no would have sufficed! [00:24:40] <TrX> apologies [00:24:48] <sstallion_work> Vagrant: to build onnv ? [00:24:54] <TrX> No, apparently not! [00:25:01] <Vagrant> sstallion_work: to compile programs in C ie ;) [00:25:10] <sstallion_work> TrX: have you double checked your media ? [00:25:43] <TrX> AM currently setting up a different transfer method just to check if it's the protocol [00:25:47] <sstallion_work> at this point, it would be better to isolate as many factors as possible. If you can go direct between two machines that would be much better [00:25:53] *** lisppaste3 has quit IRC [00:25:56] <sstallion_work> some switches just don't play nicely with some nic's [00:26:03] <sstallion_work> what protocol ? [00:26:04] <TrX> I checked this before, however since then we have disabled FC.. so i guess it's worth trying again [00:26:12] *** lisppaste3 has joined #opensolaris [00:26:14] <TrX> current test is the solaris CIFS service [00:26:16] <Vagrant> 11 or 12? [00:26:18] <sstallion_work> ugh [00:26:20] <sstallion_work> don't do that [00:26:20] <TrX> going to use iperf etc [00:26:25] <sstallion_work> Vagrant: it doesnt matter... use 12 [00:26:33] <sstallion_work> TrX: ftp is about as reliable as it gets [00:26:46] <TrX> was thinking apache tbh just for quickness [00:26:49] <Vagrant> but if i wanna build on [00:26:51] *** yongsun has joined #opensolaris [00:26:53] <sstallion_work> http does well enough [00:27:00] <TrX> 1 sec [00:27:15] <e^ipi> Vagrant: either is fine [00:27:16] <sstallion_work> Vagrant: you need to read the developer documentation. You need a patched version of studio 11 for onnv builds [00:27:29] <e^ipi> 11 is 'supported' [00:27:34] <e^ipi> 12 might break [00:27:34] <sstallion_work> well [00:27:40] <sstallion_work> 11 is the only thing that works in uts ;) [00:27:52] <e^ipi> file bugs [00:27:55] <sstallion_work> 12 yields some pretty 'interesting' results [00:28:01] <sstallion_work> meh, too busy fixing my own ;P [00:28:43] <sstallion_work> its not an easily reported issue. last time I attempted a full uts build with 12, the kernel built perfectly fine, but a module compiled on 12 could not be loaded into a kernel build with a different compiler [00:29:32] *** spiki has quit IRC [00:30:04] *** spiki has joined #opensolaris [00:30:55] *** bahumbug has joined #opensolaris [00:31:05] <TrX> sstallion_work, Downloading a 1.4GB file via apache, still only get 10MB/s [00:31:22] <sstallion_work> does the device support multiple PHY's ? [00:31:31] <TrX> It's strange that it's not just any bad speed, it's the exact speed i'd expect to see on a 100mbit line [00:31:39] <TrX> PHY's? [00:31:46] <TrX> physical interfaces? [00:31:50] <sstallion_work> 100mbits will hit ~12Mb/s at full duplex [00:32:08] <TrX> usually hits around 10MB/s in my experience [00:32:17] <TrX> actual transferred data that is [00:32:22] <sstallion_work> you havent been on many newer machines then ;) [00:32:26] <TrX> lol [00:32:30] <TrX> TCP offloading? [00:32:41] <Vagrant> thanks e^ipi and sstallion_work for advice [00:32:45] <Vagrant> ill try both [00:32:54] <lisppaste3> Apologies for the trouble today. The bot should stop bouncing in and out of your channel now. [00:32:55] <sstallion_work> Vagrant: sure thing. good luck. If you arent doing ON development, just use 12 [00:33:23] <sstallion_work> TrX: is the card fiber only ? [00:33:42] <TrX> yes [00:33:48] <TrX> 1x SX multimode port [00:33:58] <sstallion_work> are you certain you are using the correct media? (not all fiber is created equal...) [00:34:19] <TrX> yes, same fiber working fine for other cards on same switch (other servers) [00:34:33] *** e^ipi changes topic to "Latest builds: SXCE 97, ON 98, IPS 97 || Step one: see if SAG answers your question: http://docs.sun.com/app/docs/coll/47.16 || Clipboard: http://opensolaris.pastebin.ca" [00:34:49] <sstallion_work> at this point it would be best to isolate the device from the switch [00:35:11] <sstallion_work> (unless you have the same exact card working with the same exact media into the same exact switch) [00:35:15] *** katohisa has quit IRC [00:36:36] *** nicoatsun has left #opensolaris [00:36:44] <TrX> Same model card, tried with same piece of fibre, into same switch port. full speeds [00:37:05] <sstallion_work> what release of solaris? [00:37:10] <TrX> snv85 [00:37:12] <TrX> *95 [00:37:17] <sstallion_work> Indiana then? [00:37:20] *** timsf1 has joined #opensolaris [00:37:22] <TrX> SCDE [00:37:25] <TrX> *SXDE [00:37:32] <sstallion_work> there was no SXDE release at snv_85 [00:37:37] <sstallion_work> 79a was the last one [00:37:50] <TrX> Not internally [00:38:28] <TrX> I'm running the SNV_95 image [00:38:38] <sstallion_work> what does /etc/release reveal? [00:38:40] <TrX> Whatever they want to call that nowadays ;) [00:38:56] <TrX> Solaris Express Community Edition snv_95 X86 [00:38:56] <TrX> Copyright 2008 Sun Microsystems, Inc. All Rights Reserved. [00:38:56] <TrX> Use is subject to license terms. [00:38:56] <TrX> Assembled 28 July 2008 [00:39:04] <sstallion_work> that would be SXCE :P [00:39:10] <TrX> CE/DE, meh :P [00:39:42] *** alibb has quit IRC [00:39:44] <TrX> Still, should not really pose a problem [00:39:50] *** CIA-25 has joined #opensolaris [00:40:03] <TrX> could not see anything relating to a bad bug with 95 for network performance [00:40:38] *** mikl has quit IRC [00:41:11] <TrX> Also, how would you go about testing drive performance btw? [00:41:11] <sstallion_work> TrX: do you have other solaris hosts that use this device? [00:41:17] <TrX> Sadly not [00:41:21] *** timsf has quit IRC [00:41:50] <sstallion_work> well, the best you could probably do at this point is send an email to driver-discuss and networking-discuss [00:41:55] *** Vagrant has quit IRC [00:42:20] *** spiki has quit IRC [00:42:59] *** Dominic has left #opensolaris [00:43:19] *** spiki has joined #opensolaris [00:43:25] <TrX> No problems, thanks for all your help [00:43:53] <TrX> How would one go about testing performance read/write of a zfs pool btw? (just out of interest, unrelated) :) [00:44:14] <e^ipi> zpool iostat is a pretty shit measurement of it [00:44:22] <seanmcg> use filebench [00:44:28] <e^ipi> what seanmcg said [00:44:35] <seanmcg> or vdbench [00:45:03] <sstallion_work> vdbench... heh. [00:45:17] <e^ipi> screams "don't sit on it", don't it? [00:45:53] *** sstallion_work has quit IRC [00:47:19] <TrX> thanks seanmcg [00:47:52] <seanmcg> TrX, you should find filebench in /usr/benchmarks/... tis bundled with sxce [00:48:13] *** SpyKee has joined #opensolaris [00:48:27] <TrX> ahh awsome [00:48:32] <TrX> was just about to wget :P [00:49:27] *** spiki has quit IRC [00:49:29] <seanmcg> just be sure you're fileset is large enough that it doesn't get cached in the filecache :) [00:49:31] *** comay has joined #opensolaris [00:49:37] *** slant has quit IRC [00:49:56] *** ChanServ sets mode: +o comay [00:50:29] *** slant has joined #opensolaris [00:50:49] *** fr4g has joined #opensolaris [00:50:55] <TrX> IO Summary: 11508 ops 189.9 ops/s, (190/0 r/w) 0.4mb/s, 127us cpu/op, 5.2ms latency [00:51:06] <TrX> lol, I'm obviously doing something wrong there :P [00:51:41] *** h3sp4wn has joined #opensolaris [00:51:41] *** sstallion_work has joined #opensolaris [00:52:30] <seanmcg> the workloads I think by default use /tmp and are small in their fileset sizes [00:53:06] <TrX> Will edit the .prof to dump it's files on my zfs pool and use bigger files, [00:53:07] <TrX> thanks [00:53:39] *** SpyKee has quit IRC [00:53:52] *** SpyKee has joined #opensolaris [00:54:29] <seanmcg> theres some info on the .f filebench files on www.solarisinternals.com [00:54:29] *** luc^ has quit IRC [00:55:39] *** dunc has quit IRC [00:57:26] *** _teo_ has quit IRC [00:57:51] *** dunc has joined #opensolaris [00:58:22] *** bigjohnto has quit IRC [00:58:42] *** gm152 has joined #opensolaris [01:00:12] *** spiki has joined #opensolaris [01:01:37] *** capaz has left #opensolaris [01:04:36] *** fr4g_ has joined #opensolaris [01:07:22] *** SpyKee has quit IRC [01:08:24] *** tg has quit IRC [01:09:04] *** yongsun has quit IRC [01:10:37] *** storycrafter has left #opensolaris [01:11:24] *** tg has joined #opensolaris [01:15:28] *** niq has quit IRC [01:15:41] <TrX> another quick question on the network speeds issue [01:16:07] *** twisti has quit IRC [01:17:09] <TrX> the centos box that gets 20MB/s (looking on the switch status info's) seems to be sending out majoritivley big packets (1024) while the solaris box seems to be sending out many more smaller packets and no large packets [01:17:31] <TrX> This I am thinking could be affecting the speed somewhat. Is there a way to tell solaris to send only bigger packets [01:17:54] <TrX> (not talking jumbo frames here.. just 'filling' the frames properly as it were) [01:18:02] *** not-me-guv has joined #opensolaris [01:23:45] *** fr4g has quit IRC [01:23:56] *** bahumbug has quit IRC [01:27:28] <timsf1> star! [01:27:33] <asyd> * [01:27:34] *** timsf1 is now known as timsf [01:28:15] <seanmcg> is star a gnu tool ?-) [01:28:27] <timsf> bound to be. [01:29:16] <e^ipi> heh [01:29:54] * e^ipi not so great at phone conversation [01:31:43] <seanmcg> who is.. humans are built for face-to-face conversation :) [01:32:00] <e^ipi> i can do text plenty fine [01:34:02] <xRaich[o]2x> to bad i missed the first half hour -_- [01:35:26] *** Saltsa has quit IRC [01:35:56] *** sures1 has joined #opensolaris [01:36:26] <timsf> Yep, I think that was a worthwhile meeting [01:36:36] *** slant has quit IRC [01:36:44] <sures1> meeting format is good [01:37:05] <e^ipi> other channel [01:37:15] <e^ipi> #opensolaris-meeting [01:37:18] <e^ipi> :) [01:37:33] <timsf> they'll work it out. [01:38:14] *** Saltsa has joined #opensolaris [01:40:00] *** stanchion has joined #opensolaris [01:41:11] *** jtmuzix has joined #opensolaris [01:41:25] <stanchion> anyone experience 'zpool replace' command never completing? i did a replace yesterday and it still says replacing... (resilver completed successfully?) [01:42:16] <stanchion> looks like maybe it is a bug :( [01:45:03] *** danroberts has quit IRC [01:51:29] *** dunc has quit IRC [01:51:37] *** dunc has joined #opensolaris [01:51:40] *** NCommander has joined #opensolaris [01:52:02] <NCommander> Is there a relatively straightforward way to create a OpenSolaris chroot jail? [01:52:28] <NCommander> I'm dealing with a legacy app which excutes commands in a chroot, and I've not been able to convience it to use a jail instead [01:52:30] <NCommander> er, zone [01:52:35] * NCommander smacks the BSD out of himself [01:52:55] *** spiki has quit IRC [01:57:28] <e^ipi> NCommander: man zoneadm [01:57:35] <e^ipi> hold on, benr has a good blog entry on it [01:57:55] <e^ipi> i keep my zones on zfs snapshots, on account of they're tiny [01:58:19] <e^ipi> here you go http://www.cuddletech.com/blog/pivot/entry.php?id=751 [01:58:23] <NCommander> e^ipi: I'm dealing with an app that expects a chroot, not a jail. (its some evil bastardized inhouse thing that has been passed down the generations of sysadmins :-/) [01:58:31] <NCommander> s/jail/zone/g [01:58:47] <e^ipi> it 'expects' a zone? [01:58:52] <e^ipi> there's no real way to tell [01:59:03] <NCommander> No, it sorta expects to do chroot *address* *command* [01:59:06] <e^ipi> it looks like a regular full system [01:59:12] <e^ipi> oh, that's retarded [01:59:18] <NCommander> like I said [01:59:23] <NCommander> Bastardized script from hell that was inheriated [01:59:46] <e^ipi> you can just strip that out and have everything run on the "full" system [01:59:50] <timsf> Rewrite it and do your bit for future generations? [01:59:50] <e^ipi> but your full system is just actually a zone [02:00:16] <NCommander> timsf: the problem is I can't even figure out what extactly it does :-/. It does some build management I think, but beyond that, its a mixture of superglue and perl [02:00:42] <NCommander> I just know when it breaks, Bad Things(tm) happen [02:01:07] <timsf> Run the script, with execsnoop running in the background - worked for me in the past... [02:01:22] <bda> DTT++ [02:01:38] <NCommander> This I shall try [02:01:47] *** Openfree has joined #opensolaris [02:04:50] <e^ipi> i'm going to assume that you, like any other sane shop, have dev -> staging -> prod architecture [02:05:06] <e^ipi> so just grep the files in question for 'chroot' and strip that bit out [02:05:34] <e^ipi> you'll probably also need to look for references to files in /path/to/chroot and strip those out as well [02:05:45] <e^ipi> then just try it, see if anything breaks if you shove it in a zone [02:05:52] *** Openfree has quit IRC [02:06:05] * NCommander nods [02:06:07] <NCommander> THanks [02:06:13] <NCommander> That should be a great help [02:06:21] <NCommander> and you wish on the proper dev environment [02:06:27] <NCommander> My co-worker working on creating that though [02:06:47] <e^ipi> heh, i know, how presumptuous of me to assume a sane architecture [02:06:48] <e^ipi> ;) [02:09:22] * NCommander is also working on nexenta at the same time to mix Ubuntu goodness with Solaris awesomeness [02:10:57] <e^ipi> yech [02:11:18] <e^ipi> then you get all the gnu tools [02:11:20] <e^ipi> and those are broken [02:12:15] * NCommander isn't here to start a holy war ;-) [02:12:29] <e^ipi> at the very least if you don't want to consider engineering quality they don't support a lot of the solaris features [02:12:33] <e^ipi> ACL's for example [02:13:06] <NCommander> I'm aware [02:13:32] <NCommander> I thought Nexenta still had the sun tools available [02:13:39] <e^ipi> yeah, they're around [02:13:46] <e^ipi> hidden in a subpath ghetto [02:14:00] * NCommander nods [02:14:10] <e^ipi> I imagine when indiana ( 2008.05 ) stabilizes a little, nexenta will more or less fade away [02:14:20] <e^ipi> the main complaint people have with solaris is the package management [02:14:33] <e^ipi> I consider it a feature not to have network package repos, but i'm in the minority [02:14:38] <e^ipi> IPS brings that [02:14:42] <bda> Yes, because you're crazy. :) [02:15:25] *** libkeiser has joined #opensolaris [02:15:27] <e^ipi> perhaps [02:15:53] <NCommander> What's so special about Indiana [02:16:02] <e^ipi> NCommander: it's got sun backing it [02:16:18] <e^ipi> which hopefully will translate in to high engineering quality [02:16:39] <NCommander> meh. I wasn't that impressed when I played with it, since its package management system seemed to be based off the old Sun one [02:17:04] <bda> It incorporates the old packages. [02:17:12] <e^ipi> not really, a bunch has changed [02:17:12] <bda> Which is a pretty good thing. :P [02:17:27] <bda> But it's all new code and architecture. [02:17:29] <e^ipi> the package naming will almost certainly change fairly soon ( no more SUNWfoo ) [02:17:44] <bda> e^ipi: That does annoy me. [02:17:57] <e^ipi> indeed [02:18:35] *** ajmcello has quit IRC [02:19:53] <Bartman007> e^ipi: when will dpkg and apt-clone be added? [02:19:59] * Bartman007 ducks [02:21:01] *** danroberts has joined #opensolaris [02:21:08] <e^ipi> heya dan [02:21:51] *** danroberts has quit IRC [02:22:00] <e^ipi> okiedokie [02:23:26] *** Fullmoon has joined #opensolaris [02:24:00] <Praematura> Having used Indiana for a few months now (and keeping it updated pretty frequently), I have to say I'm extremely happy with the new package management system. [02:24:08] <Fullmoon> So, what bad things happen when I create a user with >8 chars username? [02:24:18] *** jtmuzix has quit IRC [02:24:19] <e^ipi> i dunno, try it [02:24:41] <Fullmoon> I see no bad effects, but adduser complains (but adds the user anyway) [02:24:44] <Praematura> Absolutely nothing, unless you have another user with the same first 8 characters. [02:25:16] <e^ipi> alternately, put together a patch and an arc case and we'll get it in the tree [02:26:03] <jamesd_> or you run some application that takes only 8 characters for the user name and doesn't catch the over load and it silently corrupts your user data or much worse your config file or the database that backs your application and brings down your systems during end of year reports... [02:26:25] <Praematura> .... that too. [02:30:44] *** alanc is now known as alanc_away [02:31:22] *** Odin- has quit IRC [02:32:42] *** comay has quit IRC [02:35:48] *** xinkeT has quit IRC [02:36:33] *** sstallion has joined #opensolaris [02:40:29] *** stanchion has quit IRC [02:45:01] *** xinkeT has joined #opensolaris [02:49:25] *** sures1 has quit IRC [02:50:07] *** LordIllpalazo has quit IRC [02:50:22] *** lolmac has quit IRC [02:50:48] *** Fullmoon has quit IRC [02:52:33] *** myrkraverk has quit IRC [02:54:01] *** Octan1 has joined #opensolaris [02:54:04] *** NCommander has quit IRC [02:54:59] *** Octan1 has left #opensolaris [02:55:18] *** Fullmoon has joined #opensolaris [02:59:06] *** dnm has joined #opensolaris [02:59:48] *** crichardso has quit IRC [03:03:17] *** timsf has quit IRC [03:04:41] *** duxex has joined #opensolaris [03:06:02] *** RElling1 has joined #opensolaris [03:07:51] *** duxex has quit IRC [03:09:25] *** derchris^ has quit IRC [03:09:30] *** derchris has joined #opensolaris [03:10:27] *** RElling has quit IRC [03:14:00] *** ker2x has joined #opensolaris [03:20:52] *** timsf has joined #opensolaris [03:23:04] *** bhall has quit IRC [03:23:43] *** neonum6_ has joined #opensolaris [03:27:04] *** bhall has joined #opensolaris [03:35:09] *** xRaich[o]2x has left #opensolaris [03:37:55] <dclarke> anyone out there have a screaming fast Sparc ? [03:38:08] <dclarke> or really really slow ? like less than 60MHz ? [03:38:28] * dclarke wonders if that is confusing [03:38:36] <timeless> pizzaboxes :) [03:38:52] * timeless wonders why dclarke would be interested in those options [03:39:20] <dclarke> just looking for a benchmark data point [03:40:25] <jamesd_> dclarke, i have a ss5 70mhz that still boots... and maybe a 40mhz ss10 if it still powers up... my fastest sparc is a blade 1500, 1.05ghz or a blade 1k with dual 750's... [03:40:30] *** neonum6 has quit IRC [03:40:49] *** likaijun has joined #opensolaris [03:40:50] <dclarke> just wget http://www.blastwave.org/dclarke/stuff/mpfr_speed_sparc_v7 [03:41:23] <dclarke> but you need libgmp and libmpfr somewhere [03:42:41] * timeless sighs at 3/4 letter lib names [03:43:04] <dclarke> okay .. gmp and mpfr then [03:43:30] <timeless> yeah, still doesn't help me understand what they are :) [03:43:55] <dclarke> ohhh .. sorry [03:43:59] <jamesd_> dclarke, i may find time to power on the old slow boxes this weekend. [03:44:06] <dclarke> gmp = GNU MKultiple Precision lib [03:44:12] <dclarke> gmp = GNU Multiple Precision lib [03:44:30] <dclarke> mpfr = Multi Precision Floating Point library [03:46:21] <jamesd_> of course i have old slow boxes that do better than sparc for those kind of things.... SGI indigo2 R4400-200 and SGI indy R5000-150, SGI indy R4300-133mhz IIRC. [03:46:43] *** kohju is now known as KOHJU [03:47:51] <dclarke> see http://www.blastwave.org/dclarke/stuff/mpfr_speed.txt [03:48:17] <dclarke> it calculates pi to the number of digits specified [03:48:33] <dclarke> so .. for loop 1 you get 2^(3+1) digits [03:49:06] <jamesd_> oh... i have mathmatica for my SGI boxes for that... just have to tell them its still 1994 for the license to still be good ;-) [03:49:08] <dclarke> for loop 5 you get 2^(5+1) digits [03:49:20] *** not-me-guv has quit IRC [03:49:24] <dclarke> mathematica .. those shiesters [03:49:50] <dclarke> I have mathematica 5.5 here and to install it I have to call them up and explain that I'm going to install it and can I have a license key pl;ease [03:49:56] <dclarke> after I paid through the node for it [03:50:03] <jamesd_> it came with the BOX... 2x indy's for $12 plus $20 for shipping about 5-7 years ago.. its a steal even now. [03:50:07] *** sartek has quit IRC [03:50:08] <dclarke> they always ask .. why are you installing this on a new machine ? [03:50:15] <dclarke> total steal [03:50:18] <dclarke> big time [03:50:24] <dclarke> heck .. it is $3500 !! [03:50:47] <dclarke> anyways .. I'm exercising the mpfr and gmp libs [03:50:51] <dclarke> thats the point there [03:51:18] <dclarke> and a Sparc 20 does half a million digits in 871.46282 [03:51:47] <dclarke> same task on an 2GHz AMD Opteron is 4.4 secs [03:51:54] <dclarke> talk about speed [03:52:03] <dclarke> imagin looking back to those days [03:52:17] <dclarke> oh boy .. a 33MHz 486DX2 [03:52:25] <dclarke> sorry .. 66MHz 486DX2 [03:52:40] <dclarke> DX2 actaully meant somthing and computers had .. turbo buttons [03:52:45] <jamesd_> my first linux box was a 25mhz 386sx .... [03:52:50] <timeless> turbo buttons were great [03:52:54] *** bourgois has joined #opensolaris [03:53:08] <jamesd_> i still have the motherboard... not in a case anymore i'm affraid.. [03:53:10] <dclarke> up until recently I still had a running 486 here [03:53:20] <dclarke> it would prbably boot oo [03:53:34] <dclarke> 386sx .. I think sx means .. no FP ? [03:54:15] <jamesd_> dclarke, yeap... and since the linux build process used awk which used fp, meant that a kernel build took over 18 hours even with 5MB of ram. [03:55:29] <dclarke> 5MB of RAM back then .. would be .. thousands of dollars [03:55:31] <dclarke> er ... [03:55:38] <dclarke> no .. many hundreds though [03:56:00] <dclarke> I have a machine from 1995 that has 96MB of RAM and it was a killer expensive box [03:56:09] <jamesd_> nah... $40 a MB... [03:56:19] <dclarke> with PCI SCSI and IDE [03:56:29] <jamesd_> now $40 gets you $1GB on a good day.... [03:56:35] <jamesd_> er 1GB [03:56:40] <dclarke> yeah [03:57:04] <dclarke> oh .. now I recall the Atari 1040ST .. it was the box that broke the $1 per meg barrier [03:57:09] <jamesd_> this was only ide, and actually i had an mfm 40MB drive in it as well .. i was really poor back then. [03:57:11] <dclarke> that was .. 1989 or so I think [03:57:18] <dclarke> gotta go look it up [03:57:50] <dclarke> http://en.wikipedia.org/wiki/Atari_ST [03:58:07] *** karrotx has joined #opensolaris [03:58:10] <dclarke> Release date 1985 [03:58:16] <dclarke> crap .. now I feel old [03:58:26] *** ninjaslim has joined #opensolaris [03:58:27] <jamesd_> weirdstuff back in those days totally rocked (probably was, but it was local to me back then) got a 9600 baud modem for $5 and a 40mb harddrive with controller for $5 more. [03:59:37] <benley> yay weirdstuff [04:00:06] <benley> I should have gone to visit them when I was in the area last week [04:00:32] <ninjaslim> what type of users is sxce intended for? [04:00:51] <benley> crazy ones [04:01:12] <ninjaslim> no seriously [04:02:07] <e^ipi> ones that really like S10 [04:02:21] <e^ipi> and people that want to work on ON [04:02:41] *** sartek has joined #opensolaris [04:02:50] <jamesd_> people that want the stability of solaris, and but willing to sacrafice a bit in stability ( theorectical at least) for new features. [04:04:08] <ninjaslim> ok [04:04:25] <ninjaslim> well i'm trying to get SXCE to boot in vbox, but loading Java can't take over 15 min [04:04:54] <jamesd_> and the number one feature of SXCE currently is XVM/Xen and maybe crossbow. [04:05:14] <benley> I'm a fan of the cifs server [04:05:16] <jamesd_> ninjaslim, give it more memory and processing power. [04:05:31] <jamesd_> benley, yeah its nice.. but samba is good enough most of the time [04:05:32] <ninjaslim> well it's running with 512MB ram [04:05:37] <benley> yeah I guess [04:05:42] <benley> ninjaslim: get more ram [04:05:54] <benley> ninjaslim: especially if you want to use zfs [04:06:03] <ninjaslim> not running ZFS in a vm [04:06:20] <jamesd_> ninjaslim, use an iso instead of a burned disk if possible... harddrives are faster than cd/dvd's [04:06:33] <ninjaslim> well all i've used is an iso [04:06:56] <ninjaslim> i guess i'll have to tinker with it as vbox is very picky in terms of guest os [04:08:07] *** sommerfeld has quit IRC [04:08:48] <jamesd_> allocate a second cpu/core if you have it.... but if you gave it two allready, try with one... all depends on your hardware which is better. [04:10:51] *** sergiusens has joined #opensolaris [04:13:43] <ninjaslim> vbox supports smp? [04:15:55] *** PicCard has quit IRC [04:15:58] *** sstallion has quit IRC [04:16:56] *** bourgois has quit IRC [04:17:48] <jamesd_> don't know... i use xen or vmware for that kind of thing.. and they both do. [04:19:35] <ninjaslim> i see i think i better invest in vmware fusion 2 as vbox is still a wip [04:20:22] <likaijun> hello, all.I make a customed distribution ISO,but my boot_archive (120M) is very large after installation. can someone give me some suggestion reducing its size? [04:24:05] *** timsf has quit IRC [04:24:18] *** skillet has joined #opensolaris [04:25:18] *** tomservo291 has joined #opensolaris [04:26:54] <tomservo291> hey.. having some serious issues getting MySQL 64bit (coolstack) running properly (InnoDB particularly) [04:27:10] <tomservo291> how can I verify that open solaris is currently running in 64bit mode? [04:29:15] <im_alone> tomservo291, not running 64-bit user app => not running 64-bit kernel [04:29:33] <im_alone> try to run a 64-bit app [04:29:45] <tomservo291> i know .. i just wanted to verify kernel was running in 64 bit mode [04:29:59] <im_alone> uname -a [04:30:04] <tomservo291> it is.. "isainfo -vk" shows 64-bit amd64 kernel modules [04:30:30] <tomservo291> uname -a shows i386/i86pc [04:32:03] <tomservo291> not really sure what the problem is... i let the csk mysql install itself before, and I just got around to re-doing it manually with a different data dir. when i use the db_install_mysql script, I'm getting an error about resolveip not being able to resolve the hostname, or localhost.. and have to use --force, and then things just get worse when trying to actually run mysql [04:32:25] <tomservo291> most info i can find on the problem point to people trying to run 64bit mysql on a 32bit kernel... which does not appear to be the problem [04:32:31] <im_alone> int main() { printf("Hello world.\n"); return 0; } gcc -march=k8 hello_world.c -o hello_world ; ./hello_world [04:32:37] <tomservo291> both "resolveip hostname" and "resolveip localhost" do in fact work [04:34:15] <tomservo291> your example works [04:36:43] *** djgregor has quit IRC [04:37:09] *** hotpockets has joined #opensolaris [04:37:43] <hotpockets> WHASSSSSSSUUUUUP [04:38:08] <hotpockets> Who do I complain to about sunstudio12 segfaults? [04:38:12] *** MrBIOS_ has joined #OpenSolaris [04:38:47] <hotpockets> I have x64, all the patches,etc [04:40:04] <hotpockets> on x86 also, it segfaults on make for SUNWgnome-media.spec [04:42:37] *** skillet has quit IRC [04:42:59] *** CIA-25 has quit IRC [04:43:23] *** KOHJU is now known as kohju [04:44:48] *** djgregor has joined #opensolaris [04:56:14] *** skillet has joined #opensolaris [04:56:27] *** tomocha6 has quit IRC [05:01:50] *** coffman_ has joined #opensolaris [05:03:50] *** tomservo291 has quit IRC [05:14:20] *** rtor has quit IRC [05:18:57] *** coffman has quit IRC [05:19:38] *** skillet has quit IRC [05:21:42] *** stukag has quit IRC [05:24:08] *** Ditegen has quit IRC [05:27:29] <Fullmoon> Are there any documents on how to lay out your file system, especially what goes into /usr and what into opt when installing new software? [05:30:45] *** ken_ has joined #opensolaris [05:34:40] *** alibb has joined #opensolaris [05:38:54] *** stukag has joined #opensolaris [05:39:15] *** stukag has quit IRC [05:42:22] *** cchapman has joined #opensolaris [05:43:35] <jamesd_> Fullmoon, simple if it came with your OS it belongs in /usr if it came from else where it goes in /opt ... so /usr/local that doesn't ship with your OS is techically incorrect [05:45:41] *** jrsharp has joined #opensolaris [05:48:28] *** kohju is now known as KOHJU [05:50:53] <jrsharp> how does one set the default route when using nwam? [05:51:11] *** postwait has joined #opensolaris [05:51:39] <postwait> So sunsolve is a piece of shit... it's giving me an error message on the change contract link. [05:51:52] <postwait> anyone know the steps to reset/set an ILOM password on an x4150? [05:54:09] * jrsharp only has eLOM machines [05:54:51] *** MrBIOS- has joined #OpenSolaris [05:55:32] *** MrBIOS-_ has joined #OpenSolaris [05:55:57] <postwait> elom is sparc? [05:56:10] <postwait> Or does the x4150 have an elom and I'm just slow? [05:56:10] <McBofh> no [05:56:28] <McBofh> have you looked on docs.sun.com? istr there being some useful doco there [05:56:37] * McBofh goes looking.... [05:56:48] <postwait> I found a doc, but sunsolve is being pissy about both of my active support contract numbers. [05:56:51] <postwait> 1-61-210475-1 [05:56:53] <jrsharp> postwait: I have an x2100 [05:58:18] *** mgedisman has joined #opensolaris [05:58:35] <mgedisman> Hello everyone! =) [05:59:43] <McBofh> mgedisman: hi [06:00:04] <McBofh> postwait: that doc is for sparc LOMs [06:00:18] <mgedisman> I got distracted with MySpace and now I can't play with Solaris. =( [06:00:18] <postwait> Oh.. well. [06:05:24] *** lloy0076 has joined #opensolaris [06:05:31] <McBofh> postwait: can you login at all to the ilom? [06:05:52] <postwait> No... I think I got it. [06:06:06] <postwait> info in the the PDF version of the x4150 install guide [06:06:11] *** lloy0076 has quit IRC [06:06:13] <postwait> page 33. [06:06:24] <postwait> I've forwarded it to the "man on deck", hope that delivers salvation. [06:06:33] <postwait> Thanks for helping McBoth [06:06:37] <McBofh> yw [06:06:53] <postwait> Would have spent hours trying to get that damn doc only to find it was sparc. [06:07:06] * McBofh specialises in roadblock removal [06:07:06] <postwait> Now I get to call Sun tomorrow and chew them out. [06:07:19] <postwait> 3rd time my sunsolve account hasn't worked. [06:07:35] <postwait> how hard is it to run a @#$^ing support site. [06:12:43] *** IvanR_ has joined #opensolaris [06:14:29] *** sah-work has joined #opensolaris [06:17:39] *** Ditegen has joined #OpenSolaris [06:22:48] *** CIA-25 has joined #opensolaris [06:27:22] <e^ipi> postwait: incredibly? [06:27:59] <postwait> e^ipi: heh. shouldn't be that hard. [06:28:25] <e^ipi> do you have any idea how many products Sun has? [06:28:38] <postwait> less that 10MM [06:28:43] <jamesd_> e^ipi, store.sun.com can give you an idea. [06:28:48] <postwait> and likely less than 100MM accounts [06:29:01] <postwait> so.. smaller than many of the sites I work on... and they don't break all the time. [06:30:05] <postwait> Given the amount people pay for sun support, it should provide a good basis for having the service available online all the time. [06:30:19] <postwait> I know I pay a significant price. [06:31:51] <postwait> I like Solaris and I like their hardware products... just the docs and sunsolve site could work better. [06:32:04] <postwait> And by better I mean "QoS and availability" [06:33:34] *** Ditegen_ has joined #OpenSolaris [06:34:22] *** dystopia has joined #opensolaris [06:39:19] *** Gekz has quit IRC [06:40:57] *** Ditegen has quit IRC [06:40:58] *** Ditegen_ is now known as Ditegen [06:41:40] <Praematura> Having been involved in supporting a support site, they are certainly not easy to run by any stretch of the imagination, even when one has only one product. [06:42:02] *** Gekz has joined #OpenSolaris [06:45:49] *** mikl has joined #opensolaris [06:47:39] *** alibb has quit IRC [06:49:24] *** hotpockets has left #opensolaris [06:51:08] <dystopia> i cant get my areca to work :( [06:51:36] *** anilg has joined #opensolaris [06:54:55] <McBofh> dystopia: let me summon somebody know knows about areca [06:54:59] * McBofh summons [06:55:11] *** jmcp has joined #opensolaris [06:55:45] <jmcp> somebody mentioned Areca [06:55:51] <e^ipi> magic. [06:55:58] <jmcp> :) [06:56:50] *** yoursdai has joined #opensolaris [06:57:14] <yoursdai> hi [06:58:18] <jmcp> hello [06:58:29] *** tomocha6 has joined #opensolaris [07:01:57] <yoursdai> I am using the zfs to replicate zvol from server A to server B.I use it in my program with a thread.Now I could not judge whether it is transfer all data.Anybody could help me? [07:02:27] *** airjump has joined #opensolaris [07:02:51] <dystopia> jmcp: yes, i would like to use my areca controler on (open)solaris, but once i load the driver, the computer either crashes or it can't read the volume [07:03:18] <dystopia> its a 1210 btw [07:03:38] <jmcp> which version of the driver? [07:03:54] <jmcp> and which OpenSolaris build? [07:04:00] *** sah-work has quit IRC [07:04:27] *** Molle has quit IRC [07:05:17] <dystopia> driver, 1.20.00.15-71010 installed after installation of OS hmm the OS, i tried Solaris 10/Q3, OpenSolaris 2008/05, Nexenta 1.0.1, Nexenta 2 beta, NexentaStor [07:05:40] *** Molle has joined #opensolaris [07:06:28] <jmcp> Solaris 10/Q3 ? [07:06:38] <jmcp> I can't comment on the Nexenta versions [07:07:41] <jmcp> so for you "can't read the volume" case, which OpenSolaris build/distro/... were you using? [07:07:51] <dystopia> i dont remember the exact solaris version it was 10/something, i grabbed it from sun.com in july [07:08:07] <dystopia> OpenSolaris 2008/05 [07:08:13] <jmcp> that's *not* Solaris 10 [07:08:24] <dystopia> i tried all of the mentioned OSes [07:08:48] <dystopia> upgraded firmware of the card, no can do... [07:09:10] * jmcp repeats::::: so for you "can't read the volume" case, which OpenSolaris build/distro/... were you using? [07:09:47] <jmcp> in addition, when "the computer crashes" what do you mean by that exactly? [07:10:07] <jmcp> I'm trying to figure out more about your problem [07:11:10] <dystopia> it totally hangs, need to do a reboot by pressing the hardware power off 4 sec [07:11:26] *** cchapman has quit IRC [07:11:26] <dystopia> i have to look up the exact version i guess, let me grab the cd [07:11:28] <jmcp> again on the volume - what do you mean by "can't read the volume?" is that "the volume shows up in the card bios, but not in OpenSolaris"? or something else [07:12:59] <dystopia> well with dmesg i see the firmware version and so on, but then ZFS doesn't detect anything, it can't import the zpool. i look around at the devices list, but the correct device isn't listed [07:13:59] <jmcp> are _any_ arcmsr-attached devices listed? [07:14:26] *** Mdx4 has joined #opensolaris [07:14:27] <dystopia> only arcmsr0 [07:14:44] <dystopia> and if i understand that isn't attached [07:14:59] <jmcp> could you pastebin the output from "prtconf -v" please [07:15:45] <jmcp> which release/distro/build are you running at the moment, btw? [07:17:01] <dystopia> none cause i couldnt get areca controller workong i'll brb getting it [07:17:06] <jbk> evening [07:17:09] <jmcp> hi jbk [07:17:17] <jbk> so nice to have power, running water, and internet [07:17:26] <jmcp> jbk: you survived! [07:17:37] <jbk> yeah [07:18:18] <jbk> so far from what i've seen on the news, the death count has been in single digits (which is far better than it could have been) [07:18:38] <jmcp> I heard 25 [07:18:43] <jmcp> at least I thought I did, anyway [07:18:57] <jbk> oh [07:18:57] *** KOHJU is now known as kohju [07:19:01] <jbk> might be [07:19:16] <jbk> i've only had spotty coverage [07:19:21] <jbk> from what i could read on my phone from time to time [07:19:56] *** gm152 has quit IRC [07:24:27] *** mgedisman has quit IRC [07:24:43] *** rtor has joined #opensolaris [07:26:33] <dystopia> phew, found the cd [07:27:39] <dystopia> how do i see the version when i have the CD, without booting it? [07:29:18] <jmcp> depends [07:29:33] <dystopia> i'm booting it in a VM, it says 2008.05 [07:29:36] <jmcp> ok [07:30:53] <jbk> but still, considering 40% of galveston stayed despited being told to leave, (as well as similar numbers in other areas) [07:31:30] *** stux|work is now known as stux [07:32:00] <dystopia> when i tried format, it did see 'something' raw data of my whole raid array, but it did not see my ZFS [07:32:29] <jmcp> did format see the same number of targets+luns that you have defined in the card's bios? [07:32:42] <dystopia> yes [07:32:58] <dystopia> the filesystem which was on the RAID was one zpool, version 6, made on freebsd 7.0.. iow no raidz because was done in hardware.. [07:33:24] <jmcp> right [07:33:28] <jmcp> did you try to zpool import ? [07:33:48] <dystopia> yes [07:34:03] <jmcp> and? [07:34:09] <dystopia> it didnt find it [07:34:31] <dystopia> when i booted freebsd 7, it did see it, and linux fuse-zfs as well [07:34:50] <jmcp> so can we restate the problem as "zpool created on freebsd7 is not imported on OpenSolaris 2008.05" ? [07:34:52] *** ninjaslim has quit IRC [07:35:13] <dystopia> ok, but 'not imported'.. the whole machine sometimes crashed [07:35:36] *** rtor has quit IRC [07:35:48] <jmcp> does a zpool scrub come back clean on freebsd7 after you tried importing it on OpenSolaris2008.05? [07:35:52] <dystopia> either right during boot after isaw firmware, or when i tried importing, or neither (but then i could only format it) [07:36:10] *** uebayasi has quit IRC [07:38:14] <dystopia> its snv 86 from january 2008 [07:43:03] *** axisys has joined #opensolaris [07:45:33] <dystopia> interestingly in the meanwhile finally a new version of the driver was released. i used 1.20.00.15-71010 while begin august 2008 1.20.00.16-80731 got released the changelof is ftp://ftp.areca.com.tw/RaidCards/AP_Drivers/Solaris/DRIVER/ReleaseNote.txt do you think this would solve my problem? [07:46:18] <dystopia> i've been able to create a zpool using 4 sata drives, this all worked fine, so i figured the rpbolem was my areca controller [07:46:34] <jmcp> it might be, I don't konw [07:46:35] <jmcp> know [07:48:00] *** tomocha6 has quit IRC [07:48:24] <dystopia> hmm, if its ok if i try it out the card with ro ext3fs i'm willing to give it a try, i guess with the latest opensolaris build + the latest areca driver [07:48:32] <jmcp> I do know, however, that I pushed 1.20.00.15Sun to build 99 yesterday [07:49:13] <dystopia> great, thanks a lot! that means no more installing manual of driver :-) [07:49:24] <jmcp> that's right :) [07:49:30] <jmcp> as soon as build99 comes out... [07:49:34] <dystopia> in the HCL, only one areca card is mentioned [07:49:38] <dystopia> i will wait for that then [07:49:40] <jmcp> and it'll be in the next OpenSolaris2008.* release [07:49:56] <jmcp> it closed yesterday, so perhaps 2 or 3 weeks, should be available [07:50:13] <jmcp> hmm [07:50:24] <jmcp> the version i pushed included the fixes that Erich lists for 1.20.00.16 [07:51:45] <dystopia> sweet so i'll wait a few weeks and can try, thats great, then i can backup my current data to 4 software raid, and use a clean ZFS on Areca [07:52:08] <jmcp> heck, you can install OpenSolaris _to_ it, from the installer, once the build is out there [07:53:09] <dystopia> that is what i want to do, and with ZFS i can easilly assign zones and brandz and use NFS or CIFS for the VMs [07:54:00] <dystopia> the only thing i will miss is OpenBSD PF, but i guess i can live with IPF [07:54:09] <jmcp> I'm sure you'll be fine [07:57:08] <dystopia> and the package system i'm not used to, but that can be learned as well, i really love how well NFS/CIFS/iSCSI are integrated in ZFS i'm thrilled to use that! :-) [07:57:35] <jmcp> yeah, that integration is *really* cool [07:58:49] <dystopia> i'll also use ZFS on my Mac once it has rw support [07:59:13] *** vk5foss has joined #opensolaris [07:59:21] * jmcp runs away from the mac user [07:59:27] <dystopia> :P [07:59:42] <dystopia> sorry;0) [07:59:46] <jmcp> :-) [08:00:21] <dystopia> thanks for your feedback though :-) attempting to pateintly looking forward to build 99 [08:00:35] <jmcp> yw [08:00:41] <Bartman007> dystopia: technically it has rw support already [08:01:05] *** postwait has quit IRC [08:01:25] <dystopia> oh, hmm, i didnt know that, since which OS X version ? [08:01:31] <Bartman007> http://zfs.macosforge.org/ [08:02:07] *** MrBIOS-- has joined #OpenSolaris [08:02:44] *** MrBIOS--_ has joined #OpenSolaris [08:03:15] <dystopia> hmm interesting! can't boot from it yet though.. at least it does support PPC [08:03:48] <Bartman007> dystopia: booting is (probably) a long ways off [08:03:50] *** tomocha6 has joined #opensolaris [08:04:04] <Bartman007> AFAIK freebsd can't boot from ZFS either. [08:04:33] <dystopia> true, i used a seperate SSD to have my FreeBSD installation on [08:05:56] *** div8 has joined #opensolaris [08:06:51] *** jbasse has joined #opensolaris [08:10:18] *** twisti has joined #opensolaris [08:11:12] *** not-me-guv has joined #opensolaris [08:11:27] *** skillet has joined #opensolaris [08:11:47] *** kgoetz has quit IRC [08:17:01] *** anathematic has quit IRC [08:19:08] *** axisys has quit IRC [08:20:19] *** dunc has quit IRC [08:22:50] *** anathematic has joined #OpenSolaris [08:24:20] *** tsoome has quit IRC [08:25:37] *** jbasse has quit IRC [08:26:07] *** sophokles has joined #opensolaris [08:26:30] *** sophokles has left #opensolaris [08:27:11] *** fr4g_ has quit IRC [08:29:37] *** yongsun has joined #opensolaris [08:30:31] *** uebayasi has joined #opensolaris [08:33:18] *** uebayasi is now known as uebayas1 [08:34:37] *** anathematic has quit IRC [08:34:48] *** sophokles1 has joined #opensolaris [08:35:51] *** Fullmoon has quit IRC [08:38:29] *** MrBIOS--_ has quit IRC [08:41:04] *** Fullmoon has joined #opensolaris [08:46:47] *** twisti is now known as twisti_home [08:46:59] *** tsoome has joined #opensolaris [08:48:17] *** GNUWorld has joined #opensolaris [08:49:30] <GNUWorld> Anybody have experience with build software using pkgbuild and SFE spec files? [08:49:50] <e^ipi> not if it means there's a follow-up question to that [08:50:20] <GNUWorld> Of course, it would be rather silly to ask otherwise :P [08:50:38] <e^ipi> it's pretty silly to ask in either case [08:51:04] <e^ipi> much more efficient to ask the actual question and wait for replies [08:51:07] <trochej> GNUWorld: I think that for sfe related questions there is an sfe channel :) [08:51:08] <GNUWorld> Regardless, I am having some issues and I am looking for some insight. [08:51:29] <GNUWorld> That would make sense, I apologize if I had posted this in the wrong room. [08:51:39] *** skillet has quit IRC [08:51:42] <trochej> GNUWorld: Never mind [08:51:45] *** kohju is now known as KOHJU [08:51:49] *** KOHJU is now known as kohju [08:51:51] <trochej> GNUWorld: Don't ask metaquestions, just ask you question. [08:52:01] <trochej> GNUWorld: If anyone knows the answer, they will tell you. :) [08:52:16] <GNUWorld> Okay, sounds like a fair point to make. [08:52:36] *** Fullmoon has quit IRC [08:54:29] *** sophokles1 has left #opensolaris [08:55:21] <spike-> Anybody use opensolaris + freenx (or any nx for that matter) [08:55:40] <GNUWorld> I have used NX in the past, not with OpenSolaris mind you. [08:55:58] *** MrBIOS--_ has joined #OpenSolaris [08:56:27] *** MrBIOS- has quit IRC [08:57:23] <cypromis> #pkgbuild [08:58:29] *** Fullmoon has joined #opensolaris [09:00:36] *** FallenHitokiri has joined #opensolaris [09:08:46] *** MeP3aBeu has joined #opensolaris [09:10:20] *** dunc_ has joined #opensolaris [09:24:14] *** CIA-25 has quit IRC [09:29:09] *** twisti_work is now known as twisti [09:34:50] *** zack has joined #opensolaris [09:35:01] *** Erwann has joined #opensolaris [09:40:14] *** Fullmoon has quit IRC [09:42:48] *** DTEIT has joined #opensolaris [09:42:53] <DTEIT> hi all [09:42:58] <DTEIT> one question... [09:43:15] <DTEIT> yesterday i tried to install opensolaris with zfs root [09:43:32] <e^ipi> opensolaris always installs with zfs root [09:43:40] <DTEIT> really? [09:43:44] <Macabee> yep [09:43:49] <DTEIT> the default was with ufs [09:43:52] <e^ipi> nevada / SXCE does not [09:44:00] <e^ipi> but "OpenSolaris" does [09:44:09] <e^ipi> with nevada, you have to go through the text installer [09:44:24] <DTEIT> uhm...then i'm in doubt if i have installed opensolaris or nevada :-) [09:44:32] *** jbasse has joined #opensolaris [09:44:42] <e^ipi> cat /etc/release [09:44:51] *** phimic has joined #opensolaris [09:45:02] <DTEIT> express [09:45:29] <e^ipi> okay, so if you used the default graphical installer, you got ufs [09:45:36] <e^ipi> if you use the text installer, you got a choice [09:45:48] <DTEIT> infact...i always use text installer [09:45:58] <DTEIT> but the problem is another one [09:46:16] <DTEIT> since i had a problem with 5 hd tryin to create a zpool [09:46:25] <DTEIT> i created a zpool on a single hd [09:46:40] <DTEIT> isn't possibile to make it at least a mirror? [09:46:47] <e^ipi> yes, man zpool [09:47:09] <e^ipi> the 'attach' subcommand [09:47:14] *** CIA-26 has joined #opensolaris [09:47:32] *** Fullmoon has joined #opensolaris [09:47:48] *** Trede has joined #opensolaris [09:48:42] *** McBofh has quit IRC [09:51:22] <DTEIT> ahh...found the problem...i was using c1t0d0 instead of c1t0d0s0 [09:51:39] <e^ipi> yeah, there's a whole process involved in doing it right [09:52:26] <e^ipi> http://malsserver.blogspot.com/2008/08/mirroring-resolved-correct-way.html [09:52:36] <e^ipi> with prtvtoc and fmthard goodness [09:53:02] <e^ipi> so you can boot when the first disk fails [09:53:18] <e^ipi> (note the intentional use of the word 'when' rather than 'if' ) [09:54:45] *** anilg has quit IRC [09:55:11] *** jbasse has quit IRC [09:55:15] <DTEIT> i read that after attach you need to installa grub on the second disk [09:55:21] <DTEIT> CR 6668666 [09:55:38] <e^ipi> yes [09:55:43] <e^ipi> see the link i /just/ posted [09:57:36] *** GNUWorld has quit IRC [09:57:57] <DTEIT> yeah...i did it without prtvtoc and the partition table are different [09:58:11] <e^ipi> eh, detach, try again [09:58:46] <DTEIT> yes...sure...i wanted just to understand...so an "attach" is not going to clone exactly the original disk [09:58:49] *** jbasse has joined #opensolaris [10:01:42] <e^ipi> no, it just attaches it to the pool [10:01:55] <e^ipi> all the partition setup stuff you have to do yourself [10:02:39] <Macabee> the attach clones the filesystem [10:02:40] <Macabee> not the disk [10:02:44] <Macabee> (is another way of putting it) [10:02:53] <DTEIT> ah ok [10:02:58] <e^ipi> one can be pedantic about it [10:03:06] <trochej> Coffee [10:03:08] <e^ipi> which is a good thing when it comes to writing system software [10:03:12] <trochej> People, drink coffee [10:03:17] <e^ipi> however, i believe meaning was conveyed sufficiently [10:03:24] <Macabee> trochej: i just poured my first one of the day [10:03:28] <DTEIT> i already drunk three of them trochej :-) [10:03:42] <DTEIT> and it's only 10:00 am [10:03:59] <e^ipi> i had a 1L beer stein for coffee a few years back [10:04:02] <e^ipi> but it broke :( [10:04:09] <Macabee> sucks [10:05:14] <Macabee> e^ipi: what do you work on, if you don't mind me asking [10:05:39] <e^ipi> i'm a transient force and my manager lends me out all over the place [10:05:53] <asyd> \_o< [10:06:03] *** Dar has joined #opensolaris [10:06:11] <e^ipi> comparisons between gnu/solaris userland tools, lately scm migration [10:06:14] <e^ipi> community putbacks [10:06:46] *** gerard13 has joined #opensolaris [10:13:40] *** sletz has joined #opensolaris [10:14:48] <sletz> Hi, i'm trying to boot OpenSolaris Live CD 2008.05 on a new Dell X420 machine (which is supposed to work according to HCL list) [10:15:13] <sletz> the boot stop with "Console login service(s) cannot run" [10:15:31] <sletz> what is that? what can i do? [10:17:18] <Macabee> if its a SATA DVD drive you might need to get a build above 95 to boot [10:18:20] *** cypromis has quit IRC [10:18:36] <sletz> is burning a Solaris Express Community Edition CD or DVD disk the way to go? [10:18:51] <sletz> from here http://opensolaris.org/os/downloads/? [10:19:19] <sletz> i undersatnd "Solaris Express Community Edition" is more uptodate yes? [10:19:27] <Macabee> its different [10:19:35] <Macabee> http://www.genunix.org/ has ISOs of the more recent OS2008.05 builds [10:19:36] <DTEIT> uhm....a little problem [10:19:47] <DTEIT> i detached the wrong device :-P [10:20:01] <DTEIT> now when i try to re-attach the device [10:20:06] <DTEIT> it says: [10:20:13] <sletz> Macabee: do i should better get a version at http://www.genunix.org/ yes? [10:20:14] <DTEIT> cannot attach c1t0d0s0 to c1t1d0s0: device is too small [10:20:20] *** noyb has joined #opensolaris [10:20:38] <Macabee> sletz: Solaris Express is something completely different [10:20:47] <sletz> Macabee: sorry lost newbie here... [10:20:49] *** xinkeT has quit IRC [10:20:59] <Macabee> sletz: get http://www.genunix.org/distributions/indiana/osol-0811-97.iso [10:21:02] <Macabee> and see if that boots [10:21:07] <sletz> Macabee: ok thanks [10:21:21] <Macabee> nps :) [10:22:29] *** xinkeT has joined #opensolaris [10:24:52] <sletz> Macabee: is global.iso better? what is "supports all language"? [10:26:46] <sickness> what I really can't understand is why the version with more languages takes actually more space than the version with fewer languages :P [10:27:14] <sickness> this should be something like the steve jobs reality distortin field :P [10:36:17] *** alka has joined #opensolaris [10:39:20] *** MattMan has joined #opensolaris [10:39:58] *** Fullmoon has quit IRC [10:47:09] *** kohju is now known as KOHJU [10:51:19] *** derchris has quit IRC [10:51:48] *** derchris has joined #opensolaris [10:55:32] *** simonleinen has joined #opensolaris [10:57:53] *** CIA-26 has quit IRC [11:03:09] *** gergap has joined #opensolaris [11:03:12] <gergap> hi [11:03:27] <gergap> does somebody know how to build qt4 on opensolaris? [11:03:52] <e^ipi> it built fine last time i tried [11:04:29] <gergap> I got XX command not found error but fixed that already [11:04:44] <gergap> this was becaue configure needs the #!/bin/bash to work [11:04:58] <gergap> but now I get CC command not found error somewhat later [11:05:06] <gergap> I still don't know why [11:05:38] <gergap> e^ipi: how did you compile it? [11:07:23] <gergap> e^ipi: did you use the commercial or open source version? maybe there is a difference in the build scipts. [11:07:45] <e^ipi> just the open source one [11:08:07] <e^ipi> i used the studio compilers last time i tried [11:08:08] <gergap> did you set CC CXX variables or did it work without that? [11:08:20] <e^ipi> i think i just had them first in $PATH [11:08:30] <gergap> that's not GNU C compiler? [11:08:41] *** LordIllpalazo has joined #opensolaris [11:09:27] <gergap> e^ipi: I'm new to Solaris, sorry. But what is this Studio exactly? Is this installed per default? [11:11:22] <sickness> isn't it the compiler? it's not usually installed per default [11:11:44] <gergap> ok, I've found sunstudio in the packet mananger and will try that instead of gcc. [11:12:36] <sickness> if you search on opensolaris.org there are various articles/blog posts about using sun studio instead of gcc (unfortunately it's a difficult task because a lot of software is full of gnuism this days...) [11:12:53] <e^ipi> not Qt [11:13:17] <e^ipi> they at least try to be cross platform most of the time [11:13:35] <gergap> sickness: I would expect that it's easier to build qt with gcc than with another compiler because gcc is also used on linux and that works without problems. [11:14:13] <e^ipi> gergap: you'd expect wrong, because the Qt folks expect solaris = studio [11:14:28] <gergap> ok [11:14:29] <e^ipi> probably for the best too, studio is a much better compiler than GCC [11:14:51] <gergap> what is better? [11:14:58] <e^ipi> studio [11:15:05] <e^ipi> code generation, much much better than gcc's [11:15:28] <gergap> yes, faster executbale, smaller, or just faster to build? [11:15:34] <e^ipi> yes, yes and yes [11:16:23] <gergap> does this still apply to current gcc versions? gcc improved a lot over the last years? [11:16:34] <e^ipi> not enough to make a difference [11:17:28] *** McBofh has joined #opensolaris [11:17:43] <e^ipi> afternoon McBofh [11:17:43] *** yongsun_ has joined #opensolaris [11:18:09] <gergap> peformance is important for my progs. I'm anxious to see my benchmark result comparing mit gcc linux app compared of the solaris version :-) [11:18:28] <e^ipi> does it do a lot of threading ? [11:18:50] <gergap> I configure my prog to run multithreaded or singlethreaded. [11:19:17] <e^ipi> if it does threading, and you have more than about 8 cores on the machine, solaris beats the tar out of linux [11:19:20] <gergap> It uses POSIX threads. Is that fair to compare? [11:19:24] <e^ipi> single threaded, linux often comes up on top [11:19:31] <e^ipi> in solaris it doesn't matter [11:19:45] <gergap> Does Solaris have native POSIX threads or is this just a wrapper? [11:20:20] <e^ipi> ? [11:20:23] <e^ipi> it has threads [11:20:51] <e^ipi> solaris threads, and posix threads... the differences between the mostly boil down to the arguments you call to thread_create/pthread_create and friends [11:21:06] <e^ipi> how you invoke threads is almost entirely irrelevant to the kernel [11:21:09] <gergap> an linux POSIX thread are implemented nativly in the kernel. Other OSs just wrapper the posix calls to their own thread implementation [11:21:33] <e^ipi> i... guess they're native then? [11:22:30] <gergap> I think its' not so important. This mainly influences the creation of threads. If a thread is already running it's no problem any more I htink. [11:23:22] <e^ipi> I think you're making assumptions based on the way linux does things that they will somehow correspond to the way solaris does things [11:23:29] <e^ipi> this is quite often not the case [11:26:43] *** jbasse has quit IRC [11:26:46] *** IvanR__ has joined #opensolaris [11:27:44] <oxygene> the linux implementation tends to allow stuff that's illegal or undefined in the standard (eg. locking a mutex, passing it to a thread, unlocking it there) [11:27:51] *** luc^ has joined #opensolaris [11:29:49] <e^ipi> oxygene: "linux" and "standard" are often disjoint sets [11:30:28] <e^ipi> best to assume that linux violates the standard everywhere, and be pleasantly surprised on the rare occasion it doesn't [11:31:13] <sickness> gergap: yeah that's the point... [11:33:44] <_mary_kate_> gergap: linux threads are equally as native as solaris threads. linux doesn't have a pthread_create() syscall, it has a low-level syscall that the pthread library wraps [11:33:59] <_mary_kate_> gergap: just like solaris doesn't have a pthread_create syscall, it wraps thr_create [11:34:04] <_mary_kate_> (or lwp_create, whichever) [11:34:28] <_mary_kate_> gergap: just because it's called 'native' posix threads library doesn't mean it's any more native than any other Unix [11:34:36] *** darmster has joined #opensolaris [11:34:41] <_mary_kate_> it's just actual kernel threading, not the horrid hack linux had before nptl [11:35:25] <gergap> yes, before it was totally different [11:35:25] <_mary_kate_> (so, compared to linux 2.4, it is native threading) [11:35:25] <gergap> now it just "wrapps" the syscall [11:35:36] <gergap> but e.g. on windows the POSIX wrapper is also horrible, because the native threads works so different [11:36:08] <_mary_kate_> windows is only one os, you make it sound like every OS except Linux works that way [11:36:17] <darmster> is there really a difference in opensolaris then debian [11:36:39] <gergap> _mary_kate_: I didn't expect that, I just was asking :-) [11:37:21] <gergap> I installed sunstudio now. Do I have to set any envirment variables now to be able to compile Qt? [11:37:31] <gergap> suncc some to be not in the PATH [11:37:50] *** IvanR_ has quit IRC [11:38:12] <gergap> Is export PATH=/opt/SunStudioExpress/bin:$PATH all that I need? [11:41:19] *** balbirs has joined #opensolaris [11:41:49] <balbirs> how can I check whether my HDD is SATA or IDE ? [11:42:39] *** domutaka has joined #opensolaris [11:43:58] <gergap> on linux the device nodes are named different hda->IDE sda->SATA/SCSI, has Solaris simalar naming conventions? [11:44:50] <domutaka> yes but they usually have one character in names to identify the device , bus and so on [11:44:53] <balbirs> solaris it is c#t#d#s# and for x86 architecture it is c#d#s#.. so not sure [11:45:42] *** alibb has joined #opensolaris [11:47:06] <gergap> e^ipi: with suncc it seems to work, it least Qt is still compiling :-) [11:47:36] <balbirs> I had ubuntu and my disk's naming convention was /dev/sda# now in solaris it is [11:47:43] <balbirs> lrwxrwxrwx 1 root root 51 Sep 17 13:56 /dev/dsk/c3d0s0 -> ../../devices/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0:a [11:47:43] <balbirs> lrwxrwxrwx 1 root root 51 Sep 17 13:56 /dev/dsk/c3d0s1 -> ../../devices/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0:b [11:47:43] <balbirs> lrwxrwxrwx 1 root root 51 Sep 17 13:56 /dev/dsk/c3d0s0 -> ../../devices/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0:a [11:47:43] <balbirs> lrwxrwxrwx 1 root root 51 Sep 17 13:56 /dev/dsk/c3d0s1 -> ../../devices/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0:b [11:47:47] *** CIA-26 has joined #opensolaris [11:48:10] <balbirs> so, here I see IDE as the controller.. [11:48:22] <balbirs> shouldn't it be SCSI or sd ? [11:48:26] <gergap> pci-ide, that's what wnated to know I think [11:48:47] <domutaka> no [11:48:55] <gergap> but then it should habe been hda on linux ... [11:49:11] <domutaka> yes [11:50:06] <gergap> I have a nother Solaris newbie question: Why are my POS1, END, DEL keys not working in the terminal? and how can I get it working? [11:50:38] <balbirs> but, I know my HDD is SATA and it was mounted as /dev/sda# in ubuntu.. it is X86 [11:52:39] <Macabee> that's because linux treats SATA as SCSI [11:53:14] *** FallenHi1okiri has joined #opensolaris [11:55:09] *** domutaka has quit IRC [11:55:35] *** domutaka has joined #opensolaris [11:55:44] <balbirs> so here, solaris is using IDE controller for my sata hdd, ? [11:57:22] <balbirs> this one is scsi ? dev_path=/pci@0,0/pci-ide@1f,2/ide@1/sd@0,0:h,raw [11:58:02] *** yongsun has quit IRC [11:58:05] *** domutaka has quit IRC [11:58:13] <_mary_kate_> pce-ide@1f,2/ide@1? yeah, that totally looks like scsi.. [11:58:33] *** domutaka has joined #opensolaris [11:58:34] <_mary_kate_> your sata controller is probably in ide-compatible mode [12:01:51] *** darmster has left #opensolaris [12:03:13] *** myrkraverk has joined #opensolaris [12:04:00] *** likaijun has quit IRC [12:05:11] <DTEIT> uhm....i'm following this steps to add a disk in a zpool [12:05:16] <DTEIT> http://malsserver.blogspot.com/2008/08/mirroring-resolved-correct-way.html [12:05:27] <DTEIT> but i get this [12:05:35] <DTEIT> fmthard: Partition 0 overlaps partition 2. Overlap is allowed only on partition on the full disk partition). [12:05:46] <DTEIT> the same for partition 8 [12:07:34] <balbirs> _mary_kate_: how you are interpreing that line ? how it looks scsi ? [12:07:55] <_mary_kate_> balbirs: i was being sarcastic [12:08:13] <balbirs> ok, :) [12:08:39] *** FallenHitokiri has quit IRC [12:12:11] *** Trede has quit IRC [12:13:34] *** alka has quit IRC [12:14:53] *** tsoome has quit IRC [12:17:08] *** Trede has joined #opensolaris [12:17:57] *** Trede has quit IRC [12:17:59] *** Trede has joined #opensolaris [12:21:09] <DTEIT> ok...did it....fdisk problems [12:22:57] *** jbasse has joined #opensolaris [12:23:39] *** jbasse has quit IRC [12:23:56] *** stux is now known as stux|work [12:24:00] *** Rarok has joined #opensolaris [12:27:13] *** Cookie` has joined #opensolaris [12:30:32] *** calumb has joined #opensolaris [12:30:44] *** myrkraverk has quit IRC [12:33:11] *** LordIllpalazoThe has joined #opensolaris [12:35:57] *** yongsun_ has quit IRC [12:37:32] *** netj has joined #opensolaris [12:38:55] *** evocallaghan1 has joined #opensolaris [12:39:30] *** LordIllpalazo has quit IRC [12:40:00] *** netj has quit IRC [12:40:38] <gergap> after a while of idleing the screen gets black for power saving. But if a press a key the screen doesn't come back. Is this a bug of my graphics driver, or do I need a special key combination to activate it again on Solaris? [12:42:12] <domutaka> solaris do have great problms with specially graphics. disable power saiving feature of screen and use your monitor power button instead [12:42:22] <gergap> ok [12:42:58] *** sergiusens has quit IRC [12:42:58] <gergap> I can log in via SSH. Can I restart X somehow? [12:43:22] *** jbasse has joined #opensolaris [12:44:45] <hile_> svcadm disable -t cde-login && svcamd enable cde-login [12:44:54] <gergap> thx [12:45:15] <hile_> (modulo typos of course -- i'm on bluetooth net acce) [12:46:47] *** kjackson has joined #opensolaris [12:46:50] <gergap> cde-login' doesn't match any instances [12:46:57] *** kjackson has quit IRC [12:46:58] <gergap> how can I list the services? [12:47:39] <hile_> svcs -a [12:47:53] <gergap> ok [12:48:21] <hile_> There's this wonderful "se also" in most manpages. It's quite useful [12:48:38] *** evocallaghan has quit IRC [12:49:01] <gergap> works [12:49:23] <gergap> it was just gdm instead of cde-login [12:50:01] *** mui has joined #opensolaris [12:50:04] <gergap> I didn't know this svcs thing. Services work quite different on Solaris. I'm used to use the scripts in /etc/init.d [12:50:05] <hile_> ugh [12:50:09] <balbirs> what these five file denotes /dev/dsk/c0d0p[0-4] ?? [12:50:11] <mui> hi, is there solution for irq16 being share by other driver [12:50:14] <mui> it causes random panics [12:50:25] <mui> shared even :P [12:50:35] <_mary_kate_> how did you discover that's the cause of the panics? [12:50:37] <mui> is it bios pnp related problem or what :P [12:51:41] <hile_> gergap: vanilla init scripts are still supported; however smf(5) is preferred. [12:51:53] *** sophokles has joined #opensolaris [12:52:13] <hile_> *sigh* do noe new administrators not read documentation anymore? [12:52:24] *** sophokles has left #opensolaris [12:52:39] <oxygene> hile_: when they come from linux, they're not used to documentation [12:52:46] <oxygene> hile_: at best, they find forum posts via google [12:53:31] <sletz> i am in the process of partitionning a disk to install Linux and Solaris; do i need to setup the future Solaris partition with a given filesystem? or do let it empty? [12:53:49] <oxygene> let it empty - it has to be a primary partition [12:54:00] *** tsoome has joined #opensolaris [12:54:17] <balbirs> are /dev/dsk/c0d0p[0-4] primary partition as fdisk[fix disk] ? [12:54:21] *** IvanR_ has joined #opensolaris [12:54:34] <sletz> primary partition: what does that mean? [12:55:15] *** IvanR__ has quit IRC [12:55:35] <oxygene> sletz: if you partition in linux, keep free one of hd[a-z][1-4] (and if you create hd[a-z]5 or later, it also allocates one in the 1-4 range to host those extended partitions [12:56:20] <balbirs> so solaris can't be installed in logical partition ? [12:56:26] <hile_> oxygene: Admittedly, I am in a bit of the same boat when it comes to screenOS vs IOS, but at least I put it out there as "I've configured foo; now i'm trying to do bar. in IOS land, i would do <cmd>; what is the equivalent syntax" [12:56:28] *** netj has joined #opensolaris [12:57:05] <hile_> Why would you want it to be? [12:57:32] <hile_> dual-boots are of dubious value anyway [12:57:52] <oxygene> balbirs: it wouldn't find its disklabel there [13:03:16] <domutaka> problem: my friend installed solaris 10 , x86 edition on vmware . ok . but after reboot. its struck at only grub prompt . what can be wrong [13:03:50] <oxygene> almost everything [13:04:27] <domutaka> tell me pointwise . no.1 is ? [13:04:44] <oxygene> it can't find /boot/grub/menu.lst for some reason [13:05:22] *** tsoome has quit IRC [13:05:23] *** tsoome1 has joined #opensolaris [13:05:25] *** stukag has joined #opensolaris [13:05:37] <domutaka> can something be done . on this: grub> [13:06:03] <domutaka> i am not expert in grub [13:06:46] <oxygene> grub has tab completion. devices start with "(" (eg. "(hd0,0)") [13:07:33] <oxygene> so try around configfile "(<tab>" - look for your device, complete, it will probably add a comma, so you're at something like "(hd0," press tab again, etc [13:07:42] <oxygene> erm.. "configfile (<tab>" [13:07:55] <oxygene> then look for where /boot/grub/menu.lst is hidden [13:08:20] *** yoursdai has quit IRC [13:10:20] <domutaka> thanks man . its enough for him. [13:12:09] <balbirs> oxygene: any idea on /dev/dsk/c0d0p[0-4] , what are the significance for these device, do we have these in solaris plateform also ? [13:13:01] <oxygene> balbirs: they're significant if you want to access msdos disks (eg. usb storage). on "pure-solaris" disks, you're more likely interesting in the *s[0-7] devices [13:13:21] *** InvaderNOP has joined #opensolaris [13:13:24] <oxygene> domutaka: hmm.. did he create a separate /boot slice? that's the only reason that occurs to me why there's a problem [13:14:18] <domutaka> he just done default settings during installation. [13:14:59] <domutaka> default was set by the os [13:15:11] <oxygene> weird, I just did such an install (sol10 on vmware), and it works. sorry to be of no help [13:15:33] <oxygene> try figuring out where that configfile is lying around, that's all I can suggest [13:16:30] <domutaka> mee too. done that install and all fine. which config file u r talking about. grub.conf ?? [13:17:22] <oxygene> menu.lst [13:17:35] <oxygene> in /boot/grub [13:18:02] <domutaka> ok. when i go to his home i will check out. but thank again. [13:19:15] <balbirs> ohh.. thanks oxygene , me trying to findout which file denotes the all four primary partion ? so suspected those file might be.. [13:19:28] <domutaka> "no OS is good until the person who teaches it is good." that's why things like MS exists [13:19:45] *** sletz has quit IRC [13:19:46] *** gergap has quit IRC [13:21:15] <balbirs> c3d0 denotes my SATA HDD, now I have fours primay partition over there.. how to see which four file denotes them ? [13:21:49] <oxygene> balbirs: do c3d0p[0-3] exist? [13:23:01] <balbirs> yes, it is five, c3d0p[0-4] [13:26:02] *** gergap has joined #opensolaris [13:26:15] *** Dar has quit IRC [13:26:49] *** kokoko1 has joined #opensolaris [13:27:00] *** domutaka has quit IRC [13:27:27] *** domutaka has joined #opensolaris [13:28:55] *** zack has quit IRC [13:29:31] <jmcp> p0 is the entire physical device [13:29:39] <jmcp> p[1234] are the primary "partitions" [13:30:49] <balbirs> oh, thank you :) [13:31:15] *** Rarok has quit IRC [13:31:51] <jmcp> balbirs: iirc, you also wanted to know whether OpenSolaris could be installed in a DOS-style Logical (aka Extended) Partition [13:31:58] <jmcp> the answer is no, but it's being worked on [13:32:19] *** stran has joined #opensolaris [13:33:35] *** Gekz has quit IRC [13:33:39] *** Rarok has joined #opensolaris [13:33:49] *** Rarok has joined #opensolaris [13:34:10] <balbirs> I guess it is because it also needs to hold the VTOC table, which can't be achieved if it is logical partition [13:35:21] <jmcp> I don't think that's the reason, but I don't know for sure [13:37:07] <oxygene> a logical partition is the same as a primary one, just with its table entry at a different position. but parsing the chained tables is not implemented yet [13:39:30] *** Gekz has joined #OpenSolaris [13:39:39] <gergap> I try to install Sun Studio 12 now instead of the express version. [13:40:08] <gergap> The batch_installer shows up an error: showrev not found [13:40:20] <gergap> what do I need to install for that? [13:40:25] *** tsoome1 has quit IRC [13:40:36] *** tsoome has joined #opensolaris [13:42:58] <TomJ> no showrev? that's pretty core [13:43:51] <jmcp> gergap: are you running OpenSolaris2008.05? [13:43:57] <gergap> yes [13:44:02] <jmcp> TomJ: there ya go [13:44:25] <jmcp> gergap: showrev is the command used to show patch revisions. os200805 doesn't have any patches [13:44:57] *** Fullmoon has joined #opensolaris [13:45:20] <TomJ> You'd think they'd include a stub so stuff like Studio wouldn't break [13:45:43] *** stukag has quit IRC [13:45:46] <gergap> how can I get showrev? [13:46:03] <TomJ> Only in hackish ways, like grabbing a version from Solaris 10, which probablh wouldn't run [13:46:07] <jmcp> I wouldn't worry [13:46:11] <TomJ> but there must be a way to install studio on opensol [13:46:15] <TomJ> cant you just ignore the error? [13:46:16] <jmcp> i doubt it would run properly on os200805 anyway [13:46:20] <oxygene> compatibility is so un-linuxy - no need to keep that in opensolaris [13:46:35] <jmcp> gergap: do you have to use the batch installer, or can you use the gui? [13:46:41] <jmcp> ... might just be easier [13:46:54] <gergap> the gui doesn work at all. all checkboxes are grayed [13:47:05] <gergap> batchinstaller breaks with this error [13:47:05] <jmcp> alternatively, have you tried "pkg install sunstudio" (or whatever the package name is) ? [13:47:12] <TomJ> have you googled for instructions on installing studio on opensolaris? [13:47:18] <TomJ> oh yeah maybe there's a pkg [13:47:34] <gergap> installed sunstudio before with the gui installer [13:47:42] <gergap> this seams to install the express version [13:47:49] <gergap> Qt didn't compile with that [13:48:43] <gergap> btw: is there any documentation which explains this pkg stuff? [13:49:00] <jmcp> there's a Getting Started guide which is on your os200805 desktop [13:49:03] <jmcp> and on the livecd [13:49:54] <gergap> ok [13:53:11] *** kokoko1 has quit IRC [13:53:15] <DTEIT> packages for solaris 10 are able to run on solaris sxce? [13:53:27] *** KOHJU is now known as kohju [13:53:55] <gergap> so /usr/bin/pkg is the default tool to install software und the GUI tool as a frontend to this. Or does this work different? [13:55:15] *** sletz has joined #opensolaris [13:57:09] <TomJ> DTEIT: most likely [14:02:56] *** wms has joined #opensolaris [14:05:53] <DTEIT> thanks [14:07:26] *** karrotx has quit IRC [14:08:31] *** stukag has joined #opensolaris [14:13:16] *** Odin- has joined #opensolaris [14:13:56] *** norman has quit IRC [14:16:22] *** JoergB has quit IRC [14:17:58] *** domutaka has quit IRC [14:18:35] *** balbirs has left #opensolaris [14:22:12] <DTEIT> i read that LD_LIBRARY_PATH should never be used...where can i set then other places for shared library? [14:23:33] <_mary_kate_> DTEIT: when compiling, link with -R/foo/lib:/bar/lib:/usr/openwin/lib:etc [14:23:46] <_mary_kate_> that will embed the path in the library [14:24:26] <DTEIT> ok...so just in the case of compiling [14:25:25] <DTEIT> is there a tool to see actual occupation of a cpu/core? [14:25:34] <_mary_kate_> mpstat [14:26:01] <DTEIT> wow...thanks [14:26:21] <_mary_kate_> (use it like vmstat, 'mpstat 5' or so) [14:31:02] *** calumb is now known as calLNCH [14:37:47] *** airjump has quit IRC [14:40:16] *** karrotx has joined #opensolaris [14:45:21] *** jbasse has quit IRC [14:47:21] *** JoergB has joined #opensolaris [14:55:33] *** tsoome has quit IRC [14:57:31] *** anilg has joined #opensolaris [14:57:32] *** timsf has joined #opensolaris [14:58:07] *** zenbalrog has joined #opensolaris [14:58:57] *** PicCard has joined #opensolaris [14:59:06] *** Openfree has joined #opensolaris [15:04:24] <sletz> for dual boot Linx/Solaris, the Solaris boot loader will erase Linux one yes? [15:04:59] <trochej> yes [15:06:10] <sletz> then Linux boot infor has to be set again later on yes? [15:06:28] <trochej> yes [15:08:40] *** Pioneer42 has joined #opensolaris [15:09:13] <sletz> and what about triple boot with Windows also? [15:09:31] *** Pioneer42 has quit IRC [15:10:45] *** _teo_ has joined #opensolaris [15:12:36] *** dnm has quit IRC [15:19:13] *** Alasdairrr has quit IRC [15:21:50] *** sgowda_ has joined #opensolaris [15:22:28] *** MrBIOS--_ has quit IRC [15:25:01] <TomJ> sletz: solaris grub should be installed, that can then "chain load" Windows, and either boot Linux direct or chain load linux grub [15:25:10] <TomJ> but solaris grub has to be first, because solaris can boot linux but linux can't boot solaris [15:26:36] <sletz> TomJ: "chain load" Windows: what does that mean? [15:27:10] <TomJ> it's a grub term, it means it doesnt boot a kernel directly, it just boots another boot loader [15:27:16] <trochej> sletz: That means that Soalris loader loads Windows loader and windows loader loads windows [15:27:49] <sletz> TomJ: would that mean; install Solaris first, then others? [15:28:04] <sletz> but Windows then Linux, or Linux then Windows? [15:28:15] <TomJ> actually it might be easier to install Windows, then Linux, then Solaris [15:28:23] <TomJ> because by default each will put its own boot loader over the disk [15:28:27] <TomJ> and you want Solaris' to be the one that's left [15:28:37] <TomJ> I dont know for sure though, I've never dual booted Solaris [15:29:01] <TomJ> you can tell Linux not to install its boot loader on install, I imagine, but it might just be eaiser doing a normal install then installing Solaris afterwards [15:29:10] <TomJ> I am sure there are pages about this, try some google [15:29:12] *** Rotarye1 has joined #OpenSolaris [15:29:28] <sgowda_> ls [15:29:53] <TomJ> donkey_porn.avi [15:31:21] *** yongsun has joined #opensolaris [15:31:38] *** yongsun has quit IRC [15:33:11] *** tsoome has joined #opensolaris [15:34:25] *** Rarok has joined #opensolaris [15:35:32] *** sgowda_ has quit IRC [15:38:25] *** suresh has joined #opensolaris [15:38:35] *** Rarok has quit IRC [15:38:41] *** Rarok has joined #opensolaris [15:39:09] *** shankar_ has joined #opensolaris [15:40:01] *** MrBIOS-- has quit IRC [15:40:04] *** MrBIOS-_ has quit IRC [15:40:51] *** MrBIOS_ has quit IRC [15:41:01] *** shankar_ has quit IRC [15:44:30] *** sergiusens has joined #opensolaris [15:48:09] *** sgowda_ has joined #opensolaris [15:48:40] <suresh> Do anybody know if this channel is logged (archived) somewhere ? [15:48:40] *** lkthomas-home has joined #opensolaris [15:49:10] <lkthomas-home> guys, I need to hear some suggestion about how to deal with zfs backup [15:51:03] <suresh> http://blogs.sun.com/timf/entry/zfs_backup [15:51:13] <suresh> Did you check that out ? [15:51:38] <lkthomas-home> I can't find it on google, thanks for pointing it out, reading [15:51:50] <h3sp4wn> lkthomas-home: For what type of usage - zfs-auto-snapshot I use at home [15:52:09] <lkthomas-home> we plan to put data to tape [15:55:43] *** calLNCH is now known as calumb [15:57:15] <h3sp4wn> lkthomas-home: take no notice of what I said then [15:57:33] <lkthomas-home> of course we also did auto snapshot :P [15:57:59] <holcomb> there was an article on bigadmin recently about using netbackup with snapshots [15:58:10] <h3sp4wn> But I guess with more care than I have [15:59:28] *** alka has joined #opensolaris [16:04:09] <lkthomas-home> LOL!!!!] [16:04:19] <lkthomas-home> I found something very interesting which will resolve my question [16:04:23] <lkthomas-home> zfs set readonly=on [16:05:22] *** kristian___ has joined #opensolaris [16:05:34] <kristian___> hi [16:06:05] *** kristian___ has left #opensolaris [16:06:59] *** xRaich[o]2x has joined #opensolaris [16:08:17] *** yongsun has joined #opensolaris [16:08:18] *** not-me-guv has quit IRC [16:08:21] *** yongsun has quit IRC [16:09:01] *** jbasse has joined #opensolaris [16:16:16] *** coffman_ is now known as coffman [16:18:23] <lkthomas-home> ok, I just read sun docs about zfs send receive, but small part confused me [16:19:21] <Macabee> ... [16:19:28] *** alka has quit IRC [16:19:31] <lkthomas-home> if my zfs got 10GB of data, and snapshot just token then I use zfs send data@justsnapshot > backup1, does it backup zero byte because I did not change anything ? [16:19:56] <TomJ> no it'll send the whole contents of the snapshot [16:20:02] <holcomb> if you send an incremental (send -i) with no changes, yeah, it's 0 (or probably just some metadata) [16:20:06] <TomJ> if you want incremental, you need zfs send -i snapshot1 snapshot2 [16:20:09] <TomJ> then it just sends the difference [16:20:20] <TomJ> which in this case would be near 0 byte (not exactly 0 bytes, there'd be headers and stuff) [16:22:01] <lkthomas-home> so, for space saving, I just need to do following step: take snapshot1, backup snapshot 1 without -i, then backup snapshot 2 by: zfs send -i snapshot1 snapshot2 > backupfile, then I could destory snapshot 1 ? [16:22:23] *** sletz has quit IRC [16:22:36] *** suresh has quit IRC [16:22:44] *** calumb has quit IRC [16:25:15] <lkthomas-home> nevermind, it seems works this way [16:25:37] *** Alasdairrr has joined #opensolaris [16:26:11] <lkthomas-home> another confused part, if I have gzip-9 enabled, when I am using zfs send to backupfile, I don't need to compress it anymore because zfs did it already, am I correct ? [16:27:05] *** sgowda_ has quit IRC [16:29:28] *** jbasse has quit IRC [16:31:37] *** Aria has joined #opensolaris [16:33:53] *** jbasse has joined #opensolaris [16:34:58] <asyd> heya jbasse [16:35:58] *** norman has joined #opensolaris [16:36:50] *** fr4g has joined #opensolaris [16:36:52] <stran> are there any solaris hosting options that are competitive with linux hosting? [16:39:41] <coffman> joyent.com [16:40:10] *** Rarok has quit IRC [16:40:28] *** sletz has joined #opensolaris [16:41:01] *** sletz has quit IRC [16:42:05] *** cypromis has joined #opensolaris [16:44:14] *** MeP3aBeu has quit IRC [16:48:32] *** Rarok has joined #opensolaris [16:49:32] *** div8 has quit IRC [16:53:47] *** jwit has quit IRC [16:55:12] <gergap> I try to port some lines of inline assembler from gcc to suncc, but I get the error "invalid register" when using placeholder "%o0,%o1,..." like described in the docs. Any suggestions? [16:55:34] *** chaataka has joined #opensolaris [16:58:49] *** gehmehgeh has joined #opensolaris [17:00:27] *** calumb has joined #opensolaris [17:01:27] *** yippi has joined #opensolaris [17:02:17] *** storycrafter has joined #opensolaris [17:03:32] <gergap> ah, this syntax seems to work only on SPARC [17:04:17] <jbasse> asyd: salut petit [17:04:24] <jbasse> asyd: sorry training [17:04:31] <asyd> oh:) [17:04:32] <asyd> about? [17:04:48] <jbasse> asyd: Positive Power and Influence >> /dev/null [17:05:08] <asyd> ph3ar [17:05:30] <sstallion_work> gergap: erm. first experience with asm ? [17:06:18] <_mary_kate_> gergap: %o0 is SPARC asm, so yes, it only works on sparc ;) [17:06:27] <_mary_kate_> gergap: i386 registers are named like %eax, %ebx.. [17:06:43] <sstallion_work> IA32 registers are ;) [17:06:47] <sstallion_work> x86 gets a bit ugly [17:06:57] <_mary_kate_> yeah, well there's %ax and whatever too [17:17:13] *** gerard13 has quit IRC [17:17:54] <gergap> _mary_kate_: I know x86 asm, but sparc is new. [17:18:18] <sstallion_work> gergap: are you familiar with RISC archs ? [17:18:27] <gergap> I was just irretated by the suncc samples [17:18:44] <gergap> sstallion_work: this principles are clear with RISC, but I'm not an expert [17:18:59] <gergap> passing params over registers, rotating registers, etc. [17:19:00] <sstallion_work> gergap: *sigh* read the asm documentation for SPARC [17:19:47] <gergap> at the moment I need x86, I was just not sure about the inline assembler syntax. [17:20:19] <sstallion_work> a quick hint: using inline asm is never portable; you would be better off rewriting that in C (assuming there are no real performance requirements that the optimizer cannot obtain) [17:20:45] <gergap> it's not a matter of performance, I'm not writing better code than the compiler [17:20:53] <gergap> but some things are only possible in asm [17:21:02] * sstallion_work blinks. [17:21:23] <e^ipi> like what exactly? [17:21:58] <gergap> evaluating a carry after add -> integer overlfow [17:22:18] <gergap> efficient atomic increments... [17:22:29] <sstallion_work> gergap: of which you should be using syslibs for [17:23:18] <gergap> a call to another function is much slower than lock; cmpxchg :-) [17:23:30] <oxygene> gergap: for atomic operations, there are library calls [17:23:39] <sstallion_work> gergap: not if they are inlined. [17:23:42] <oxygene> gergap: and in case of studio, magic to inline them [17:24:07] *** lkthomas-home has quit IRC [17:24:15] <gergap> oxygene: what's the name of this function? [17:24:20] <sstallion_work> gergap: its generally not a good idea to try and outsmart the compiler... you end up with brittle (and annoying) code [17:24:23] <e^ipi> gergap: studio is smarter than you. [17:24:58] <oxygene> gergap: see atomic_ops(3C) [17:25:18] <gergap> compare the output of suncc -S. I don't believe any studio can do that faster than with this two opcodes. [17:25:55] <oxygene> gergap: compare after linking (ie. decompile). at least on sparc, the linker is quite clever, too [17:26:10] <oxygene> not sure if they caught up on x86 already [17:26:39] <gergap> btw: man doesn't show me any C funtion on Solaris. Are there no man pages available? [17:26:50] <sstallion_work> gergap: oxygene gave you the reference [17:26:51] <e^ipi> docs.sun.com [17:26:54] <e^ipi> they're all there [17:27:01] *** Rarok has quit IRC [17:27:21] <e^ipi> (i'm assuming you're on 2008.05 which has a bunch of missing pages) [17:27:36] <sstallion_work> bleh [17:27:46] * sstallion_work wishes people would stop trying to port to Indiana [17:28:28] <twisti> sstallion_work, Because...? [17:28:31] *** _mary_kate_ has quit IRC [17:28:37] *** calumb has quit IRC [17:28:40] <sstallion_work> twisti: its not ready. As soon as it has parity with SXCE it will be ready [17:29:02] <twisti> But libraries will stay mostly the same, I suppose. [17:29:05] <oxygene> indiana is missing pages? fits in the "let's appease linuxers" motto - they wouldn't be able to cope with proper docs [17:29:23] <sstallion_work> twisti: assuming they arent closed, maybe. [17:29:40] <sstallion_work> if you port to solaris, port to solaris. Indiana isn't quite there yet. [17:30:52] <twisti> That's a point but as Sun is always very deliberate on being binary compatible, I guess it will work out. [17:31:17] <sstallion_work> twisti: you are missing my point. Not everything has made it into Indiana yet. There are quite a few closed bits floating around. [17:31:25] <oxygene> twisti: compare "API stability" notices in solaris 10 and solaris express manpages. [17:31:39] <twisti> sstallion_work, Sorry, missed that. [17:38:54] *** gehmehgeh has quit IRC [17:42:28] *** loke__ has quit IRC [17:44:38] *** RavenSlay3r has joined #opensolaris [17:45:25] *** sletz has joined #opensolaris [17:45:37] *** perlmongo has joined #opensolaris [17:45:53] *** _mary_kate_ has joined #opensolaris [17:46:19] *** twisti is now known as twisti_work [17:46:21] <sletz> I'm trying to install OpenSolaris form LiveCD , after boot Install Solaris doe not see any hard disk [17:46:26] *** loke has joined #opensolaris [17:46:31] <sletz> and pfexec format also not... [17:46:47] <sletz> the disk already contains a Linux system [17:49:12] *** apersson has left #opensolaris [17:49:52] *** crichardso has joined #opensolaris [17:49:57] <Alasdairrr> sletz: Then solaris hasn't found your SATA/IDE controller [17:49:59] *** comay has joined #opensolaris [17:50:13] <sletz> Alasdairrr: so what can i do? [17:50:24] <sletz> Alasdairrr: it a SAYA disk [17:50:24] *** ChanServ sets mode: +o comay [17:50:28] <sletz> SATA [17:50:45] *** stevel has joined #opensolaris [17:50:45] *** ChanServ sets mode: +o stevel [17:51:05] <Alasdairrr> Go into your motherboard's bios and set the SATA type to IDE instead of AHCI [17:51:21] <jbk> morning [17:52:23] <sstallion_work> jbk: yo [17:52:27] <sstallion_work> drying out yet ? [17:53:28] *** rab has joined #opensolaris [17:54:08] <jbk> yeah [17:54:31] *** jwit has joined #opensolaris [17:54:40] <jbk> the carpet by the balcony door got wet [17:55:06] <sstallion_work> balcony ? [17:55:25] <Alasdairrr> sletz, did that work? [17:55:36] *** calumb has joined #opensolaris [17:55:43] <jbk> yeah i'm in an apartment [17:55:48] <sstallion_work> ahh [17:56:37] <gergap> sstallion_work: I need to implement something like http://doc.trolltech.com/4.3/atomic-operations.html#1 [17:57:01] <gergap> atomic_cas looks like what I want, but doesn't return if the operation succeeded. [17:57:22] <gergap> I don't see an alternative. [17:58:31] <sstallion_work> atomic_cas looks fine to me [17:58:48] <jbk> soaked up 3 towels that night [17:59:05] <jbk> but just got power last night again [17:59:11] <jbk> went out around 2-3am sat morning [17:59:12] <sstallion_work> gergap: just don't be lazy... compare the returned value with the value you passed in [17:59:14] <jrsharp> is anyone running postfix on opensolaris? [17:59:40] <gergap> sstallion_work: that will not work, If I compare it afterwars another thread could have changed it. [17:59:49] <sletz> Alasdairrr: trying.. [17:59:51] <gergap> So I don't know if the operation work or not [17:59:59] <sstallion_work> the return value [18:00:13] <sletz> Alasdairrr: the disk was configurated with a "RAID on" stuff... [18:00:57] <gergap> It returns always the old value no matter of it was changed or not. how can this help? [18:01:02] <sletz> Alasdairrr: not seen anything else to change in the BIOS related to SATA disk [18:01:04] <sstallion_work> eh ? [18:01:13] *** ajmcello has joined #opensolaris [18:01:29] <sstallion_work> think about it [18:01:30] <ajmcello> i am getting Property exec for method inetd_start of instance svc:/network/rpc/gss:default is invalid at bootup time from inetd. anybody know i can fix it? [18:01:36] <jbk> sstallion_work: http://www.flickr.com/photos/jasonbking/sets/72157607337997731/ you can (or I guess can't) see i10 towards downtown [18:01:53] <Alasdairrr> sletz, turn the "RAID" option off [18:01:58] <Alasdairrr> unless you're using raid :) [18:01:59] <sstallion_work> jbk: oh nice [18:02:25] *** ahe has joined #opensolaris [18:02:28] <Alasdairrr> RAID/AHCI requires specialised drivers, if you turn that off, most modern motherboards have an "IDE Emulation" mode, so SATA drives can be made to appear as IDE drives [18:02:31] <gergap> sstallion_work: e.g. value=0; ret = atomic_cas_uint(&value, 0, 1); -> ret = 0; [18:02:36] <Alasdairrr> Which works with almost all operating systems [18:03:01] <sstallion_work> gergap: if (atomic_cas_uint(&value, 0, 1) != 1) { ... } [18:03:15] <Stric> and IDE emu is most often set as default.. resulting in way slow speed. [18:03:16] <Alasdairrr> The alternative is to try and work out what SATA chipset you have, find out if Solaris supports it, and somehow get Solaris to recognise it - if you're using the latest OpenSolaris release and it doesn't detect it, then there's a good chance it doesn't support your chipset [18:03:39] <gergap> sstallion_work: I've to rethink that. [18:03:44] <Alasdairrr> There's a hardware compatibility list (HCL) you can check [18:03:58] <_mary_kate_> gergap: atomic_cas returns the old value [18:04:10] <sletz> Alasdairrr: i did, i have a Dell XPS 420 that is reported to work [18:04:27] <gergap> sstallion_work: I can assume the if atomic_cas_uint returns the expected value, it was swapped, otherwise not, right? [18:04:33] <_mary_kate_> gergap: yes [18:04:38] <sstallion_work> gergap: yes. read the man page. [18:04:48] <gergap> I did [18:04:58] <_mary_kate_> gergap: store the old value, do the CAS, test if it returned the old value you expected - if yes, it succeeded [18:05:14] <gergap> ok [18:05:24] <_mary_kate_> do { old = *v; } while (atomic_cas(v, old, newval) != old); [18:05:26] <gergap> I just had a knot in my brain :-) [18:05:37] *** stran has quit IRC [18:05:47] <sletz> Alasdairrr: the DVD SATA drive was not recognized also, could both problem be related? [18:06:04] *** yippi has quit IRC [18:06:04] <gergap> _mary_kate_: excatly that's what wanno do :-) [18:06:13] <sstallion_work> sletz: ATAPI isnt implemented on many SATA controllers [18:06:18] <sstallion_work> I think AHCI is the only one so far [18:06:28] <sstallion_work> rather in many SATA drivers [18:07:29] <sletz> sstallion_work: and so? will it work? [18:07:51] *** jfndi_ has joined #opensolaris [18:08:05] *** storycrafter has quit IRC [18:08:12] <sstallion_work> sletz: on an ahci controller, yes. My Ultra 24 has an SATA DVD-rom with an AHCI controller. [18:08:26] <_mary_kate_> gergap: now wait until you have to port to an arch that doesn't have atomic CAS, that's fun ;-) [18:08:34] <_mary_kate_> (SPARCv7, PA-RISC..) [18:09:46] <gergap> my software has a portlayer for that. at the moment I'm implementing the Solaris one. The first time as you see... [18:10:16] <gergap> the good thing is 99% of it is identically with linux. [18:10:23] <gergap> Windows is completely different. [18:10:33] <_mary_kate_> i thought Windows had _InterlockExchange or something [18:10:55] <_mary_kate_> (x86 lock; cmpxchg is an atomic cas... although you don't see it used as much since x86 also has interlocked increment and suck) [18:11:03] <gergap> _mary_kate_: yes it has, but this function call is not as fast as using inline assembler. [18:11:07] <_mary_kate_> s/suck/such/ (although x86 has lots of suck too) [18:11:12] <_mary_kate_> gergap: the compiler can't inline it? [18:11:17] <gergap> that's wy I replace it with an own implementation [18:11:55] <gergap> _mary_kate_: it's not statically linked on windows. its a call to kernel32.dll [18:12:04] <_mary_kate_> it's a syscall? [18:12:04] *** Erwann has quit IRC [18:12:49] <gergap> The definition is not clear to me on windows. It's a windows API call. Don't know if it's really kernel code, but kernel32.dll sounds so :-) [18:14:04] <_mary_kate_> according to MSDN, the compiler's allowed to inline it [18:14:20] <_mary_kate_> which i assume VS does (http://msdn.microsoft.com/en-us/library/ms683590(VS.85).aspx) [18:14:22] <sletz> Alasdairrr: ok a disk is recognized now... now i'll check that my linux still works...before going on with Solaris [18:14:43] *** storycrafter has joined #opensolaris [18:14:45] <gergap> my performance measurement says something different :-) [18:14:47] <Alasdairrr> okie dokie sletz [18:15:53] <gergap> _mary_kate_: where it says that it#s inlined? [18:16:02] <gergap> the page also refers to kernel32.dll [18:16:11] <_mary_kate_> "This function is implemented using a compiler intrinsic where possible. For more information, see the header file and _InterlockedExchange." [18:16:17] *** edgy has joined #opensolaris [18:17:14] *** nitrile has quit IRC [18:17:28] *** jbasse has quit IRC [18:17:39] *** jafari has joined #opensolaris [18:17:43] <gergap> and when it's possible? [18:18:26] <gergap> maybe you need a special compiler switch for that [18:18:40] <gergap> the x86 asm code works always [18:18:42] <_mary_kate_> /O ? ;) [18:19:04] <gergap> btw. trolltech uses also asm for their atomic operations. [18:19:46] *** phimic has quit IRC [18:19:46] <oxygene> because they want their tree to build on a large set of operating systems, so it's easier for them to ship their own [18:19:52] <gergap> _mary_kate_: I found /Oi enables intrinsic operations [18:19:54] <sletz> Alasdairrr: linux still work, so reboot on LiveCD and try to install Solaris on the hard disk... [18:19:58] *** stevel has quit IRC [18:19:59] <_mary_kate_> yeah, i imagine they already had x86 asm, so they just reused it [18:20:06] <edgy> Hi, may some one explain this phrase to me: "It is NOT possible to perform a snapshot of a file system if the file system is in use by system accounting." [18:20:20] <sletz> Alasdairrr: on a separated partition i made with Linux [18:20:29] <_mary_kate_> edgy: i imagine, if you have acct(1m) enabled, you can't snapshot the filesystem [18:21:02] <edgy> _mary_kate_: why? isn't the point of snapshot is to backup an active filesystem? [18:21:21] <oxygene> edgy: might be that they haven't quite figured out yet how to balance accounting with snapshots [18:21:49] <_mary_kate_> acct is kind of old and crappy [18:22:09] <oxygene> "you have 100mb quota, 105mb of that in snapshots. good luck getting into a situation where you can write again" [18:22:31] <_mary_kate_> edgy: is this zfs or fssnap(1m)? [18:22:40] <edgy> _mary_kate_: fssnap [18:24:01] <edgy> _mary_kate_: it's an exam question: see http://pastebin.com/m1b3b940d [18:24:12] <_mary_kate_> you're still doing that exam? heh [18:24:39] <edgy> _mary_kate_: yes, I would sit for the exam after a couple of days so I am reviewing questions from the net ;) [18:24:59] <edgy> _mary_kate_: A and D is definetely wrong [18:25:02] <_mary_kate_> i'm not sure that's how it's meant to work [18:25:26] <gergap> _mary_kate_: I cannot find a info what MS compilers support this intrinsic. I need to support VS6 to current VS 2008. [18:26:05] <_mary_kate_> gergap: if you already have the asm, you might as well just use that [18:26:53] <edgy> _mary_kate_: I would try to enable acct and check fssnap myself ... [18:28:00] <gergap> _mary_kate_: yes I'll do that. The advantage would just be that would not need to care about different processor architectures. [18:28:17] *** vmlemon_ has joined #opensolaris [18:28:42] *** alibb has quit IRC [18:29:47] <gergap> _mary_kate_: back to atomic_case. The manpage doesn't says anything about supported architectures. Do you know where I can find that info or can I assume that it works on every arch where OpenSolaris itself works. [18:29:58] <_mary_kate_> gergap: it works on both of them (x86 and sparc) [18:30:07] <gergap> good [18:30:25] *** domutaka has joined #opensolaris [18:30:36] <gergap> thx a lot [18:30:57] *** Fullmoon has quit IRC [18:35:15] *** logic has joined #opensolaris [18:36:05] *** twisti_home has quit IRC [18:39:00] *** MrBIOS_ has joined #OpenSolaris [18:39:30] *** ahe has quit IRC [18:40:46] *** vmlemon_ has quit IRC [18:41:43] <edgy> _mary_kate_: how can I check whether acct is enabled, please? by checking whether command like who, users are working? [18:42:21] <ajmcello> does anybody know how to fix these problems with inetd and SMF? i keep getting Property exec for method inetd_start of instance svc:/network/rpc/gss:default is invalid and Invalid configuration for instance svc:/network/rpc/gss:default, placing in maintenance [18:42:26] <Stric> edgy: run 'acctadm' [18:42:31] <ajmcello> and bootup..here is the exact mesages: http://pastebin.com/m3f49ead4 [18:43:00] *** rab has quit IRC [18:45:07] <Alasdairrr> Oh it's LOSUG tonight [18:45:25] <Alasdairrr> In 15 minutes in fact [18:45:28] *** Zplay has joined #opensolaris [18:47:32] *** bowdengl has joined #opensolaris [18:48:41] *** rab has joined #opensolaris [18:48:44] <edgy> Stric: thanks a lot. I enabled it now using acctadm -e basic but don't know how to disable it yet ;) [18:48:58] <holcomb> acctadm -d i think [18:49:33] *** Dominic has joined #opensolaris [18:49:34] *** Fullmoon has joined #opensolaris [18:51:46] <edgy> holcomb: thanks a lot [18:52:33] *** balbirs has joined #opensolaris [18:53:40] *** Pietro_S has quit IRC [18:57:06] <Fullmoon> What is the way to map distant NFS user id 501 to a local user with another UID? (but same username) [18:57:25] <Fullmoon> idmap dump does not lists the user [18:57:57] *** DTEIT has quit IRC [18:58:46] <Fullmoon> It seems the client does only support NFSv3, so mapping is via UID [19:00:57] <Stric> I think the answer is "you don't" [19:01:58] <Fullmoon> SMB it is :) Thanks [19:04:16] *** domutaka has quit IRC [19:04:30] *** Molle has quit IRC [19:04:36] *** edgy has quit IRC [19:06:03] *** Fullmoon has quit IRC [19:06:13] *** Molle has joined #opensolaris [19:12:06] *** twisti has joined #opensolaris [19:13:24] *** vmlemon_ has joined #opensolaris [19:17:19] <twisti> About that atomic instruction discussion before, does Sun Studio inline the atomic calls? [19:17:23] <twisti> It does not seem so. [19:20:01] <twisti> The generated code is also very interesting: [19:20:02] <twisti> http://rafb.net/p/AgD83h14.html [19:21:15] *** MrBIOS- has joined #OpenSolaris [19:29:49] *** bowdengl_ has joined #opensolaris [19:29:58] *** fr4g has quit IRC [19:31:19] *** calumb has quit IRC [19:32:01] *** jfndi_ has quit IRC [19:32:13] *** dunc has joined #opensolaris [19:33:13] *** MattMan has quit IRC [19:33:57] *** cypromis has quit IRC [19:34:19] *** Rarok has joined #opensolaris [19:36:44] *** bowdengl has quit IRC [19:37:51] *** gergap has quit IRC [19:45:40] *** stevel has joined #opensolaris [19:45:40] *** ChanServ sets mode: +o stevel [19:47:02] *** Fullmoon has joined #opensolaris [19:47:09] *** __teo__ has joined #opensolaris [19:53:05] *** gassy_bugger has joined #opensolaris [19:53:22] <gassy_bugger> so hey is the sun javasystem webserver fster than tomcat? [19:54:44] <e^ipi> i think the newer versions are [19:58:25] <sstallion_work> certainly more featureful [19:59:13] <jbk> i didn't think tomcat was a webserver [19:59:29] <sstallion_work> servlet container [19:59:34] <gassy_bugger> It says 8x faster than apache2.0+tomcat [19:59:40] <gassy_bugger> but I cant find any numbers [19:59:44] <gassy_bugger> sounds solid tho [19:59:45] <sstallion_work> although it can be configured to serve static content [19:59:46] <e^ipi> gassy_bugger: you're talking about glassfish, right? [19:59:51] <e^ipi> ( SJS app server ) [19:59:56] <gassy_bugger> servlet 2.5 jsp 2.1 security virt hosts etcetcetc [19:59:57] <e^ipi> or ws7 [19:59:59] <gassy_bugger> no no [20:00:05] <e^ipi> which also can be a jsp server [20:00:19] <e^ipi> but not a full app server [20:00:38] <gassy_bugger> http://www.sun.com/software/products/web_srvr/index.xml [20:00:41] *** _teo_ has quit IRC [20:00:45] <gassy_bugger> yea 7 [20:01:05] <gassy_bugger> so ok about glassfish, is it a total pain to use or easy? [20:01:10] <gassy_bugger> is glassfish even faster? [20:01:51] <sstallion_work> gassy_bugger: that will not support servlet 2.5 [20:01:56] <sstallion_work> it requires jdk 1.5... [20:02:30] <sstallion_work> odd... it must be container only. the java ee 5 spec requires java se 5 [20:04:01] *** capaz has joined #opensolaris [20:04:39] *** anilg has quit IRC [20:05:49] <e^ipi> glassfish is faster than tomcat. [20:05:58] <e^ipi> and it is retarded easy to use [20:06:34] <h3sp4wn> gassy_bugger: zeus will most likely be the fastest [20:06:36] <e^ipi> a monkey can deploy to glassfish [20:06:57] <e^ipi> h3sp4wn: what? [20:07:16] <e^ipi> zeus is a swing library, glassfish and tomcat are app servers [20:07:36] <Stric> zeus is a web server too [20:07:41] <Stric> (non-java stuff) [20:08:54] <DerJoern> damn, my solaris 8 iso have vanishe [20:08:55] <DerJoern> d [20:09:47] <e^ipi> how silly of me, when we were talking about JavaEE, to not associate a non-javaee server with the conversation [20:11:21] <_mary_kate_> you could put zeus in front of glassfish, but i am not sure it'd be faster ;) [20:11:37] <_mary_kate_> glassfish already has a pretty decent http engine compared to tomcat [20:11:46] *** domutaka has joined #opensolaris [20:13:06] *** balbirs has quit IRC [20:14:22] *** CIA-26 has quit IRC [20:16:53] <h3sp4wn> Stric: Just interested in how things compare. Zeus with whatever jsp server they recommend [20:18:51] <_mary_kate_> i'd be more interested in ZXTM with glassfish [20:19:10] *** beckyb has joined #opensolaris [20:20:02] <gassy_bugger> so glassfish or javaserver7 are faster? [20:20:11] <gassy_bugger> javaserver 7 says its servlet 2.5 compliant? [20:20:13] *** __teo__ has quit IRC [20:20:15] <_mary_kate_> sjs web server you mean? well they do different things [20:20:26] <_mary_kate_> WS7 has a servlet container (only), glassfish is a full J2EE app server [20:20:45] <gassy_bugger> A high-performance, in-process Java container provides a Java EE 5 compliant implementation of Java Servlets 2.5, JSP 2.1 and more [20:20:49] <gassy_bugger> ok [20:21:00] <_mary_kate_> right, servlets are part of EE, but not all of it [20:21:01] <gassy_bugger> so glassfish is higer performance as well? [20:21:04] <gassy_bugger> k [20:21:08] <_mary_kate_> glassfish provides the rest of java ee as well [20:22:12] <gassy_bugger> hmmmmm [20:22:15] <gassy_bugger> sounds solid [20:22:23] <e^ipi> gassy_bugger: yes, especially the beta/alpha whatever versions [20:22:43] *** sergiusens has quit IRC [20:22:44] <_mary_kate_> if you want to deploy only a servlet, and you're putting a web server in front anyway (which is common), you might find WS7 is a little faster [20:22:48] <_mary_kate_> (but i haven't tested it) [20:23:26] <gassy_bugger> argg [20:23:34] <gassy_bugger> is ws7 > tomcat6? [20:23:55] <_mary_kate_> well if you're going to use a separate servlet container, just use glassfish [20:23:56] <gassy_bugger> btw is opensolaris > archlinux [20:24:02] <_mary_kate_> it's better than tomcat no matter what you're doing [20:24:05] <gassy_bugger> ok [20:24:15] <gassy_bugger> so if tomcat then glassfish [20:24:20] <e^ipi> gassy_bugger: almost certainly [20:24:24] <_mary_kate_> no, glassfish _instead of_ tomcat [20:24:32] <gassy_bugger> yes thats what i meant [20:24:42] <gassy_bugger> if I see tomcat rip it out and put glassfish [20:24:52] <gassy_bugger> I wonder if thell go for it here at work [20:25:01] <gassy_bugger> does glassfish bring a crapload of complixty? [20:25:03] <_mary_kate_> well obviously you need to test it with your application first [20:25:16] <_mary_kate_> just replacing software for no reason because some guy on IRC told you it was better is a good way to break your systems [20:25:19] <gassy_bugger> we have dozens of commodity boxes [20:25:19] *** sactodave has quit IRC [20:25:22] <gassy_bugger> running tomcat [20:25:35] <gassy_bugger> doh [20:25:59] <e^ipi> gassy_bugger: just try it [20:26:08] <e^ipi> take your setup, benchmark it [20:26:15] <e^ipi> then try it with glassfish [20:26:27] <e^ipi> and try your setup with solaris instead of linux [20:26:30] <e^ipi> then try it with glassfish [20:27:09] *** capaz has quit IRC [20:27:15] <e^ipi> especially on multi-core systems solaris shines, but don't just migrate because it seems like a cool thing to do [20:27:21] <e^ipi> test to make sure you gain something from it [20:28:26] *** pjfloyd has joined #opensolaris [20:30:51] *** beckyb has quit IRC [20:30:57] <holcomb> it is a cool thing to do though [20:31:56] *** Fullmoon has quit IRC [20:33:33] *** bigjohnto has joined #opensolaris [20:33:34] *** logic has quit IRC [20:34:18] *** bahumbug has joined #opensolaris [20:36:29] *** sletz has quit IRC [20:43:13] *** sletz has joined #opensolaris [20:45:43] *** MrBIOS- has quit IRC [20:45:48] *** niq has joined #opensolaris [20:54:41] *** CIA-25 has joined #opensolaris [20:54:50] *** jgracin has joined #opensolaris [20:56:30] *** bhall has quit IRC [20:56:33] *** bigjohnto has quit IRC [20:57:44] <DerJoern> dammit. was solaris 8 2/04 update 8? [20:58:33] *** bhall has joined #opensolaris [20:59:12] *** dunc has quit IRC [21:01:56] *** MrBIOS_ has quit IRC [21:03:59] *** timsf has quit IRC [21:05:26] <gassy_bugger> is the freebs ule scheduler makign freebsd faster than opensolaris on multicore? [21:05:39] *** dunc has joined #opensolaris [21:06:39] <sstallion_work> gassy_bugger: is the sky bluer in canada ? [21:06:56] *** sah-work has joined #opensolaris [21:07:54] <smtms> sstallion_work, it is [21:08:01] *** AxeZ has joined #opensolaris [21:08:05] <oxygene> freebsd surely crashes faster (for me, ymmv).. so if mean time between crashes is any indication, ... ;) [21:08:10] <sstallion_work> smtms: tough to know if you don't try and find out for yourself ;) [21:08:50] <e^ipi> iirc freebsd7 was faster on a handful of cores but it hits a wall much sooner [21:08:59] <e^ipi> but i could be way off on that one [21:09:06] *** domutaka has quit IRC [21:09:15] <davidX-> freebsd crashes? [21:09:20] <e^ipi> plenty [21:09:25] <davidX-> sure [21:09:49] <davidX-> i've never had that problem [21:09:53] *** domutaka has joined #opensolaris [21:11:37] <e^ipi> and the plural of 'anectdote' isn't 'evidence' [21:11:40] <e^ipi> what's your point? [21:11:46] <oxygene> davidX-: given the right set of hardware, you get any operating system to crash. and given the right set of hardware, you can skew any benchmark to do your bidding [21:14:19] <sletz> i am trying to add a Linux oart in my Solaris grub file (menu.txt) [21:14:40] <sletz> i get "(Error 17 cannot mount selected partition)" when starting on Linux [21:15:11] <sletz> i installed Linux without changing the previous Solaris grub [21:15:24] <sletz> then foud th UUID of the Linuc partition to boot on [21:15:35] <ajmcello> does anybody know how to fix these problems with inetd and SMF? i keep getting Property exec for method inetd_start of instance svc:/network/rpc/gss:default is invalid and Invalid configuration for instance svc:/network/rpc/gss:default, placing in maintenance [21:15:37] <sletz> but not working... [21:15:42] <ajmcello> at bootup..here is the exact mesages: http://pastebin.com/m3f49ead4 [21:15:44] <oxygene> sletz: what kind of esoteric linux filesystem are you using? [21:15:52] <sletz> ext3 [21:16:28] <oxygene> sletz: hmm.. simple fix would be to get into linux (with your boot cd) and install grub (or lilo or whatever) into the linux bootblock (not the MBR!) [21:17:02] <oxygene> sletz: and then add a chainload (YourDisk,Partition)+1 entry to solaris' grub config [21:17:14] <oxygene> sletz: that way, no matter what you do on linux, the solaris grub doesn't need to know [21:17:16] <sletz> how that? [21:17:23] <gassy_bugger> hm [21:17:31] <gassy_bugger> never heard of freebsd crashing [21:17:48] <Stric> now you have ;) [21:17:50] <gassy_bugger> ok so solaris seems to be more vetted when it comes to larege numbers of cores? like 8+ [21:18:06] <sletz> oxygene: how to add "chainload (YourDisk,Partition)+1".. [21:18:09] <oxygene> gassy_bugger: that likely depends on the workload [21:18:48] <oxygene> sletz: "title Linux" "chainload (hd4,6)+1" (replace hd4,6 with your linux partition). those two lines somewhere in solaris' /boot/grub/menu.lst should be enough to add a new entry that runs the bootsector of your linux partition [21:20:38] *** alanc_away is now known as alanc [21:21:02] *** t_[^^]z has joined #opensolaris [21:21:02] *** crichardso has quit IRC [21:21:06] <gassy_bugger> is the opensolaris developer edition beocming more stable-----doing a global update on that caused me some fits [21:21:08] <oxygene> gassy_bugger: there might be workloads where freebsd is faster on 16 cores, there might be workloads where solaris is faster on 2 cores. [21:21:18] <gassy_bugger> ok [21:21:33] <gassy_bugger> so why move to solaris from archlinux or freebsd? [21:21:39] <sletz> oxygene: LiveCD does not find a configurated bool loader... [21:21:49] <gassy_bugger> Im a linux admin with soem freebsd exp [21:21:58] <gassy_bugger> I can get a box up with xorg and icewm window mgr [21:22:03] <gassy_bugger> I can run apache or taomcat [21:22:08] <gassy_bugger> compile stuff [21:22:11] <gassy_bugger> write absh [21:22:13] <gassy_bugger> bash [21:22:20] <gassy_bugger> so Im thinking hmm opensolaris [21:22:26] <gassy_bugger> maybe there is something here [21:23:54] *** Rarok has quit IRC [21:24:16] <oxygene> gassy_bugger: I can tell you why I prefer not doing linux anymore: it annoyed me that they replaced some part of the system with some new part (marginally better, with new exciting bugs and regressions, always incompatible) all the time. solaris tends to be more stable in that regard. [21:24:41] *** sergiusens has joined #opensolaris [21:24:48] <Alasdairrr> All unices are pretty similar, almost all applications run on any unix varient. It's basically comes down to needs and personal preference. [21:25:11] <gassy_bugger> Ive gotten good enuf that linux or freebsd are kinda liek old gloves. [21:25:35] <benley> comfy and nicely worn in [21:25:36] <benley> yet, [21:25:38] <gassy_bugger> Is it pretty easy to setup networking tricks like bonding two nics in opensolaris? [21:25:44] <benley> the fingers have holes in them [21:25:51] <Aria> Some. [21:25:53] *** wms has quit IRC [21:25:56] <gassy_bugger> so they work as one and give me faster throughoput? [21:26:13] *** chaataka has quit IRC [21:26:21] <Alasdairrr> 802.3ad [21:26:22] <gassy_bugger> is diskless stuff available in opensolaris? [21:27:00] <gassy_bugger> although have all servers in a farm dependant on nfs mounted root dir is kinda scary [+dhcp] [21:27:13] <turtle> THE YEAR WAS 802.3AD! [21:27:22] *** sletz has quit IRC [21:27:32] <gassy_bugger> the nice bit is having 1 image they all boot from so no config deltas [21:27:42] <Aria> ad. Not AD. It's a spec, not informational. [21:27:44] <gassy_bugger> I cant wait for the wirelss datacenter [21:27:47] <gassy_bugger> no more wiring bs [21:28:34] <oxygene> what else, microwaves? [21:28:37] <gassy_bugger> ati radeon supported ok in opensolaris? [21:28:42] <Alasdairrr> I'm not sure if you can bond solaris nic's together for greater throughput, but you certainly can do active/standby, I forget what it's called in solaris [21:28:44] <oxygene> wires have the nice advantage of isolating communication between two points [21:28:55] <gassy_bugger> hm [21:29:10] <gassy_bugger> maybe I ll throw dev edition on this box [21:29:11] *** mib_ioykrz has joined #opensolaris [21:29:14] <gassy_bugger> for fun [21:29:15] <gassy_bugger> ;) [21:29:25] *** beckyb has joined #opensolaris [21:29:39] <gassy_bugger> crap i wonder how I exprt my thunderbird [21:29:56] <Alasdairrr> OpenSolaris's package management is much better [21:30:04] <mib_ioykrz> my opensolaris box stopped responding for requests from remote subnets, where should I start troubleshooting? [21:30:13] <Alasdairrr> (than Solaris 10) [21:30:15] <oxygene> Alasdairrr: that's up for debate [21:30:16] <bda> gassy_bugger: See dladm(1M), re: link aggregation (bonding). [21:30:31] <bda> mib_ioykrz: Your netmask? [21:30:53] *** noyb has quit IRC [21:30:53] <mib_ioykrz> doesn't seem to be a routing issue because other machines work fine [21:31:25] <bda> Use snoop(1M) to see if the traffic gets to the box? [21:31:42] <Alasdairrr> Also "arp -an" and "netstat -rn" are always handy [21:32:19] *** cypromis has joined #opensolaris [21:32:32] <gassy_bugger> ah ha!! [21:32:35] <Alasdairrr> Why don't you like IPS oxygene? [21:33:03] *** comay has quit IRC [21:33:18] <bda> Alasdairrr: It has had some problems. It's also a very basic change from how packages and patching are done now. [21:33:21] <oxygene> Alasdairrr: I dislike magic that affects my system configuration [21:33:30] <gassy_bugger> ok [21:33:37] <gassy_bugger> im about to reboot into solaris express [21:33:42] <gassy_bugger> is it pretty easy to partition [21:33:47] <gassy_bugger> nothign crazy liek freebsd? [21:33:48] <Dominic> gassy_bugger: Thunderbird/Firefox both have profile directories (probably .thunderbird or .mozilla). Just copy that over IME. [21:34:00] <gassy_bugger> ok [21:34:11] <gassy_bugger> someone kick em down the knowhow on partitioning [21:34:12] <oxygene> Alasdairrr: and "oh, let's just download all kind of stuff and update it behind your back, just because you requested some minor tool" belongs in my "magic" category [21:34:16] <gassy_bugger> is it clikk ngo? [21:34:29] <gassy_bugger> cuz i dont have 2nd pc for thsi adventurelol [21:34:36] <gassy_bugger> it suports dhcp eh? [21:36:14] <oxygene> Alasdairrr: in such a case, I prefer the install of that minor tool to fail, with some notice that stuff is updated. it annoyed me on linux (and broke several debian installs over the year) and is one of the worse imports of linuxisms to opensolaris [21:36:18] <oxygene> (imho) [21:36:19] <bda> oxygene: At least you don't hate it just because it works over the network. ;) [21:36:26] * Stric gets a headache frmo al teh misspellings nlol omgwtfbbq cuz u r etc.. [21:37:54] <oxygene> bda: the lack of a formalized file format feels weird, too. but "tarballs with a certain file layout inside" seems good enough for now [21:38:41] <bda> Heh. [21:38:48] *** gassy_bugger has quit IRC [21:39:12] <oxygene> and that seems to be possible by jumping through some hoops [21:39:14] <mib_ioykrz> here is the relevant portion from snoop: http://mibbit.com/pb/Odp7Ja [21:39:42] <mib_ioykrz> thunderhead is not responding to pings from craig [21:43:56] *** codestr0m has quit IRC [21:44:15] *** fr4g has joined #opensolaris [21:44:29] *** codestr0m has joined #opensolaris [21:44:40] <mib_ioykrz> this is a better (filtered) paste of the snoop results: http://dpaste.com/78718/ [21:45:40] <mib_ioykrz> it looks like it is replying but then making a failed DNS request. any thoughts? [21:47:20] <oxygene> mib_ioykrz: and where is it not responding to pings? it looks okay [21:49:03] <oxygene> bda: the lack of postinst scripts makes some things hard (eg. update of gnu info directory files). and I still don't see the rationale for developing something new, instead of advancing the existing system. well... [21:49:35] <mib_ioykrz> craig (and all the other machines on it's network) don't get the responses [21:49:55] <oxygene> that discussion never left SWAN (if it took place at all), and now it's too late in any case *shrug* [21:50:04] <mib_ioykrz> ...but I can successfully ping from craig's subnet to any other machine on thunderhead's subnet [21:51:54] *** lisppaste3 has quit IRC [21:52:13] *** lisppaste3 has joined #opensolaris [21:52:54] <oxygene> mib_ioykrz: subnets.. what's the distance (in network terms) between those boxes? [21:53:46] <mib_ioykrz> 7ms ping times [21:54:03] <mib_ioykrz> 1 hop [21:54:50] *** vmlemon_ has quit IRC [21:54:53] <oxygene> so one machine inbetween? what does that one see? [21:54:54] *** vmlemon_ has joined #opensolaris [21:55:33] <mib_ioykrz> they are connected w/ a T1 and cisco 1700's on each end [21:55:55] <bda> Are both of the Ciscos configured to forward ICMP? [21:56:15] <bda> If A sends ICMP to B and B seeds and responds but A never sees the response... [21:56:21] *** CIA-25 has quit IRC [21:56:22] <mib_ioykrz> ICMP works for every other machine but the solaris box [21:56:27] <bda> Did you enable ipf on B and enable ICMP filtering? [21:57:24] <mib_ioykrz> i haven't done anything afaik [21:57:32] <mib_ioykrz> stock opensolaris nv72 [21:57:52] <mib_ioykrz> I did turn off network magic, but it has been working for a while since then [21:58:28] *** nitrile has joined #opensolaris [21:58:41] <bda> Just to quibble, the first OpenSolaris was b86. You're presumably running Solaris Express. Longshot, but check for networkig bugs in b72? [21:58:48] *** Mdx4 has quit IRC [21:59:28] *** capaz has joined #opensolaris [22:00:03] <mib_ioykrz> ok [22:00:14] <mib_ioykrz> fwiw, it is not just ICMP [22:00:30] <mib_ioykrz> ssh doesn't work either [22:00:30] <bda> So it answers no extra-subnet requests, but answers intra-subnet requests just fine? [22:00:37] *** PicCard has quit IRC [22:00:37] <mib_ioykrz> correct [22:00:50] <bda> That really sounds like netmask, or the Ciscos Doing Stuff. [22:01:46] <mib_ioykrz> going back through our logs, this all started after rebooting our solaris box [22:01:50] *** ninjaslim has joined #opensolaris [22:02:26] <mib_ioykrz> which makes me thing the ciscos aren't an issue (that and all other computers seem to be working fine) [22:02:56] <mib_ioykrz> A is 192.168.2.0/24, B is 192.168.1.0/24 [22:03:06] <bda> It could be a network configuration that wasn't committed. [22:03:19] <bda> hm. [22:03:27] <mib_ioykrz> this is the ifconfig on B: inet 192.168.1.245 netmask ffffff00 broadcast 192.168.1.255 [22:03:30] <bda> When did /etc/inet/hosts and /etc/inet/ipnodes get merged? [22:04:13] <mib_ioykrz> I do have an interface on B 10.10.1.1/24 [22:04:43] <bda> mib_ioykrz: nopaste /etc/inet/hosts, /etc/inet/ipnodes, /etc/hostname.if, /etc/defaultrouter, /etc/inet/netmasks. [22:04:53] <bda> Also, netstat -rn. [22:06:00] <mib_ioykrz> defaultrouter is empty [22:06:26] <bda> So you have no default route set on reboot. [22:06:31] <bda> (unless you get it from DHCP?) [22:07:14] <mib_ioykrz> nope [22:07:32] <mib_ioykrz> sounds like this is the issue... [22:07:39] <mib_ioykrz> what do I do? [22:08:32] <mib_ioykrz> this looks good: http://www.cyberciti.biz/tips/howto-solaris-how-do-i-setup-a-static-route-ie-static-router-ip-address.html [22:08:43] <bda> echo <DEFAULTROUTER> > /etc/defaultrouter [22:09:00] <bda> Where <DEFAULTROUTER> is 192.168.1.1 or whatever your router is. [22:09:18] <bda> Then reboot and ensure it's the default route via netstat -rn. [22:09:26] <bda> If you can't reboot but just need to fix it now, add it via route. [22:09:49] *** Rarok has joined #opensolaris [22:09:55] <bda> Yeah, that's what that page says. :) [22:10:37] *** stevel has quit IRC [22:11:11] *** stevel has joined #opensolaris [22:11:11] *** ChanServ sets mode: +o stevel [22:11:24] <mib_ioykrz> working! thx so much for your help bda [22:11:42] <mib_ioykrz> I assume this was the first reboot since pulling network magic [22:12:48] *** beckyb has quit IRC [22:13:17] <mib_ioykrz> so... how scared should i be of upgrading to Indiana? I'm using the machine as a NAS (ZFS over iSCSI and NFS) [22:13:44] <mib_ioykrz> sounds like the recommended path is ZFS export, then wipe OS and install? [22:14:13] <bda> Your zpools are in their own partitions or on dedicated disk? [22:14:15] *** RElling1 has quit IRC [22:14:25] <bda> But yeah, that's the way to do it. [22:14:45] <mib_ioykrz> 1 zpool containing an 8 disk raid-z [22:14:52] <bda> I just helped a friend move a tb NAS box over to OpenSolaris from FreeBSD. Migrating the data was a pain, but everything else was easy. [22:15:25] <mib_ioykrz> is ZFS import/export totally reliable? [22:15:28] <bda> (he was using ZFS on FreeBSD, and it was Weird) [22:15:38] <bda> I've never had any issues with it. [22:15:59] <bda> A lot of my systems have mirrored SATA disks, UFS on p1, zpool on p2. [22:16:12] <bda> Whenever I upgrade/reinstall them, I export the pool, blow away p1, and then import the pool back. [22:16:27] <bda> (which is great, because a lot of the system config is on the zpool, so there's basically no work, except recreating the zones) [22:20:28] <mib_ioykrz> yeah, that is similar to what we have (hardware RAID-1 for OS, JBOD RAID-Z for zpool) [22:21:00] <bda> I use SVM, but same idea. [22:21:45] *** pjfloyd has left #opensolaris [22:25:45] *** yarihm has joined #opensolaris [22:25:52] *** _teo_ has joined #opensolaris [22:26:20] *** _setuid_H has joined #opensolaris [22:26:35] <_setuid_H> Evening all [22:27:18] *** karrotx has quit IRC [22:27:37] *** lolmac has joined #opensolaris [22:27:47] *** chaataka_ has joined #opensolaris [22:28:13] *** yippi has joined #opensolaris [22:28:23] <_setuid_H> How many percent of opensolaris users does use indiana? I looked on project observers and saw a few friends that were against indiana :-) [22:28:37] *** beckyb has joined #opensolaris [22:32:42] *** dunc has quit IRC [22:33:09] *** fr4g has quit IRC [22:34:20] <t_[^^]z> I'm trying to get ssh to be able to log in without a password, I've made the public key and added it to the server's authorized_keys yet I am still asked for a password [22:34:22] *** _setuid_H has quit IRC [22:35:28] *** dunc has joined #opensolaris [22:36:47] <bda> t_[^^]z: Edit /etc/ssh/sshd_config, PermitRootLogins. [22:36:48] <bda> svacdm restart ssh [22:37:08] <t_[^^]z> thanks [22:37:58] <t_[^^]z> there's already PermitRootLogin (no s) and it's set to yes [22:40:00] *** _setuid_H has joined #opensolaris [22:40:29] <_setuid_H> Hi could someone help me with my ultrabay harddrive? [22:41:35] <_setuid_H> Looks like it speak with some dialect unknown to my system :-) [22:41:52] <_setuid_H> Output throws just read errors [22:42:01] *** lolmac has quit IRC [22:42:04] <_setuid_H> but disc is fine it worked ok underl inux [22:42:06] <_setuid_H> linux [22:43:18] *** CIA-25 has joined #opensolaris [22:44:22] *** _setuid_H has quit IRC [22:45:36] <bda> t_[^^]z: Check syslog for errors. Check perms on ~/.ssh. [22:46:00] <bda> t_[^^]z: Also, I misread your question; RootLogin is only needed if you want to log in as root. Normal uses don't need that option (if you care). [22:49:26] *** anathematic has joined #OpenSolaris [22:49:26] *** xRaich[o]2x has left #opensolaris [22:50:30] *** Trede has quit IRC [22:51:48] *** derchris has quit IRC [22:51:53] *** derchris has joined #opensolaris [22:52:25] *** slava has joined #opensolaris [22:52:36] <slava> hi guys, is there something like bsd ports for solaris? [22:52:50] <slava> i need to install a few tools (gcc, git, ...) and sunfreeware is a bit of a pain [22:53:59] <cypromis> pkgtool.sf.net./ [22:54:01] <cypromis> sorry [22:54:40] *** jgracin has quit IRC [22:54:48] <bda> slava: pkgsrc [22:54:48] <cypromis> http://pkgbuild.sourceforge.net/spec-files-extra/ [22:54:53] <cypromis> would be the right one [22:54:55] <cypromis> or pkgsrc [22:55:26] <slava> ok thanks [22:55:41] *** erast has joined #opensolaris [22:56:15] *** AxeZ has quit IRC [22:56:30] <alanc> gcc is built into S10 & later, just hidden under /usr/sfw/bin [22:56:47] <alanc> git isn't there yet (coming soon to opensolaris/sxce) [22:56:56] <SplasPood> it builds very easily tho [22:57:08] <bda> git without X bits is kind of annoying to build. [22:57:13] <alanc> (where there == "built into solaris" - it's in spec-files-extra) [22:57:27] <alanc> X bits without git is very annoying to work on [22:57:27] <SplasPood> bda: hrm, I don't think I had a problem, but maybe the systems I built on happened to have it all... [22:57:34] <bda> Or was last time I did it (quite a few months ago). [22:57:53] <bda> SplasPood: I may be misremembering, but iirc it was kind of tedious. Not hard. [22:59:06] *** jrsharp has quit IRC [22:59:21] <oxygene> yay, just got the OK for my bachelor thesis, which involves a migration of a bunch of vmware+linux instances to solaris+zones [22:59:30] <bda> oxygene: What's the thesis? [23:00:04] *** ahe has joined #opensolaris [23:01:31] <SplasPood> bda: I generally give up if its more than a couple deps and I don't recall doing so :) [23:01:45] <bda> Co-worker wanted it. shrug. [23:01:50] <bda> None of the prod boxes have X bits. [23:03:01] <oxygene> bda: design and implementation of a more scalable experimentation platform for the "IT security" course. ie. this migration, the rationale (besides "we had a sparc t2+ box lying around"), porting code, porting firewall setup, .. [23:04:24] <bda> oxygene: Interesting. Though in an IT security course, wouldn't you want to use LDOMs rather than zones? [23:05:53] <Alasdairrr> What other useful 3rd party package management things are there for Solaris 10? [23:06:30] <bda> pkgbuild. pkgsrc. openpkg. blastwave. sfe. sfw. [23:06:31] <oxygene> bda: it's only part of the course, and it consists of a small intrusion detection scenario (within $HOME), some ip firewall setup, squid (for http access control). I figured zones would be good enough for that, and we had scalability problems before (too many students) [23:06:32] <bda> I think that's all? [23:06:42] <bda> oxygene: Oh, that's easy, then. [23:06:55] <bda> oxygene: And you can just clone the template when rebuilding the zone set for a class. [23:07:09] <oxygene> bda: that's the idea :) [23:07:17] <bda> Cool stuff. [23:07:33] <oxygene> there's also pmpkg (my project). though I didn't update it for a year (until two days ago) [23:07:39] <oxygene> (re: packaging) [23:10:04] <Alasdairrr> I see [23:10:34] <Alasdairrr> *copies down the list* [23:10:39] <Alasdairrr> I'll have to investigate those. [23:10:41] <sponix> Think with Sun doing VirtualBox now, they could at least come up with one audio driver within it that Solaris supports ? *Grin* [23:10:54] <sponix> Think they should purchase OSS next ! [23:11:20] *** hspaans has joined #opensolaris [23:11:31] <oxygene> sponix: is it for sale? [23:11:42] <hspaans> g'day all [23:12:05] <sponix> oxygene: I have my doubts, but maybe I should research that and put that comment in the drop box ;) [23:12:44] <sponix> Really would be a huge step toward Desktop Adoption imho [23:14:14] *** stukag has quit IRC [23:14:58] *** smartboyathome has joined #opensolaris [23:15:53] *** pksu_ has joined #opensolaris [23:17:50] *** snejk has joined #opensolaris [23:18:37] <snejk> dclarke: there ? [23:20:46] <smartboyathome> I'm using OpenSolaris 2008.11 pre, and am trying to compile the latest Murrine SVN from SVN, and everything goes fine until I get to line 22004, when it says: syntax error at line 22004: `(' unexpected. What is on line 22004 and beyond is in this pastebin. http://pastebin.com/pastebin.php?dl=m4f35e8d7 [23:21:08] <smartboyathome> sorry, wrong link. correct one: http://pastebin.com/m4f35e8d7 [23:22:39] *** yippi has quit IRC [23:23:31] *** vmlemon_ has quit IRC [23:26:19] *** RElling has joined #opensolaris [23:26:24] <Zplay> smartboyathome: Where did you download the pre release? I can't find it [23:26:48] <sponix> Zplay: genunix.org [23:27:26] <Zplay> thank you sponix [23:28:08] <sponix> Zplay: no, thank you... very seldom someone asks a question about Solaris simple enough for me to answer ;) [23:29:41] <hspaans> sponix: if you want I have question for you about Solaris, cluster, containers and SPS ;-) [23:31:29] *** Rarok has quit IRC [23:37:56] <sponix> hspaans: well, I at least know what the first 2 are used for, the 3rd I've never heard of at all afaik [23:38:02] <sponix> what is a SPS ? [23:39:00] *** sergiusens has quit IRC [23:40:48] <hspaans> sponix: http://www.sun.com/software/products/service_provisioning/ [23:42:36] <hspaans> sponix: and sorry to say, its a perfect way to create work and chaos [23:43:36] <ninjaslim> is there an easier way to do a minimal installation of SXCE with just X11 and the necessities [23:47:08] *** piwi has joined #opensolaris [23:49:14] *** ahe has quit IRC [23:49:55] <Alasdairrr> there doesn't seem to be a way to do a minimal install of SXCE [23:50:04] *** domutaka has left #opensolaris [23:51:21] *** chaataka_ has quit IRC [23:51:26] <sponix> sure there is, install it on a box, strip it down, and then tarBall it up, to extract it on another box ;) [23:52:16] <ninjaslim> sponix: well no i mean during installation, it's realy tedious having to go through all the packages and dependencies [23:53:27] *** anathematic has quit IRC [23:55:12] <jmcp> Alasdairrr: there are documented minimisation procedures at sun.com/blueprints [23:55:15] <jmcp> look for "JASS" [23:55:45] <Alasdairrr> Oh cool, thanks for the tip jmcp [23:56:05] <jbk> hey jmcp [23:56:09] <jmcp> yw [23:56:11] <jmcp> gday gday [23:56:40] <snejk> jmcp: look for JET [23:56:48] <snejk> dunno if its maintained tho [23:56:57] <jmcp> snejk: that's not necessarily the same thing as minimisation [23:57:01] <jmcp> and yes, I believe it is maintained [23:57:11] <snejk> nop, it isnt, but it supports JASS [23:57:15] <snejk> during install [23:57:20] <sponix> jmcp: think you could drop a hint to the right people to acquire opensound.com OSS ? would make a nice addition, then Solaris inside of VirtualBox could have a working default sound driver *Grin* [23:58:02] *** fgd__ is now known as fgd [23:58:06] <jmcp> sponix: talk to gdamore some time [23:58:29] <sponix> jmcp: gdamore, haven't spoke with him prior, is that his lane ? [23:58:38] <gdamore> gdamore == Garrett D'Amore. [23:58:51] <jmcp> geez, that was quick! [23:59:11] <gdamore> heh. [23:59:34] <gdamore> on a call... buzz me in about 10 mins.