August 13, 2007 
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31
August 13, 2007 [00:00:29] <CIA-17> ksadhukh: 6574102 Need to add extended family/model/stepping info to cpuid_pass1() for Intel processors [00:00:47] <Shinden> hm [00:01:03] <Shinden> chattr dosn't work on solaris ? [00:02:08] <jmcp> Shinden: I don't have a manpage entry for it, so perhaps it doesn't [00:02:13] <jmcp> what does iit do? [00:02:34] <dlg> change attributes [00:02:42] <dlg> like chmod but lower level in the fs [00:02:47] <jmcp> ah, right [00:03:02] <jmcp> dlg: the first comment goes with "duh" :) [00:03:20] <nachox> isnt that an ext# linux command? [00:03:21] <tomww> getfacl/setfacl ? [00:03:31] <richlowe> different kinds of attributes. [00:03:32] <Shinden> yes it's ;] [00:03:35] <richlowe> nachox: and no, the BSDs have it, too [00:03:59] <richlowe> or maybe I'm just crazy [00:04:04] <jmcp> richlowe: maybe? [00:04:13] <nachox> i think so, it was actually improperly ported from bsd [00:04:23] <richlowe> jmcp: I meant in specific regard to this. [00:04:32] <richlowe> jmcp: "crazy" is lacking appropriate granularity [00:04:48] <Shinden> y chattr works on linux and *bsd [00:04:49] <Shinden> :} [00:04:56] <jmcp> richlowe: heheh :) [00:05:01] <nachox> chattr is used to set the inmutable and append only bits among others iirc, could be wrong though [00:05:10] <Shinden> y [00:05:11] <jmcp> Shinden: is it a command or a function? [00:05:40] <Shinden> http://www.linuxdevcenter.com/linux/cmd/cmd.csp?path=c/chattr [00:05:56] <jmcp> I assume that means it's a comment [00:05:58] <jmcp> command [00:06:20] <nachox> well, i dont think solaris has those things [00:06:31] <Shinden> chattr +i file make's inmutable to the file [00:06:32] <Shinden> :D [00:06:33] <nachox> never actually heard of them in solaris [00:06:39] <gejza> Do I have chance of mounting an reiserfs partition into my sxce? [00:06:54] <jmcp> gejza: no [00:07:01] <gejza> :( [00:07:11] <jmcp> gejza: unless you nfs export it from another system, and mount it via nfs on your SXCE system [00:07:53] <gejza> ok,what other filesystems are supported in sxce? [00:08:09] <jmcp> Shinden: the chattr command seems to me to be a way of seriously screwing with the underlying filesystem. For a userlevel command I have significant doubts about its safety [00:08:27] <jmcp> gejza: ufs, zfs, lofs (loopback), hsfs (iso9660), udfs [00:08:38] <jmcp> and vxfs ($$) and afs [00:09:05] <gejza> ok,so no linux FS at all :( [00:09:08] <dlg> qfs [00:09:25] <jmcp> I think somebody has ported ext2fs but it's definitely not an integrated feature of solaris [00:09:35] <jmcp> dlg: yup, that one too [00:09:42] <tsoome> pcfs [00:09:44] <jmcp> and pxfs if you're running Solaris Cluster [00:09:50] <jmcp> tsoome: we don't like to talk about that ... [00:09:51] <jmcp> :) [00:09:51] <dlg> jmcp: i assume thats cos there's some form of quality control on code going into solaris [00:09:54] <nachox> jmcp, also ntfs with some tools [00:10:29] <richlowe> dlg: you'd hope. [00:10:45] <richlowe> and pcfs [00:10:53] <richlowe> damn, tsoome got that one. [00:10:55] <tsoome> gejza: you should use more open solutions than linux:P [00:11:11] <gejza> ok thank you guys for this info :) [00:11:11] <coffman> there is pcfs... [00:11:15] <dlg> richlowe: its one of its best features [00:11:28] <richlowe> coffman: and udfs, which they both should deal happily with. [00:11:52] <richlowe> dlg: keep saying that, loudly and clearly. :) [00:12:13] <dlg> richlowe: i do all the time. i read a bit of code, its pretty obvious [00:12:25] <jmcp> dlg: yeah, quite possibly :) [00:12:28] *** tsoome1 has joined #opensolaris [00:12:47] <jmcp> dlg: you've been known to write a bit of code too, from time to time [00:12:57] <dlg> yeah, but it hasnt made it in [00:13:05] <jmcp> dlg: I know :( [00:13:12] <dlg> i hope thats more cos of politics than quality though [00:13:15] <coffman> there is experimental ext2 http://mount-ntfs.sourceforge.net/ [00:13:26] <jmcp> dlg: that's my take on the situation [00:13:52] <coffman> but hey, he uses reiser, he does not care about his data anyways :P [00:13:57] *** migi has joined #opensolaris [00:14:07] <gejza> ? :) [00:14:24] <gejza> what's wrong with reiserfs? [00:14:29] <g4lt-mordant> coffman, ...or his lif, apparntly ;P [00:14:55] <jmcp> gejza: as I understand it, there have been data integrity issues with it [00:15:15] <nachox> gejza, didnt you know? the problem is that it is in linux and we hate linux ;) [00:15:21] <gejza> lol [00:15:24] <gejza> :) [00:15:31] *** tsoome_ has joined #opensolaris [00:15:34] <g4lt-mordant> more than a few. I lost one of my toyboxes ten times in a row due to corruption [00:16:06] *** tsoome has quit IRC [00:16:48] <richlowe> dlg: if it was quality, the process would have got further. [00:17:04] <nachox> g4lt-mordant, if it was the same one i suspect hard drive failure, do not blame the fs [00:17:07] <coffman> gejza: dude, storing data on reiser is like sleeping with a street hooker (without a condom) [00:17:38] <jmcp> coffman: ouch! [00:18:19] <g4lt-mordant> nachox, it was a toybox, eevery time I had extra crap it went inside, including one full-disk migration [00:20:24] *** kFuQ has joined #opensolaris [00:23:08] *** dunc has quit IRC [00:23:52] *** swa has joined #opensolaris [00:25:51] *** yarihm_ has quit IRC [00:25:54] <coffman> im just trying to follow the spirit of indiana [00:26:12] *** ichigo has joined #opensolaris [00:28:21] <coffman> making the transition smooth for the linux people (speaking in picture language, etc) [00:29:39] * coffman hides [00:29:42] <coffman> n8 [00:29:48] *** coffman is now known as coffman_zzz [00:31:15] *** migi has quit IRC [00:31:32] *** tsoome1 has quit IRC [00:32:45] *** bunker has quit IRC [00:32:56] <nachox> coward, he insults indiana and then leaves to sleep... [00:32:57] <nachox> :P [00:42:36] <Shinden> we wanna apt-get on opensolaris [00:42:37] <Shinden> ;] [00:42:43] <jmcp> I don't [00:42:46] <jmcp> I *really* don't [00:42:52] <Shinden> [; [00:43:09] <Shinden> nextenta works good with apt-get [; [00:43:31] <jmcp> I want to be able to net-install from $local-netinstall-server, but for patched stuff - no f!@$!@ng way do I want to just depend on apt-get [00:44:12] <Shinden> u may use both things [; [00:45:23] <gejza> what's up with the changing 3423523files in order to change my hostname... [00:45:40] <jmcp> gejza: what's up with not rtfm on how to change your hostname? [00:46:09] <gejza> I was rtfm,but it's quite complicated [00:46:16] <bda> You mean three? [00:46:16] <gejza> to only change the stupid hostname [00:46:32] <jmcp> gejza: did you go to docs.sun.com and try reading the Solaris System Administration Guide? [00:46:56] <gejza> no, I'm only familiarizing with the sytem [00:46:58] <gejza> system [00:47:05] <bda> /etc/nodename, /etc/hosts, /etc/hostname.if. [00:47:09] <gejza> but I'll look into it,thanks [00:47:19] <bda> /etc/inet/ipnodes if you're not actually on ONNV (where it's a symlink to /etc/hosts). [00:47:36] <bda> Hope you had a backup of those other 3423520 files. :) [00:48:26] <bda> (or rather /etc/inet/hosts, I suppose, but whatever) [00:49:06] *** pg_user has joined #opensolaris [00:49:45] <pg_user> SFTP, is it like ftp or more like ssh? [00:50:03] <bda> gejza: The admin guides are worth reading. They'll familiarize you far better with the system than getting frustrated and deciding it sucks (it doesn't). [00:50:15] <gejza> :) [00:50:16] <jmcp> gejza: the admin guides are *very* well written [00:50:23] <pg_user> in other words, if I run an SFTP will my ISP notice like if it were ftp? [00:50:37] <jmcp> pg_user: more like ssh [00:50:40] <bda> It took me about a week to get through them. It was very much worth it. [00:50:55] <bda> sftp is just an ssh hack, like scp. [00:51:05] <pg_user> jmcp: will SFTP allow users more rights to my machine then? [00:51:14] <jmcp> pg_user: why would it? [00:51:25] <jmcp> perhaps you should rtfm sftp(1) [00:51:26] <jbk> so anyone a guru with the sparc instruction set? :) [00:51:45] <pg_user> jmcp: sftp man page isn't good for beginners. [00:51:54] <jmcp> pg_user: so keep re-reading it [00:52:00] <gejza> bda & jmcp: which one would you recommend me? there's "no System Administration Guide" document [00:52:07] <jmcp> pg_user: no, sftp won't allow users any more rights to your machine [00:52:10] <jmcp> gejza: I'll get you a url [00:52:11] <pg_user> ok [00:52:11] <jbk> it appears from everything i can read, something involved with compiling the kernel is generating a few (I found 2) invalid instructions [00:52:22] <jmcp> jbk: using studio12 ? [00:52:23] <pg_user> thank you jmcp [00:52:24] <gejza> jmcp: I appreciate it [00:52:55] <jbk> jmcp: whatever was used to build b63 on sparc [00:53:04] <jbk> what's strange is dis disassembles them [00:53:22] <jbk> but the 2005 ultrasparc docs explicitly say it should generate an invalid instruction trap [00:53:23] <pg_user> it seems like all the cool things are happening on solaris now [00:53:37] <jbk> in the one case, in the other, it's highly implied [00:53:54] <jmcp> gejza: s10 system admin guide http://docs.sun.com/app/docs/coll/47.16, SX system admin guide http://docs.sun.com/app/docs/coll/47.23 [00:54:04] <jmcp> pg_user: I certainly agree with you there :) [00:54:18] <jmcp> jbk: so this isn't a build you've done yourself, but a downloaded one? [00:54:35] <gejza> jcmp: thank you [00:54:42] <jbk> i believe so -- it's whatever dclarke used to install this box -- i'm guessing it was a sxce release [00:55:22] <jmcp> jbk: example? [00:55:28] <pg_user> so what's Bill Joy doing these days? [00:55:33] <jmcp> pg_user: good question [00:55:36] <jmcp> VC stuff, I think [00:55:50] <pg_user> i wonder if he's tried solaris 10. [00:56:02] <jbk> the two instructions are 0x20726574 and 0x20526563 [00:56:03] <gejza> jmcp: I'm using SXCE so I guess I should read the 2nd one? [00:56:09] <jmcp> gejza: yes [00:56:11] <gejza> ok [00:56:21] <jbk> they use invalid integer condition code registers [00:56:27] <jmcp> jbk: sorry, that doesn't actually help me. can you give me a snippet from dis please [00:56:29] <jbk> i'm gonna try to find the location in a sec.. [00:56:32] *** pg_user has quit IRC [00:57:52] <coffman_zzz> jbk: cat /etc/release [00:58:10] <coffman_zzz> nachox: there is nothing to insult, its hot air [00:58:40] <coffman_zzz> and one day, there will be code, from one to the other day [00:58:56] <coffman_zzz> when the gates are open [00:59:07] <coffman_zzz> that kind of stuff stinks [00:59:20] *** NikolaVeber has quit IRC [01:00:12] <jbk> jmcp: http://pastebin.ca/654951 [01:00:33] <coffman_zzz> sure you can make your own solaris, you only need this and this and this and this (...) in binary" "oh sure, you cant redistribute that stuff" [01:00:53] <coffman_zzz> narf [01:01:00] <coffman_zzz> n8 [01:02:03] <jmcp> jbk: I think that's worth an email to mdb-discuss - with as much context and explanatory notes as you can muster [01:02:36] <jbk> ok.. [01:02:39] <Shinden> go to sleep folks [01:02:43] <jbk> i know it confused me :) [01:02:51] <jmcp> Shinden: why? it's only just gone 9am here [01:02:58] <Shinden> :D [01:03:22] <Shinden> in europe is 01:07 [; [01:03:46] <Shinden> 12h clock sux ;p [01:04:47] *** noyb has quit IRC [01:05:07] <Shinden> bye [01:06:11] <jmcp> ciao [01:09:48] <reflect> did you read Google report on disk failures? http://209.85.163.132/papers/disk_failures.pdf and a commentary on for those who doesn't want to read it all > http://storagemojo.com/?p=378 [01:10:46] <jmcp> jbk: it could be that you're looking at the space between the functions, not sure that you should be bothering, really [01:10:47] <reflect> wow, really crappy english there [01:10:56] <jmcp> jbk: unless you've got something which jumped there .... [01:11:02] <jmcp> reflect: that paper is really old news [01:11:18] <reflect> jmcp: yeah, but it's new news for those who haven't read it. [01:11:21] <jmcp> true, true [01:11:52] <reflect> sure opened my eyes [01:12:30] <jmcp> "all hardware sucks, and all software sucks" [01:12:49] <reflect> if you're negative about it, sure [01:13:14] <jmcp> uh ... not all hw and sw has equal degrees of suckfulness [01:13:15] <jmcp> fortunately [01:13:55] <reflect> more like, all manufacturers gives you the wrong MTBF numbers, temperature isn't as bad as you thought, software can only predict some two thirds of the errors we encounter.. [01:14:03] <jmcp> exactly [01:14:27] <jmcp> so all you can really say is that you should expect failures, and expect to replace kit when it fails [01:14:29] * jmcp shrugs [01:15:36] <reflect> or, you could make plans.. raise the temperature in your storage cabinets.. and start thinking about replacing drives after 3-4 years [01:15:55] <reflect> if you think the report is insignificant, sure - think that. [01:16:05] <jmcp> no, I don't think it's insignificant [01:16:29] <reflect> you sure make it sound like it. [01:16:42] <jmcp> I've got years of cynicism about disk reliability reports under my belt [01:17:16] <jmcp> what is great about the google paper is that they've got a statistically significant number of disks to look at [01:17:21] <jmcp> most customers don't [01:18:03] <reflect> also gives the people of SMART something to think about [01:18:06] <jbk> well part of the problem i'm having is dis craps out if it hits an illegal instruction [01:18:10] <jmcp> what I love about that paper is that its conclusions challenge people's assumptions about MTBF, firmware smarts etc [01:18:21] <jmcp> jbk: oh no, no dissing dis! :) [01:18:34] <jbk> but the closed libdisasm.so seems to disassemble stuff it shouldn't [01:18:40] *** LeftWing has quit IRC [01:18:45] <jbk> so it's adding complexity to my testing [01:19:00] <jmcp> jbk: are you running dis on genunix? [01:19:24] <jbk> not yet, so var on /usr/bin, and just tried it on unix [01:19:34] <jmcp> ok [01:19:37] <jbk> i'm gonna try to clean up the spacing a bit so it matches better for branches [01:19:38] *** dmarker has quit IRC [01:19:48] *** dmarker has joined #opensolaris [01:19:55] <jbk> to try to cut down on the stuff it spits out as different [01:20:13] <jmcp> ok [01:20:24] <jmcp> oh, you're working on the sparc disassembler? [01:20:28] <jbk> yes [01:20:34] *** noyb has joined #opensolaris [01:20:38] <jbk> i'm pretty close [01:20:39] <jmcp> light dawns .... [01:20:48] <jmcp> jbk: I'm very keen to see you succeed [01:21:03] <jbk> and really at this point just trying to make sure the output's comparable [01:21:38] <jbk> i figure if /usr/bin, /usr/sbin, and the kernel + modules all pass, that's probably a pretty good sign [01:22:11] <jmcp> agree [01:22:47] *** tomww_ has joined #opensolaris [01:23:11] *** jwit has quit IRC [01:23:14] *** Stric_ has joined #opensolaris [01:23:20] *** emergo_ has joined #opensolaris [01:24:03] <jbk> so far it's been pretty good.. the ones that didn't work were the ones i suspected they wouldn't (because the instruction's are so overloaded), but that's been fixed, now it's realy just through trial and error trying to match the output where there's some room for variance (I'm making it a setting that's stored in the disassembler handle, so you can turn it on and off) [01:24:15] *** tomww has quit IRC [01:24:22] *** WickedWicky has quit IRC [01:24:24] *** emergo has quit IRC [01:24:28] *** Stric has quit IRC [01:24:37] <jbk> i'm hoping to get some space here soon so i can put the code up for others to look at, comment, laugh, etc. :) [01:25:08] <reflect> btw, I have a 3ware controller here that I'm willing to donate to whoever can make a driver for it [01:25:40] <jmcp> reflect: there's no existing driver? [01:25:44] * jmcp looks @ dlg ..... [01:26:06] <reflect> none that I know of.. [01:26:11] <jmcp> jbk: I'll blog it as soon as you publicise it [01:26:52] <gejza> wtf,i've just followed the wiki howto change hostname but no success: http://www.genunix.org/wiki/index.php/Change_hostname_HOWTO [01:26:53] *** LeftWing has joined #OpenSolaris [01:27:13] *** jwit has joined #opensolaris [01:28:21] <jmcp> reflect: interesting - I googled for 3ware, solaris and driver, and found hits from amcc for jni stuff! [01:28:55] <dlg> jmcp: there's no 3ware driver for solaris [01:29:14] <dlg> it would be easy to write though [01:29:17] <dlg> for io anyway [01:29:21] <dlg> on the older ones [01:29:23] <jmcp> nod [01:29:26] <jmcp> reflect: what model? [01:29:59] <reflect> 8500 I believe [01:30:04] <reflect> 8 port SATA controller [01:30:13] <jmcp> can you identify any of the chips on it? [01:30:16] <reflect> PCI-X [01:30:26] <reflect> sure, let me get it.. wait [01:30:36] <dlg> i think that would work on the old driver interface [01:30:59] <rbrown_> Windows XP EOL data 01/31/08 [01:31:02] <rbrown_> that's pretty damm soon [01:31:59] *** reidms has joined #opensolaris [01:32:34] <jmcp> reflect: the 3ware fora are full of "we do not support Solaris at present, we may in the future" [01:32:36] <jmcp> gah [01:32:51] <jmcp> idiots [01:33:00] <reflect> 3ware.. escalade.. 8506 [01:33:40] <reflect> it has 2 modes of operation, jbod and their bios mode, which entails various RAID modes [01:33:58] <reflect> if the JBOD option was supported in solaris, it'd be a step forward [01:34:02] <reflect> any takers? [01:34:12] * jmcp looks @ dlg again [01:34:18] <dlg> i cant comment on the jbod mode [01:34:30] <dlg> id say the io path is the same [01:34:37] <dlg> so no joy [01:35:01] <jpdrawneek> is Tempt up? [01:35:09] <jmcp> dunno [01:35:18] <jmcp> jpdrawneek: he was on quite late last night [01:35:35] <jpdrawneek> i know - i want to harass him again :) [01:35:43] <jmcp> no rest for the wicked :) [01:35:48] <jpdrawneek> which time zone is he in? [01:36:00] <jmcp> Australia/Victoria [01:36:07] <reflect> well, this is the second time I've asked.. if there should be someone interested.. please, let me know [01:36:07] <jmcp> so at the moment it's 0936 for him [01:36:45] <jpdrawneek> ta [01:36:52] <reflect> especially if it's someone not in timezone Europe [01:37:44] *** nachox has quit IRC [01:38:38] <reflect> dlg : so.. you're not available for a 3ware driver? [01:38:47] <jpdrawneek> is tempt on most of the day? [01:39:22] <jpdrawneek> any of you guys know much about scsi and fixing issues with it? [01:39:23] <jmcp> jpdrawneek: I think so [01:39:40] <jmcp> jpdrawneek: depends on the question [01:40:04] <jpdrawneek> http://www.opensolaris.org/jive/thread.jspa?threadID=37064&tstart=0 [01:40:47] <jmcp> did you see David Hollister's response to your email? [01:40:54] <jpdrawneek> yes [01:41:17] <jpdrawneek> he said he would get back to me [01:41:35] <jmcp> which driver are you using with it? [01:41:41] <jpdrawneek> some time on monday [01:41:44] <jpdrawneek> mpt [01:42:15] <jpdrawneek> think i need to lamp mpt and use itmpt (lsi one) [01:42:32] <jmcp> what's the pci vendor/device id for that card? [01:42:43] <jpdrawneek> lsi [01:42:57] <jpdrawneek> what do y ou mean? [01:43:04] <jmcp> jpdrawneek: run "prtpicl -v | less" [01:43:11] <jmcp> search for pci1000 [01:43:12] <reflect> every card has a pci id, and a vendor id [01:43:31] <reflect> it's basically what the drivers use to attach themselves to the card in question [01:43:50] <jpdrawneek> it attaches it self to scsi@4 [01:44:10] <jmcp> jpdrawneek: nope, that's not quite it either [01:44:31] <jpdrawneek> box is off [01:44:34] <jmcp> jpdrawneek: oh [01:44:38] <jpdrawneek> back in five [01:44:56] <reflect> http://www.sun.com/bigadmin/hcl/driverlist.html <-- there's a PCI ID and Vendor ID list [01:47:49] <jmcp> jpdrawneek: this might help a bit - http://blogs.sun.com/jmcp/date/20051117 [01:48:39] *** tombhadAC has quit IRC [01:49:57] <FrostCS> Anyone need a laugh? I can honestly tell you people believe this stuff.. http://rootuser.insanejournal.com/ [01:51:31] <Tempt> Morning all. [01:51:36] <jmcp> hi Tempt [01:51:46] <jmcp> jpdrawneek has been anxiously awaiting your return [01:51:55] <Tempt> Oh dears. [01:52:02] <Tempt> I still haven't had my morning caffeine quota [01:52:47] <reflect> FrostCS: I don't see the funny part, though.. I see a rant.. [01:52:58] <jmcp> FrostCS: that's hilarious [01:53:11] * Tempt heads off to read the link [01:53:14] <FrostCS> I had someone quote that, telling me that's how it is.. [01:53:22] <FrostCS> and I was told the T1 processor sucks [01:53:23] <FrostCS> lol [01:53:31] <jmcp> they'll be laughing on the other side of their face soon enough [01:54:12] <jmcp> far out [01:54:46] <FrostCS> Someone had asked a question on hardforums about the T2 processors.. and had the thread forwarded to me.. I chimed in a bit.. and I couldn't help but laugh when I read the things people say [01:54:48] <jmcp> "Eight. PCIe. Lanes. [01:54:48] <jmcp> Oh. My. Gods. You have GOT to be kidding me. Did somebody not tell Sun that we've had PCs shipping with 40 lanes since 2005? " ... yeah dickhead, on the f!$!$ng cpu itself [01:54:53] <jpdrawneek> :binding-name pci1000,30 [01:55:00] <FrostCS> http://www.hardforum.com/showthread.php?t=1217774 [01:55:03] <jmcp> jpdrawneek: aha [01:55:12] <jmcp> jpdrawneek: did it ever work? [01:55:19] <Tempt> Loving it. [01:55:22] <FrostCS> jmcp, I had a good chuckle about it too.. [01:55:24] <jmcp> jpdrawneek: which build are you running? [01:55:31] <jpdrawneek> solaris 10 u3 [01:55:43] <Tempt> FrostCS: That's brought a laugh to my otherwise bleak monday morning. [01:55:48] <reflect> we employ some 500+ machines of linux and I don't see that kind of problems.. perhaps we're just using it differently [01:55:48] <FrostCS> there is the forum thread if anyone wants to have a reply rant.. I kind of get bored of stupidity quickly [01:55:51] <jmcp> jpdrawneek: your mpt driver is patched to current? [01:56:13] <FrostCS> reflect, that and the T2 stuff.. it's just funny [01:56:37] <jpdrawneek> give smpatch 5mins [01:56:54] <jmcp> FrostCS: ....SPARC procs aren't very well known or used,.,., [01:57:05] <jmcp> FrostCS: it's so ill-informed it's hilarious [01:57:25] <jpdrawneek> what do you expect from the unwashed masses [01:57:36] <jmcp> FrostCS: that poster is clearly a gentoo user [01:57:37] <reflect> the T2 rocks.. but it'll all depend on the pricing [01:57:55] <FrostCS> jmcp, yea, most of the guys are windows/newbie linux brats I was trying to educate a little.. [01:58:02] <jmcp> FrostCS: don't bother [01:58:10] <FrostCS> jmcp, obviously too late.. heh [01:58:11] <jpdrawneek> i seem all patched up [01:58:28] <jmcp> jpdrawneek: ok, good. so back to the other question - did this ever work? [01:58:36] <jpdrawneek> i don't know [01:58:41] <jmcp> and what hardware do you have this card installed in? [01:58:52] <jpdrawneek> netra 1405 [01:59:16] <jmcp> where did the card come from? [01:59:24] <reflect> I actually think the attitude "we're better than you are" is bad. let's face it, most OSes out there has their uses.. it all depends on what you want out of it [01:59:26] <jpdrawneek> ebay - sold as working [01:59:34] <jmcp> jpdrawneek: hmm [01:59:53] <Tempt> jpdrawneek: Didja manage to update the firmware on that card? [02:00:00] <jpdrawneek> hense the urgency to find out whats wrong [02:00:03] <jpdrawneek> no [02:00:14] <jpdrawneek> Tempt: i don't know hoe to do that [02:00:26] <Tempt> Aah [02:00:31] <jpdrawneek> Tempt: nobody told me yet [02:00:40] <Tempt> Raise a case with LSI [02:00:50] <Tempt> All you need is your serial number and they'll respond pretty quickly. [02:01:02] <jpdrawneek> ok [02:01:03] <jmcp> serial # of the card [02:01:04] <Tempt> You do not need a service contract with them; just raise a case on their website [02:01:07] <FrostCS> reflect, if anything people just don't like to think there is something out there they don't know about.. [02:01:13] <jpdrawneek> what do i ask for? [02:01:15] <Tempt> I raised one for my LSI SAS card and they responded in 9 hours. [02:01:28] <Tempt> Just cut and paste as much information as possible into the fault description. [02:01:38] <jpdrawneek> ok [02:02:45] <jpdrawneek> quick question [02:03:40] <jpdrawneek> scsi card 320 - disk 320 - array 160 - lsiutil reports fast wide (or something) ideas on whats up? [02:03:41] <reflect> FrostCS: well, then you can either fuel that fire of ignorance, or you can work against it. your choice [02:04:48] <jmcp> jpdrawneek: what does "prtpicl -v" say for the "scsi-options" property of the disk [02:04:49] <SYS64738> anyone knows in wich packages is there the /usr/cluster/bin/scinstall ? [02:04:49] <jmcp> ? [02:05:03] <jmcp> SYS64738: that'll be part of the Solaris Cluster (formerly SunCluster) software [02:05:06] <jmcp> it's not part of PN [02:05:07] <jmcp> ON [02:05:16] <FrostCS> I already added my 2 cents, I just thought you guys might like a laugh, I shouldn't ever need to write more than once on anything :-) [02:05:24] <jmcp> that's for sure [02:05:27] <dlg> reflect: i dont have hardware [02:05:30] <dlg> and time is a problem [02:05:51] <reflect> dlg : ok.. thanks anyway [02:05:51] <jpdrawneek> jmcp: which disk? [02:06:10] <SYS64738> jmcp, what do you mean for ON ? [02:06:15] <jmcp> jpdrawneek: you just mentioned a "card 320, disk 320, array 160" [02:06:16] <reflect> although, I think it works in regular PCI, if that helps [02:06:26] <jmcp> SYS64738: ON is a consolidation, is Cluster [02:06:43] <jpdrawneek> jmcp: will get back to you on that one - things need rejigging [02:07:00] <jpdrawneek> jmcp: could it be a cable problem? [02:07:15] <SYS64738> jmcp, i've installed suncluster 3.2 but I haven't the scinstall command [02:07:24] <jmcp> jpdrawneek: unlikely [02:07:31] <Tempt> SYS64738: That's odd. [02:07:40] <Tempt> SYS64738: How did you install SC? [02:07:48] <jmcp> jpdrawneek: if the array/enclosure that contains your disk doesn't support U320, then you'll never get U320 out of the disks [10:35:42] *** echelog has joined #opensolaris [10:40:29] *** theRealballchalk has quit IRC [10:45:41] *** tomww_ is now known as tomww [10:45:42] <trochej> Drat, my new laptop has a built-in camera. I feel invigilated by my own hardware :) [10:48:04] <Tempt> put a big solaris sticker over it [10:48:29] <trochej> :) [10:48:42] <Tempt> Find a sol10 mediapack; they included stickers [10:49:49] <Tempt> Say, anyone know a way to get screen to put the workspace list on the side of the display instead of the bottom? Or is this one of those "use the source" moments? [10:49:50] <trochej> I didn't get one [10:50:27] <trochej> Tempt: You want whole bottom bar? [10:51:10] <Tempt> I want it on the side. [10:51:17] <Tempt> I'm using a widescreen display [10:51:26] *** migi has joined #opensolaris [10:51:47] <Tempt> I think I might just build the latest dev version and get vertical splits [10:52:50] <LeftWing> Vertical splits, eh? Neat. [10:52:59] <Tempt> Man I wish the laptop SunRays were cheap. I'm using a $2500 laptop as an X-terminal [10:53:05] <LeftWing> lol [10:53:14] <Tempt> LeftWing: I'll package it if you want. [10:53:33] <Tempt> and when I say X-terminal, I mean XDMCP .. [10:53:40] <LeftWing> Mmm. [10:54:03] <LeftWing> Buy a Thinkpad Travel keyboard, a big battery and a lot of duct tape. [10:54:11] <LeftWing> It'll almost be a laptop. ;) [10:57:03] <dme> LeftWing: Do you have the Thinkpad Travel keyboard and, if so, is it any good? (I quite like the laptop keyboards, was thinking about the external one...) [10:58:45] *** aruiz has joined #opensolaris [11:01:33] <LeftWing> dme: I've ordered one. [11:02:06] <LeftWing> I prefer full-size, full-depth keyboards but this one is small and has a trackpoint -- I've bought it for couch/bed/media computing. [11:03:27] *** Giaco has joined #opensolaris [11:03:30] *** whaq_ has joined #opensolaris [11:08:36] *** Snake007uk has joined #opensolaris [11:11:17] *** whaq has quit IRC [11:17:06] *** KermitTheFragger has joined #opensolaris [11:20:24] *** SYS64738 has quit IRC [11:21:47] *** Mazon has joined #opensolaris [11:28:45] <asyd> \_o< [11:32:40] *** linma has joined #opensolaris [11:37:47] *** GmanAFK is now known as Gman [11:38:30] *** bnitz has joined #opensolaris [11:39:08] *** Somethingelse has quit IRC [11:39:19] *** KermitTheFragger has quit IRC [11:40:57] *** Squegie has left #opensolaris [11:44:54] *** jmcp has quit IRC [11:46:13] *** edwardoc has joined #opensolaris [11:46:26] <edwardoc> Hey people, just a quick drop by [11:46:46] <edwardoc> Tempt:Hows things? [11:47:04] <Tempt> edwardoc: Not bad. [11:47:46] <edwardoc> Tempt:I made a template on SSGD on my site ready to be filled with _stuff_^tm [11:47:51] *** Tpenta has quit IRC [11:48:32] <edwardoc> I see dclarke made a blog about Joomla on Solaris :D [11:48:47] <edwardoc> Got a bit more traffic today [11:49:44] <edwardoc> and a new user who I emailed about writing some things up on Solaris FAQs/HOWTOs [11:50:03] <Tempt> cool [11:50:11] *** tsoome has quit IRC [11:50:13] <edwardoc> Java class was good today and I am going to look at that now [11:51:06] <Tempt> fun stuff. [11:51:07] <edwardoc> When I get up to SWING etc I will start doing things to that GUI for pkg's poeple are working on over at opensolaris.org [11:51:16] <Tempt> Except for the "java" bit. [11:51:21] <edwardoc> lol [11:51:24] <dlg> Tempt: heh [11:51:28] <dlg> hi5! [11:51:38] <edwardoc> OOP=OOPs [11:51:45] <Tempt> Oops. [11:52:11] <edwardoc> man, doing ASM and C, looking at OOP is hard but I guess I will get the hang of all this class stuff [11:52:28] <oxygene> with java? hardly [11:53:02] <edwardoc> we try still [11:53:05] *** tsoome has joined #opensolaris [11:53:24] <edwardoc> I got the teacher using NetBeans [11:54:06] <edwardoc> :D so thats made things a bit better for me as I use the C/C++ pack with that IDE and I know a good chunk of were things are [11:54:57] <edwardoc> I do wish the C/C++ Pack for windows version had the gcc/gdb built into it with the path setup at install time [11:55:45] <edwardoc> I see its a bit better on 5.5_1, wonder what its like on 6, not looked at that yet [11:57:30] *** bobbyz has quit IRC [11:58:15] <edwardoc> edwards junk he had to say EOF [11:59:14] <edwardoc> Oh well I better get back to this Java while its still in my head [11:59:36] <edwardoc> All the best [11:59:39] *** edwardoc has quit IRC [12:00:19] *** bengtf_ has joined #opensolaris [12:00:24] <Doc> java in your head? that's gotta be painful [12:01:45] <kohju> $B$U!<$`(B [12:02:31] * PerterB tries to work out whether that's utf8 or a teco macro [12:02:41] <flyingparchment> ISO2022-JP [12:02:54] <PerterB> ah [12:02:59] <kohju> sorry. [12:03:14] <kohju> mistake channel. [12:04:03] <flyingparchment> if it has lots of $'s it's ISO 2022, if it has lots of accented latin characters its utf-8 ;-) [12:04:04] *** KermitTheFragger has joined #opensolaris [12:05:11] <PerterB> yeah, I should have known that, having once had to maintain some dodgy perl where most of the comments were in ISO 2022 [12:05:36] <kohju> I see. [12:06:32] <kohju> almost Japanese channel is used ISO-2022-JP. [12:07:45] <kohju> Recently, we use utf-8 codec in source-code. [12:07:52] *** calumb has joined #opensolaris [12:10:22] *** jmcp has joined #opensolaris [12:10:52] *** calumb has quit IRC [12:17:40] *** Pietro_S has quit IRC [12:17:52] *** Pietro_S has joined #opensolaris [12:18:05] *** calumb has joined #opensolaris [12:19:37] *** tsoome has quit IRC [12:21:13] *** tsoome has joined #opensolaris [12:23:37] <Gekkko> I hate people. [12:23:57] <dlg> they do suck [12:24:14] <renihs> mammals suck, it lies in their nature [12:24:59] <Gekkko> they never fail to disappoint me. [12:25:19] <Gekkko> hmm [12:25:21] <Gekkko> I'm confused [12:25:40] <Gekkko> SXCE has 3 parts. [12:25:55] <Gekkko> each is a part that needs to be "cat"ed together right? [12:26:17] <flyingparchment> i thought it had 5, a-e [12:26:26] <flyingparchment> did they make the parts larger? [12:26:32] <Pietro_S> now only 3 with 2GB size [12:26:35] <flyingparchment> anyway, yes, cat them and then unzip and you'll get an iso [12:26:36] <Pietro_S> 1GB [12:27:21] <Pietro_S> unzip first and then cat them into iso ... [12:27:32] <flyingparchment> duh, right. haven't done that for a while. [12:27:33] <Gekkko> flyingparchment: isn't it unzip then cat? [12:29:30] *** calumb is now known as calAFK [12:30:45] *** tsoome has left #opensolaris [12:30:54] *** Vanuatoo has joined #opensolaris [12:31:15] <reflect> hm, that storage report from google states that raid5 is not as safe as previously thought, cause when one drive fail, you often get correlating errors on other disks, bit errors while rebuilding the array, etc. Does ZFS completely remove this as it does checksumming? [12:31:44] <flyingparchment> checksums don't prevent a disk failing [12:31:52] <flyingparchment> all it does it let you know sooner that it failed [12:32:07] <Tempt> reflect: The checksumming and advanced parity stuff with ZFS lesson the risks. [12:32:20] <Tempt> reflect: And you don't get the silent data corruption that some other RAID solutions have suffered from. [12:32:52] <reflect> hm.. scary. perhaps one should do 6+2 raidz2 instead.. [12:33:05] <renihs> ya but i dont exactly understand how the silent corruption is avoided on zfs? [12:33:15] <Tempt> Because ZFS actually checksums stuff. [12:33:18] <Tempt> Every read [12:33:27] <reflect> and write [12:33:40] <renihs> hmm [12:33:44] <renihs> and it compares then? [12:33:57] <Tempt> If the checksum doesn't add up, it screams. [12:34:07] <Tempt> Try it. [12:34:17] <Tempt> Create a small zpool on some lofi devices [12:34:21] <Tempt> export it [12:34:26] <Tempt> write some random crap over one of the volumes [12:34:30] <flyingparchment> unfortunately it screams by panicing the system [12:34:30] *** WickedWicky has joined #opensolaris [12:34:31] <Tempt> and then import it and try some reads [12:34:46] <Tempt> flyingparchment: Nah, that's only when the whole pool evaporates. [12:34:47] <renihs> hmm ya that i dont doubt but hmm [12:34:55] <flyingparchment> hmm, yeah if you have a mirror it won't [12:34:56] *** mikefut_ has joined #opensolaris [12:35:08] <renihs> flyingparchment, panic is the safest way for shutdown [12:35:09] <renihs> :p [12:35:15] <reflect> I'm amazed at the performance, considering all these checksums [12:35:26] <timeless> renihs: depends on what you consider valuable data [12:35:31] <Tempt> That's because ZFS is made of secret magic j00j00. [12:35:33] <timeless> on average, the data i value is the data i haven't saved [12:35:40] <flyingparchment> i didn't like raidz performance much. but i don't think that was because of the checksumming. [12:35:47] <renihs> timeless, true :p [12:36:04] <timeless> renihs: speaking from a laptop that bluescreened earlier this morning [12:36:22] <timeless> i don't actually blame ms, i think my employer+dell have conspired to add some buggy drivers [12:36:24] <reflect> flyingparchment: really? I get nearly 600MB/sec in writes.. reading is actually slightly less [12:36:39] <timeless> my employer included a feature which prevents me from getting the equivalent of an os core file [12:36:51] <Gekkko> OOMG NO [12:36:52] <flyingparchment> i was copying a large number of small files from a 6-disk raidz and got fairly lacklustre performance [12:36:55] <timeless> so i can't analyze/debug the bosd :( [12:36:59] <Gekkko> My SXCE part B has a bad checksum [12:37:02] <Gekkko> I'm about to kill someone [12:37:11] <flyingparchment> Gekkko: did you try to download it by changing the url in wget? [12:37:15] <timeless> gekkko: download or disk? [12:37:27] <Gekkko> download [12:37:35] <Gekkko> flyingparchment: what's that meant to mean [12:37:36] <Gekkko> >_> [12:37:43] *** damienc has joined #opensolaris [12:37:55] <flyingparchment> Gekkko: you get a long URL ending with /sx-....-a.iso. you can wget that to download [12:38:02] <flyingparchment> but you can't change -a to -b to get the next part [12:38:20] <Gekkko> I'm using a download manager on Windows [12:38:33] <timeless> did you just click on the individual links? [12:38:36] <timeless> or did you cheat [12:38:41] <timeless> flying is used to people who cheat [12:38:43] <Gekkko> I clicked the individual links [12:38:47] <Gekkko> I had a drop out a few times. [12:38:48] <timeless> if you clicked the links, ignore flying :) [12:39:05] *** Vanuatoo_ has joined #opensolaris [12:39:11] <timeless> otoh, if you're trying resume w/ a download manager, um, [12:39:16] * timeless doesn't support that [12:39:24] <timeless> web servers suck in most instances [12:39:32] <timeless> proxies/caches make that even worse [12:39:41] <timeless> iirc sun will send you a dvd if you ask nicely [12:39:44] <Gekkko> It worked for parts 1 and 4 [12:39:45] <Gekkko> 3* [12:39:50] <Gekkko> I'm getting one [12:39:53] <Gekkko> but I want SXCE NOW [12:39:54] <Gekkko> >_>! [12:39:59] <Gekkko> wow wtf. [12:40:03] <Gekkko> I just got 250kb/s for no reason. [12:40:11] <Gekkko> and now I'm back to 50kb/s lol. [12:40:17] <Gekkko> I hate my ISP. [12:40:22] *** simford has quit IRC [12:40:26] *** postwait has quit IRC [12:41:30] *** calAFK is now known as calumb [12:44:02] <Pietro_S> Gekkko: are you using SDM? it's much faster than downloading with enything else? [12:44:17] <Gekkko> no, and I don't know [12:44:21] <Gekkko> Java doesn't play nice with me [12:45:07] <timeless> Pietro_S: i never bothered to check, is that basically a bittorrent or equivalent distributed client? [12:45:16] <timeless> or loadbalancing or... [12:45:45] <timeless> personally i get the best performance by downloading through intermediaries who are better connected to remote, myself than the path between myself and the remote [12:47:24] <renihs> ingenious :p [12:47:38] <Pietro_S> timeless: I don't know internals of it but it's much faster I guess that t done some loadbalancing ... at university the difference is quite big [12:47:45] <timeless> renihs: i've been doing that for years [12:47:58] <renihs> :p [12:48:05] <timeless> i remember in college that a friend on i2 had better access to netscape's ftp servers than i did on i2 [12:48:13] <renihs> ya me too [12:48:15] <timeless> so i downloaded to his machine and then copied to mine [12:48:19] <Pietro_S> SDM download about 5-7 Mb/s firefox is happy with 1 ... [12:48:26] <renihs> always fetched to a university, dragged from there to studenthome [12:48:55] <timeless> i2<=>i2 connections were so fast that they really didn't influence download time :) [12:49:24] <timeless> i also remember finding that downloading a gz file and extracting it was faster than getting the bz2 file and extracting it :) [12:49:46] <timeless> even if i saved 10mb of someone else's bandwidth, my cpu just couldn't deal as well :) [12:49:55] <renihs> heh, ya i never use bz2 ...so much slower [12:49:59] <renihs> lzma ftw :p [12:50:10] <renihs> looooong compression, fast decompression :p [12:50:26] *** mikefut has quit IRC [12:50:26] <renihs> better than gz, slightly worse than bz2 [12:51:12] <timeless> if it isn't portable, i don't like it :) [12:51:13] <timeless> zip. [12:51:27] *** mikefut_ has quit IRC [12:51:28] <renihs> it isnt? [12:51:39] <timeless> oh, and my definition of portable is: works everywhere w/o requiring installing something [12:51:42] <renihs> lzma is just not commonly known/used but else dunno why its so seldom [12:52:00] <renihs> timeless, hehe.... [12:52:27] <WickedWicky> sorry to interrupt but.. does anyone know how long memcheck will take with 3072MB of ram? [12:52:38] <renihs> all tests? [12:52:42] <WickedWicky> yesh [12:52:48] <renihs> pff should take a few hours i guess? [12:52:58] <WickedWicky> cheers :D [12:53:27] <renihs> but dunno, just a educated guess, last time i test 16gb i didnt wait :p [12:53:36] <renihs> not really educated though :p [12:53:39] <renihs> the guess that is [12:53:42] <WickedWicky> it's not like i am gonna stare at the screen either.. :P [12:53:43] * timeless rotfl@http://slashdot.org/pollBooth.pl [12:56:05] <WickedWicky> I assume "Pass 1%" kinda gives away the progress of the test.... [12:57:08] <Gekkko> lol [12:58:33] *** Vanuatoo has quit IRC [13:04:23] *** deather_ has joined #opensolaris [13:07:05] *** kaiwai has joined #opensolaris [13:08:44] *** MattMan is now known as MattAFC [13:10:33] <trochej> Khe [13:11:35] *** deather has quit IRC [13:11:38] *** deather_ is now known as deather [13:19:11] *** lplatypus has joined #opensolaris [13:23:19] *** phimic has joined #opensolaris [13:30:18] *** MattAFC is now known as MattMan [13:33:30] *** solar-star has joined #opensolaris [13:35:42] *** crib has quit IRC [13:35:48] <WickedWicky> hooray! [13:35:52] <WickedWicky> faulting memory [13:36:07] <WickedWicky> that'd explain a crash or two [13:36:44] *** crib has joined #opensolaris [13:36:50] <lplatypus> has anyone shot down the rumour yet that opensolaris' legal status is murky now that a court has ruled that SCO doesn't own the rights to Unix thus undermining the licensing deal between SCO and Sun? [13:38:08] <Cyrille> now it's even better, on /. they claim Novell can and should sue Sun for infringing what they think is now Novell's IP. [13:38:38] <Cyrille> oh sorry that's what you were referring to. [13:38:59] <purserj> meh [13:39:18] <lplatypus> my understanding is that SCO did have the right to sublicense Unix, even though they didn't own it [13:39:44] <purserj> Yup, however they had to funnel most of the fees from that back to Novell [13:39:44] <lplatypus> is that right? [13:39:49] <trochej> /. is rarely a belivable source of information and JP Voughan (or whatever) is rarely a real, objective journalist [13:39:58] <purserj> groklaw is better [13:40:03] *** xushi_ has joined #opensolaris [13:40:10] <lplatypus> yes I like groklaw, but a lot of people read slashdot [13:40:25] <Cyrille> I didn't say it was, but it's only on /. that I saw that sort of comments popping up [13:41:07] <lplatypus> I think it was Stephen Vaughan-Nichols who started the rumour: http://desktoplinux.com/news/NS3180933176.html [13:41:11] *** StylusEater has left #opensolaris [13:41:20] <Cyrille> there even was the obligatory "Novell should sue Sun and force it to release OpenSolaris with a Linux compatible license so that we can pilfer the best bits of it" [13:41:21] <lplatypus> clearly marked "opinion" [13:41:29] *** rbrown_ has joined #opensolaris [13:42:31] <purserj> either way, Novell's not going to spank Sun, they've just finished up one law suit and Sun has real lawyers. Instead they're going to be demanding that SCO hand over the fees that it collected from Sun [13:42:52] <lplatypus> which it has spent already :-) [13:43:01] <purserj> So SCO is now basically dead man walking [13:43:16] <lplatypus> I'm happy to see Darl's SCO go down [13:43:47] <purserj> Tarantela's still going so old SCO will go on [13:44:37] <purserj> however to be honest, anything that takes OpenServer off the market is a good thing [13:45:09] *** freakazoid0223 has quit IRC [13:45:47] <quasi> few people will miss OpenServer [13:46:56] <dlg> purserj: "they should just kill it" [13:47:05] <purserj> dlg: bleh [13:47:15] <dlg> :) [13:47:18] <purserj> Don't get me started on inter-community bitching [13:47:35] * dlg grin [13:47:49] *** halton has left #opensolaris [13:47:51] *** freakazoid0223 has joined #opensolaris [13:47:58] <Tempt> OpenSewer more like [13:48:05] <Tempt> I got stucking looking after one of those once. [13:48:07] <Tempt> it sucked. [13:48:17] *** MattMan is now known as MattAFC [13:48:29] <purserj> Tempt: I had to do two boxes for four years [13:48:44] <purserj> We actually had them scheduled to reboot every night [13:48:50] <lplatypus> what was its selling point? x86 unix? [13:48:53] <Tempt> Horrible. [13:48:58] <Tempt> Yes, x86 UNIX [13:49:13] <purserj> it came from Xenix [13:49:48] <Tempt> Xenix was as much a SCO thing as a Microsoft one. [13:50:01] <Tempt> Back when it was Santa Cruz Operation. [13:50:01] <lplatypus> did xenix run on 16-bit hardware? [13:50:09] <purserj> Tempt: yup [13:50:14] <Tempt> Always thought that was a groovy name, Santa Cruz Operation. [13:50:32] <purserj> Basically everything went to shit when Caldera bought out the Unix operations of SCO [13:50:59] <solar-star> Hi [13:51:14] <solar-star> anyone there who knows how to stop gdm? [13:51:51] <Tempt> purserj: Don't forget Caldera going after Microsoft first. With the approval of every slashweenie in history. [13:52:19] <Tempt> purserj: It wasn't until the poked at the sacred penguin that they went from a dream freedom-fighter to vexatious litigant [13:52:39] <lplatypus> solar-star: you mean to log into the text console? gdm has an option for that [13:52:46] *** jpdrawneek has joined #opensolaris [13:52:59] <solar-star> I need to shutdown X without restarting [13:53:11] <lplatypus> ctrl+alt+backspace :-) [13:53:27] <solar-star> That does not the trick [13:53:30] *** sparc-kly has joined #opensolaris [13:53:31] *** ChanServ sets mode: +o sparc-kly [13:53:33] <solar-star> I restarts again [13:53:55] <purserj> Tempt: meh, who knew what darkness lurked in the heart then [13:53:58] <lplatypus> you can probably do something like: svcadm disable gdm [13:54:12] <solar-star> yes I think that is what I need [13:54:26] <lplatypus> not sure if "gdm" is the right name of the service... use svcs to find the right name [13:54:56] <solar-star> Is there a possibility not to disable but to shutdown? [13:54:58] <WickedWicky> type svcs | more [13:55:00] <WickedWicky> and have a look [13:55:08] <Tempt> purserj: They were just as dark when they were going after Microsoft. I mean, the case was based on Microsoft stopped them levering their DR-DOS purchase in the late nineties [13:55:11] <WickedWicky> I think it's gnome-loging-manager [13:55:18] <WickedWicky> but I'm not sure [13:55:33] <solar-star> Which starts all over again? [13:55:54] <solar-star> Ok, I will give svcs a try [13:55:54] <WickedWicky> svcadm disable <name of the service that starts gdm> [13:56:05] <solar-star> Ok [13:56:20] <lplatypus> I can't remember if there is a "restart" verb... you could check "man svcadm" [13:56:24] <Cyrille> there is [13:56:27] <Cyrille> I think [13:56:39] <WickedWicky> restart wont do you much good when you want to kill X though [13:56:45] <Chipdancer> WickedWicky: cde-login, I would assume [13:56:57] <WickedWicky> he specificly mentions gdm [13:56:57] <Cyrille> and the GDM service used to be something like login-gdm2, svcs -a | grep gdm would help [13:57:07] <Cyrille> cde-login is dtlogin [13:57:19] <Chipdancer> Cyrille: right, wasn't sure whether it was dependant [13:57:45] <lplatypus> how early did unix get memory protection hardware? [13:58:18] <Tempt> By keeping each process in its own memory segment [13:58:24] <Tempt> At least on x86. [13:58:37] <Tempt> That's the way minix did it [13:59:06] <lplatypus> oh okay, using the 16-bit protected mode I presume [13:59:26] <jpdrawneek> Tempt: think i have found out why its not booting [13:59:50] <Tempt> hmm? [13:59:53] <jpdrawneek> Tempt: the new card seems to ko the trap 3e hack for sleepy disk [13:59:54] *** sponix has quit IRC [14:00:37] <Atomdrache> Does anybody know of any good CD burning applications for Solaris 10? [14:01:02] <Atomdrache> Need to burn an ISO. [14:01:08] <jpdrawneek> Tempt: ideas on why it would mess the script up? [14:01:24] <WickedWicky> Atomdrache: cdrecord [14:01:33] <WickedWicky> growisofs for dvds [14:01:43] <Tempt> jpdrawneek: Dodgy firmware probably. [14:01:48] <Tempt> jpdrawneek: What did LSI say? [14:01:55] <jpdrawneek> nothing yet [14:02:17] <Atomdrache> Oh, already there. Thanks. [14:02:54] *** calumb is now known as calLNCH [14:05:30] *** killersto has joined #opensolaris [14:07:37] <killersto> helo. i have SX64. it has bugs in Xorg. can i update only Xorg in it? i know i can just get latest version, but i'd like to know if it's possible to update system parts individually [14:07:38] <tomww> Atomdrache: and: cdrw (already there too) [14:08:35] <Atomdrache> Well this is odd. [14:08:38] <tomww> killersto: it is for SX not usual to update only parts. but if you like, you can identify all X-packages and only update those manually [14:08:47] <Atomdrache> Oh wait. [14:08:50] <Atomdrache> It was just being slow. [14:08:50] <tomww> Atomdrache: hmm? [14:09:15] <Atomdrache> I was going to say "Oh crap, it can't find my CD-RW drive, even though I've used it before", but...it was just being slow. [14:09:39] <Atomdrache> "Warning: controller creates hard SCSI failure when retrieving CD capabilities page." worries me, however. [14:10:10] *** master_o1_master has joined #opensolaris [14:12:06] *** tombhadAC has joined #opensolaris [14:14:18] <killersto> tomww: is it possible to build parts from source? [14:15:18] *** jmcp has quit IRC [14:15:56] *** jmcp has joined #opensolaris [14:16:39] <killersto> btw, maybe somebody can suggest a trick to configure Xorg? it fails to load on G965 chipset due to known bug. however, i think it can work in generic VGA mode -- i've seen gdm once intialized to normal 1024x768, colored -- but it was only once, i guess it's some random autodetection thing.. [14:16:40] <tomww> killersto: what >parts< are you interested in? [14:16:47] <killersto> Xorg [14:16:58] <purserj> Tempt: fair enough [14:16:59] <killersto> X server [14:17:10] <purserj> To tell you the truth I didn't actually pay that much attention to them [14:17:16] <purserj> Their distro sucked [14:20:04] *** cmihai has joined #OpenSolaris [14:20:50] *** master_of_master has quit IRC [14:21:49] *** hohum_ has quit IRC [14:24:03] <Atomdrache> Will it complain if I just do cdrecord -v image.iso? [14:24:14] <Atomdrache> (with a CD burner that is) [14:24:28] <Tempt> cdrecord dev=x,y,z -v -data image.iso [14:24:46] <Atomdrache> Eww, I have to specify the device? [14:24:52] <Tempt> cdrecord -scanbus [14:25:01] <Atomdrache> Did that earlier. [14:25:13] <Tempt> well, that'll give you x, y and z [14:25:15] <Atomdrache> In case I needed to. [14:26:08] <seanmcg> cdrw -i /path/to/iso [14:26:53] *** rbrown_ has quit IRC [14:27:14] <WickedWicky> or cdrecord dev=/dev/rdsk/cXtYdZs2 blergh.iso [14:27:28] <Tempt> whatever, the disc should be burnt by now [14:27:31] <tomww> killersto: should be possible, but I think you should consider just burning a new DVD and do a in-place very simle solaris upgrade. [14:27:36] <WickedWicky> Tempt: ya [14:27:41] <Atomdrache> Do I have to do the whole /dev/ thing or can I just do, say, dev=0,1,0? [14:27:44] <Atomdrache> Seems to be all right with that. [14:27:47] <WickedWicky> either will work [14:27:59] <WickedWicky> well wait [14:28:03] <Tempt> Atomdrache: Just get it cooking. It'll give you a chance to change your mind [14:28:06] <Atomdrache> Tempt: Not really. It's an old Blade 100 with an I don't even know how slow CD-RW drive. I might be waiting a while :3 [14:28:19] *** killersto has quit IRC [14:29:01] <Tempt> heh. [14:29:15] <WickedWicky> of course the faulting memory module is the only non-kingston strip [14:29:20] <Atomdrache> Printed out a bunch of stuff about my device, buffer size, and a few warnings I'm not sure I should be worried about. Now I've just got a blinking cursor and I'm hoping it'll tell me some things and let me know whether it's burning a CD or playing internet poker. [14:29:25] <WickedWicky> hooray for me and buying cheap-ass memory 4 months ago [14:29:55] <Tempt> WickedWicky: ECC memory is your friend. [14:30:06] <Atomdrache> WickedWicky: I hope the memory I got for my Ultra 80 isn't crap. Pulling out that riser card sucks. [14:30:32] <Atomdrache> Well...no, putting it back in sucks. [14:31:47] *** Gekkko[PDA] has joined #opensolaris [14:31:54] <Atomdrache> Eww. Eww. [14:31:57] <Atomdrache> Write error. [14:32:06] <Gekkko[PDA]> hey [14:32:13] <Gekkko[PDA]> still no irssi >.> [14:32:18] <Tempt> heh [14:32:23] <Atomdrache> Sure I don't need any other arguments to cdrecord? [14:32:25] <Gekkko[PDA]> it's breakin my balls.; [14:32:26] <Tempt> Pay up for a shell account somewhere [14:32:28] <Atomdrache> Maybe I'll try cdrw. [14:32:33] <Tempt> instant write error? [14:32:36] <Tempt> or a buffer underrun? [14:32:39] <Atomdrache> Yep. [14:32:44] <WickedWicky> yep to what? [14:32:54] <Atomdrache> To Tempt's question. [14:33:11] <Atomdrache> write track data: error after 0 bytes [14:33:16] <Atomdrache> Earlier: [14:33:31] <Atomdrache> Drive returns wrong startsec (0) using -150 [14:33:53] <Tempt> oh dear. [14:33:53] <Atomdrache> I/O error. write_g1: scsi sendcmd: no error [14:33:59] <Tempt> wierdo drive? [14:34:03] *** phimic has quit IRC [14:34:10] <Atomdrache> Some LG thing. [14:34:25] <Atomdrache> CED-8080B. [14:34:36] <Tempt> Hmm [14:34:43] <Atomdrache> But, yeah, it generates its fair share of error messages. [14:34:49] <Tempt> Check in the docs if there's something specific for your burner [14:34:57] <Tempt> otherwise try pulling the speed down [14:35:00] <Tempt> for good luck [14:35:05] <Atomdrache> Such as "controller creates hard SCSI failure when retrieving CD capabilities page." What's that mean? [14:35:14] <Atomdrache> I get that from a few applications. [14:35:48] <Tempt> It means your burner isn't playing by the rules. [14:36:00] <WickedWicky> which would be very LG [14:36:02] <Gekkko[PDA]> hit j [14:36:06] <Gekkko[PDA]> it [14:38:07] <Atomdrache> As for the other errors, does that suggest that (1) I need to give cdrecord funny arguments, (2), My CD burner should be removed because it's damaged, or (3) My CD burner is some crap designed for a Wintel box? [14:38:40] *** Gekkko[PDA] has quit IRC [14:38:45] <Atomdrache> (or (4) Reply hazy, sacrifice a goat to my SCSI chain and try again later?) [14:39:25] *** Gekkko[PDA] has joined #opensolaris [14:40:32] <Atomdrache> Wait, that's an ATAPI/IDE drive because it's a Blade 100. Goats do nothing. [14:42:19] <tomww> Atomdrache: something like: cdrecord -dev=ATAPI:0,0,0 -atip for testing? [14:45:15] <Tempt> Nobody ever keeps a goat and black candles anymore [14:45:22] <Tempt> Hence SCSI problems are on the increase again. [14:45:27] <dlg> Tempt: i sure do [14:45:31] <Tempt> You need a sheep and beige candles for IDE [14:45:45] <dlg> its all the same these days [14:45:50] <Atomdrache> I thought you just needed a hammer for IDE. [14:45:54] <Gekkko[PDA]> what about SATA? [14:45:57] <WickedWicky> I use IDE use /dev/rdsk/cXtXdXs2 without any fancy parameters [14:46:01] <dlg> sata is ide [14:46:04] <Tempt> Gekkko[PDA]: A bag of rice [14:46:04] <tomww> Atomdrache: ask the recorder about the empty cdrom in the drive (manufacturer, other parameters) - this is handy for checking the communication [14:46:05] <Atomdrache> Gekkko[PDA]: A fat wallet. [14:46:06] <dlg> just that the cables are thinner [14:46:21] <Gekkko[PDA]> lol rice [14:46:43] <Atomdrache> Okay, this is what -atip gives me. [14:46:58] <Atomdrache> Manufacturer is guessed because of the orange forum embargo. [14:47:17] <Atomdrache> I/O error. read_format_capacities: scsi sendcmd: no error [14:47:21] <Atomdrache> (stuff) [14:47:22] <Gekkko[PDA]> what embargo [14:47:28] <Gekkko[PDA]> >.> [14:47:29] <Atomdrache> status: 0x2 (CHECK CONDITION) [14:47:40] *** yongsun|wfh has quit IRC [14:47:41] <WickedWicky> it's an LG. At best you can try to upgrade the firmware or something [14:47:42] <Atomdrache> Sense Key: 0x5 Illegal Request, Segment 0 [14:47:51] <Atomdrache> Invalid this, invalid that... [14:47:52] <WickedWicky> the whole reason why I switched from LG to AOPEN is this kind of crap [14:48:17] <Atomdrache> I'm going to go rifle through my absent roommate's bedroom for a not shit CD burner. [14:48:29] <Tempt> Good plan [14:48:35] <Tempt> I use a firewire burner on my blade [14:48:41] <Tempt> Pioneer DVD thingy [14:48:47] <WickedWicky> the only decent stuff that LG makes are TFT screens [14:48:57] <Atomdrache> The computer store in town here, I find, is just not a good place to go for Sun parts. [14:49:32] <Atomdrache> He likes to buy bottom-shelf PC crap, often of only one model with an awful chipset. [14:49:33] <Tempt> I got my firewire box and pioneer drive from the local yum-cha pc shop [14:49:36] <Atomdrache> Hence, I often have to improvise. [14:49:56] <Atomdrache> Or drive to the Black Hole in Los Alamos and dig something from beneath a pile of old Apple manuals. [14:50:00] *** lplatypus has quit IRC [14:50:03] <Atomdrache> (eBay is usually the best choice, however) [14:50:03] * WickedWicky goes to his butcher for computer stuff [14:50:25] <WickedWicky> safes on the goat/sheep stock at home [14:50:39] <Atomdrache> (The Black Hole is better for vacuum tubes, radio crystals, huge variacs, 100kV power supplies, that kind of thing.) [14:50:48] <Tempt> mmm [14:50:49] <Tempt> huge variacs [14:50:55] <Tempt> 100kV power supplies [14:50:56] <Tempt> mmm [14:50:57] <Atomdrache> Whatever LANL throws away, basically. [14:51:02] <Atomdrache> Awesome place. [14:51:23] <Atomdrache> Just can't get very good Sun parts there. Though I did find a 411 external drive case with a CD-ROM drive. [14:51:35] <Atomdrache> And there are a couple of SPARCstation-1s and -2s. [14:51:36] <Tempt> woah, don't see this too often: [14:51:39] <Tempt> /usr/lib/cpu/sparcv9+vis2/libmlib.so.2 [14:51:44] <Atomdrache> What? [14:51:45] <Tempt> (from ldd) [14:51:59] <Gekkko[PDA]> und? [14:52:17] <Tempt> Now, guess the application that's pulling in that optimised libmlim [14:52:20] <Tempt> libmlib [14:52:21] <Tempt> .. [14:52:30] <Gekkko[PDA]> cat [14:52:33] <Tempt> Would you believe .. vim. Yes, a text editor. [14:52:39] <Gekkko[PDA]> lol [14:52:50] <Gekkko[PDA]> I was close :P [14:52:53] <Tempt> Atomdrache: Didja grab the SS2s? [14:53:11] <Atomdrache> Didn't have the cash to throw at 'em. [14:53:55] <Atomdrache> I figure I've already got an IPX and a SPARCstation-20. Until I get a fatter wallet, I don't really need more sun4c or sun4m boxes. [14:54:02] <Atomdrache> (oh, and a SPARCserver 630MP.) [14:54:04] *** tombhadAC has quit IRC [14:54:13] <Atomdrache> Or maybe that one's just sun4. [14:54:41] <Atomdrache> Actually, I wasn't there to look for anything in particular. I was going to Los Alamos to get an Alpha server out of some dude's garage. I'm actually not sure yet if it works. [14:55:00] <Atomdrache> But I only go up there so often, and it's always worth a stop at that place. [14:55:18] <Atomdrache> Ed Grothus always has some weird thing I want to buy. [14:55:56] <Atomdrache> Actually, he always has a few metric tons of things I'd like to buy but shouldn't for financial reasons and because I can't stuff more than a few hundred pounds' worth into my car. [14:56:07] <Atomdrache> (could use another boatanchor oscilloscope) [14:56:15] <Atomdrache> (mmm....dual-beam) [14:56:17] <Tempt> Wish there was some crazy place to grab stuff around here [14:56:33] <Tempt> Alphaserver? [14:56:35] <Tempt> What sort? [14:56:39] <Atomdrache> 2100. [14:57:01] <Atomdrache> Humongous monster about half as tall as a refrigerator. [14:57:15] <Atomdrache> Four processors and two memory boards the size of the average PC motherboard. [14:57:26] <Atomdrache> No hard drives or hard drive cage; need to bolt something in there. [14:57:31] <Tempt> COOL [14:57:33] <Atomdrache> Need to attach a serial terminal, too. [14:57:36] <Atomdrache> YES, VERY. [14:57:45] <Atomdrache> I just hope nothing's wrong with the motherboard. [14:57:52] <cmihai> Oh please [14:57:55] <Atomdrache> It sat in a garage face-up and got a bunch of sawdust on it. [14:57:58] <cmihai> AlphaServer 2100 has room for 8 disks [14:58:00] *** mikefut has joined #opensolaris [14:58:01] <Atomdrache> And I'm a pessimist :D [14:58:02] <cmihai> You can put 4GB SCSI disks [14:58:05] <cmihai> And it has video [14:58:11] <Tempt> There's your VMS machine [14:58:14] <cmihai> And it's hardly the size of a fridge [14:58:14] <Atomdrache> Well, the video output isn't giving me anything. [14:58:18] <cmihai> It's more like half. [14:58:23] <Tempt> Let me know if you need a Tru64 mediapack. [14:58:23] <Atomdrache> I said half the size of a fridge. [14:58:26] <cmihai> Atomdrache: well, you're doing something wrong :P [14:58:41] <Atomdrache> Somebody might've gone into the monitor and done, like...set output serial. [14:58:55] <cmihai> Anyway, they're great machines. [14:59:02] <cmihai> I've got 2. [14:59:08] <Atomdrache> I plugged it in, attached a keyboard and a monitor, and powered it on. It's entirely possible that it's hosed, considering where it *came* from. [14:59:09] *** laca has joined #opensolaris [14:59:10] *** Chihan has joined #OpenSolaris [14:59:21] <cmihai> heh [14:59:22] <Atomdrache> If I can get it to work, I'll put OpenVMS on it and scare my system administrator. [14:59:36] <renihs> vms... [14:59:40] <cmihai> That's all those buggers run. [14:59:43] <cmihai> Don't bother with BSD. [14:59:50] <cmihai> Tried it, waste of time, Alpha bug :P [14:59:52] <Atomdrache> Because I hear VMS and descendants thereof is a really well-designed OS. [15:00:03] <cmihai> OpenVMS is a nice OS, yeah [15:00:11] <Atomdrache> It's been around long enough that if it sucked, people would whine about it. [15:00:21] <cmihai> Well, if they'd use it :P [15:00:24] <Atomdrache> But no. Everybody I've ever talked to who has used it says it rules. [15:00:35] <cmihai> Consider the source :P [15:00:43] <Atomdrache> Also, it can do creepy stuff like cluster computers in different cities. [15:00:49] *** nikinana has joined #opensolaris [15:01:12] <Atomdrache> In any case, definitely worth a try. [15:01:24] <cmihai> Eh, clustering and rolling upgrades are nice... Files-11 ain't half bad either... [15:01:30] <Atomdrache> But if you have experience with Alphas, can you tell me what's going on if the front panel stops at "Probe I/O subsys"? [15:01:36] <cmihai> Pretty stable Oracle server I say. [15:02:12] <cmihai> Atomdrache: check your cards [15:02:31] <renihs> or cables :p [15:02:33] <cmihai> Maybe you have a borked EISA / PCI card. [15:02:40] <Atomdrache> I might try yanking them out and booting again. [15:02:41] <cmihai> Nope, PROBE I/O is PCI / EISA [15:02:47] <renihs> oh [15:03:06] <cmihai> It's right before the memory test. [15:03:08] <cmihai> Does it freeze there? [15:03:19] <Atomdrache> Freezes on probing the buses. [15:03:26] <cmihai> Yeah, figures. [15:03:27] <Atomdrache> Doesn't tell me anything about the memory. [15:03:33] <cmihai> Hosed EISA / PCI cards or BUS [15:03:37] <kaiwai> meh [15:03:38] <Atomdrache> Well...actually, I thought it tested them before. But maybe it just initializes it or such. [15:03:56] <Atomdrache> CPUs all test good. That's nice to seee. [15:04:13] <Atomdrache> Well, with that advice I'm going to pull out some cards, plug it in, and see what happens. [15:04:14] <cmihai> Do you have a copy of the manual? [15:04:22] <Atomdrache> Yeah, in PDF format. [15:04:30] <cmihai> Eh, will do. [15:04:41] <cmihai> If not, they still keep stuff on their ftp. [15:04:49] <Atomdrache> Considering I got it without a hard drive cage and with a partially broken off door, a manual would be a surprise ;3 [15:04:52] <cmihai> ftp.digital.com/pub/DEC/Alpha/firmware/readmes/archive/doc/EK-KN450-OP-B01.pdf - AlphaServer 2100 :) [15:05:11] <cmihai> Atomdrache: well, it can boot from an external SCSI array or network. [15:05:23] * WickedWicky still has an Alpha workstation somewhere [15:05:26] <cmihai> Or floppy or CDROM for that matter, so you hardly need a disk inside :-) [15:05:36] <WickedWicky> with an ancient CPU.. EV3 I think.. it runs linux [15:05:45] <cmihai> Plug in a console [15:05:53] <cmihai> 10$ says it freezes probing pci bridge [15:05:58] <Atomdrache> I need to visit Radio Shack to do that. [15:06:12] <Atomdrache> It's got a 9-pin male serial port, and I've got nothing that goes to that. [15:06:19] <cmihai> Eh? [15:06:33] <Atomdrache> Also, since it's male, I can't just MacGuyver something out of random wires like I can with female connectors. [15:06:36] <cmihai> AlphaServer 2100 uses a standard RS-232 5-5 2-3 3-2 cable... [15:06:51] <cmihai> Any old null modem 9600 8 N 1 tip/cu combo will do [15:06:55] <Atomdrache> I see a male 9-pin serial connector. [15:07:12] <cmihai> And you don't have a female huh? [15:07:14] <Atomdrache> I have no cables or adapters which can connect to that, and I'm going to buy one today. [15:07:15] <cmihai> What kind of sysadmin are you? [15:07:22] <cmihai> Oh, right, the common kind :P [15:07:50] <Tempt> Oh come on [15:07:52] <Atomdrache> I'm not a sysadmin. I'm a dork who accumulates creepy old hardware and makes Christmas tree ornaments out of vacuum tubes =P [15:08:01] <cmihai> Don't mind me, no sleep and 12 cans of Pepsi-X do that to a guy... [15:08:03] <Tempt> You must have a 9pin socket somewhere [15:08:13] <cmihai> Yeah [15:08:14] <Atomdrache> (I wouldn't last a day if somebody were paying me to do this stuff :3) [15:08:24] <cmihai> Sacrifice some old mice on the altar of VMS [15:08:26] <Atomdrache> Nah. I've got plenty of others, but not that. [15:08:35] <cmihai> Rip out a motherboard [15:08:36] <cmihai> Do someathing. [15:08:42] <cmihai> erm [15:08:43] <cmihai> Right. [15:08:46] <Atomdrache> Radio Shack opens in like three hours. [15:08:52] <Atomdrache> I think I'll just buy one. [15:09:02] <cmihai> Bah, where's the fun in that :-) [15:09:30] <Atomdrache> I don't have any systems I want to sacrifice for their serial ports. I got rid of my only Intel box. [15:09:55] <Tempt> Or just tack-solder some wires to rx, tx and ground [15:10:08] <Gekkko[PDA]> lol [15:10:14] <Atomdrache> That's disturbing, dude. [15:10:31] <Atomdrache> I can see sticking wires in TX, RX, and ground. I've done that. But...nasty. [15:10:37] <WickedWicky> wait till you hear his CPU-instruction reduction techniques [15:11:01] <Gekkko[PDA]> hammer, chisel [15:11:04] <Gekkko[PDA]> lol [15:11:13] <cmihai> Got pens? [15:11:16] <cmihai> Old ballpoint pens? [15:11:21] <cmihai> They actually make nice pins. [15:11:24] *** calLNCH has quit IRC [15:11:28] <cmihai> If you can't sacrifice some ancient mb or whatever. [15:11:29] *** Vanuatoo_ has quit IRC [15:11:32] <Tempt> It isn't hard to just tack solder some wires on [15:11:43] <Atomdrache> Yeah, but pins don't work on male connectors. That'd be gay or something. [15:11:43] <cmihai> I use ballpoint pen heads :-) [15:11:50] <Tempt> For extra points use a little heatshrink to insulate it [15:12:00] <Atomdrache> I've got some copper tape that might work, though. [15:12:01] <Tempt> Otherwise, got an aligator clips? [15:12:42] <cmihai> Got milk? [15:12:48] <Gekkko[PDA]> lol [15:12:51] <cmihai> Hm.. [15:12:52] <Atomdrache> Nope, no alligator clips. I keep telling myself to go get some, but then I guess I lose my train of thought and exclaim "Ooh, a dead cockroach" or something. [15:13:01] <WickedWicky> I pitty the f00 who doesnt drink milk [15:13:30] <Atomdrache> I should buy those today with the cable and fan grill for my physically hacked SPARCstation 20. [15:13:35] <Atomdrache> (by hack, I mean the kind with a saw) [15:13:43] <Atomdrache> (it will never overheat again) [15:13:56] <WickedWicky> dare I to ask what you did? [15:14:11] <cmihai> Heh, did you know it only takes 2 liters of Powershot to die from caffeine? [15:14:26] <delewis> for anyone that has experience with SNDR/AVS -- how well will async. replication work over a limited bandwidth connection (say, a few Mbps)? [15:14:26] <Gekkko[PDA]> lol p [15:14:33] <Atomdrache> After one of its SIMMs overheated, I cut a couple of holes in the top and bolted on a junction box and a couple of line-voltage case fans I pulled out of a couple of old rack-mount DC power supplies. [15:15:17] <delewis> trying to decide whether or not to use SNDR/AVS or zfs send/recv and SSH. [15:15:23] <Doc> delewis: short answer is "it depends" [15:15:34] <Doc> how much data is being changed? [15:15:45] <delewis> Doc: on the amount of writes, I'd suspect. [15:15:56] <WickedWicky> cmihai: what's powershot? Something like black booster or something? [15:16:08] <delewis> well, not a lot. But the replication set is quite large -- 500GB, roughly. [15:16:08] <Doc> if you've got enough disk you can also do a multi-stage replication - especially if it's the same data being written over and over [15:16:09] <WickedWicky> or highly concentrated coffee? [15:16:09] <cmihai> http://www.energyfiend.com/caffeine-content/powershot [15:16:16] <Doc> ie, original -> II -> SNDR -> II [15:16:20] <delewis> I suppose I need to gather some statistics on how much of the data is being changed on a daily basis. [15:16:22] <Atomdrache> I should take pictures. It's even more gruesome than the replacement cover I built for my SPARCserver. [15:16:34] <delewis> its mostly user's home directories and server backups. [15:16:41] <delewis> so, I'd suspect not *that* much is being changed. [15:17:10] <Doc> remember that if a user opens a file, changes one byte, and then saves it, then they have re-written the entire file [15:17:27] <Doc> that can result in huge changes - especially for things like email [15:17:40] <delewis> Doc: but this will be backed so ZFS, so is it not the case that only the changed blocks will need to be updated? [15:18:01] <Doc> nope. it's not a function of the OS, it's a function of the app [15:18:18] *** StylusEater_Work has joined #opensolaris [15:18:21] <delewis> lovely. would zfs send/recv + snapshots have the same effect? [15:18:46] <Doc> depends on the config [15:19:00] <Doc> if i edit a file 5 times, then so a zfs send/recv, it will only send it once (the latest copy) [15:19:21] <Doc> same if you do snapshots with II and sync then with sndr (but as i said, that needs lots of disks) [15:19:24] <delewis> right, and that's fine. The plan is to create daily snapshots of user's home directories and backups and send them over. [15:19:33] <delewis> that'll give us some level of retention we don't have at the moment. [15:19:42] <Doc> depends if you want real-time-ish replication (ie, SNDR) or fairly lazy replication (ie, ZFS) [15:19:43] <delewis> (we're using rsync for replication and no snapshots) [15:20:08] <ofu> how many files? [15:20:11] <delewis> lazy replication is acceptable, but SNDR would minimize administration/maintenance, I think. [15:20:14] <renihs> ya but snapshots gives you time to backup without having to make them unaviliable [15:20:27] <Doc> nah.. i'd call SNDR higher overhead than ZFS [15:20:28] <renihs> snapshot is done in a second, then you have all time to backup :p [15:20:45] <delewis> Doc: even once its configured and setup? [15:20:50] <Doc> renihs: no different to SNDR/II [15:21:04] <renihs> SNDR? [15:21:09] <renihs> me feels stupid :p [15:21:17] <Doc> you always seem to be kicking it when links go down or servers get reboot and the like [15:21:26] <delewis> but also, SNDR would sync the metadata, as well. using zfs send/recg would require that someone ssh into the offsite replication system and delete the older snapshot that have expired the retention (well, without doing some scripting, anyway) [15:21:50] <Doc> either way you're going to need to script some bits [15:22:10] <delewis> that's true. [15:23:23] <ofu> does Sun sell a sata-jbod with one data-path (no mux-boards neededd) and sas-host-connection? [15:23:54] <jmcp> renihs: iirc, it's StorEdge Network Data Replicator [15:24:01] <Doc> nearest would be the SE3511, unless they have come out with something recently [15:24:13] <renihs> jmcp, ya i figured meanwhile :p [15:24:25] <jmcp> Doc: ST2530 might fit the bill [15:24:27] <Tempt> another blastwave package removed; vim this time. [15:24:37] <Doc> jmcp: that's out now? [15:24:50] <dlg> Tempt: cos it uses libm? [15:24:53] <jmcp> it's been selling for at least a quarter [15:25:04] <Doc> wasnt out when i left [15:25:13] <jmcp> Doc: but no mpxio support with it yet unless you want to run snv_63++ (preferably snv_68) [15:25:14] <Atomdrache> Hmm. Yeah, tests memory before the card slots. [15:25:52] <ofu> jmcp: but 2530 uses sas-disks and has one or 2 hw raid controllers [15:26:18] <Doc> jmcp: http://www.flickr.com/photos/docbert/1103252950/ [15:26:19] <jmcp> ofu: I did qualify my suggestion, and I think you'll find it uses SATA disks inside anyway [15:26:23] <WickedWicky> 3510 is hardware raid capable, no? [15:26:24] <WickedWicky> err [15:26:25] <WickedWicky> 11 [15:26:37] <jmcp> Doc: nice! [15:26:49] <Atomdrache> So if it freezes probing the PCI/EISA bridges, is that most likely indicative of a bad card? [15:26:52] <Tempt> dlg: Naah [15:27:03] <Tempt> I'm just slowly pulling blastwave packages and building my own [15:27:10] *** flowolf has joined #opensolaris [15:27:13] <Atomdrache> I pulled out the PCI cards, but left in the EISA graphics card. Still freezes. Might be that card. I'll try it again. [15:27:14] <flowolf> hi all [15:27:30] <jmcp> Doc: which camera + lens did you use for that shot? [15:27:41] <Doc> 10D + 28-135IS [15:27:49] <jmcp> nod [15:27:50] <Doc> out of a very dirty window [15:28:31] <jmcp> yeah [15:29:01] <cmihai> Atomdrache: could be the bridge too. Then it's dead. [15:29:21] <cmihai> Atomdrache: a serial console or monitor will give you more details, get one. [15:29:25] <jmcp> I thought we removed support for EISA long ago? [15:29:33] <Atomdrache> Yes. If that happens, I'll call a friend who could use the scrap metal. [15:29:45] <Tempt> jmcp: I think we're still talking alpha. [15:29:46] <Atomdrache> But, yeah, I won't really know until I can connect the serial console. [15:30:26] <jmcp> Tempt: http://bugs.opensolaris.org/view_bug.do?bug_id=5080888 Remove EISA nexus driver from S10 [15:30:38] <cmihai> jmcp: it's not Solaris, it's Alpha. [15:30:43] <jmcp> oh, sorry [15:30:55] * jmcp endures a horrendous context shifty [15:30:55] <Tempt> As in suitable for running Tru64 or OpenVMS [15:31:00] * jmcp shudders [15:31:03] * jmcp runs away [15:31:04] <cmihai> Nope, just OpenVMS 7.3. [15:31:08] <cmihai> Tops that is. [15:31:17] <cmihai> I think his comes with 6.2, it's about 1994-ish [15:31:19] <Tempt> Well, there you go. openVMS all 'round [15:31:28] <Gekkko[PDA]> VMS became NT :O [15:31:32] <Atomdrache> Think it's likely at all for the bridge to be dead? [15:31:37] <Tempt> Gekkko[PDA]: Grow up. [15:31:48] <Gekkko[PDA]> lol. [15:32:09] <Atomdrache> Well I hear the same guy who wrote VMS did some work for NT. Forgot his name. Still, I hear VMS rocks and I'm going to try it unless I determine that this machine is dead. [15:32:10] <cmihai> Tempt: nah, he's on to something there. VMS - WNT - see the pattern? IBM - HAL? Go ask Dave Cutler [15:32:38] <cmihai> Atomdrache: it's Dave Cutler.. he made RSX-11, VAXELN, VMS and Windows NT [15:32:42] <Atomdrache> That was his name. [15:33:00] <cmihai> Grab his book, Inside Windows NT [15:33:44] <Tempt> Just grab some VMS and take a look [15:33:51] <Tempt> I know I'm currently looking at getting an ES40 for that purpose [15:34:22] <cmihai> He had this Prism deal back in the day... left DEC for Microsoft and started Windows NT and WinNT Alpha... he worked on 64 bit Windows too (2003,XP, Vista etc). He just hated x86-32 :P [15:34:27] <cmihai> (smart guy) [15:34:38] <Atomdrache> All the cards are out. Didn't come with a hard drive. If it still sits there for several minutes at "Probe I/O Subsys" without doing anything, should I be pessimistic when I connect a serial terminal? [15:35:07] <Atomdrache> Well, yeah. x86 is like Z80 with bigger address width and higher clock speed. [15:35:38] <cmihai> Atomdrache: yep, you should be, probably hosed PCI bridge. [15:35:46] <Atomdrache> Ah, it got past it, [15:35:50] <cmihai> Phew :-) [15:35:55] <Atomdrache> once I yanked the EISA graphics card. [15:35:57] <cmihai> Guess it was the cards after all. [15:35:58] <Atomdrache> "Model 4/200". [15:36:07] <cmihai> Eh, mine are 4/233 [15:36:14] <Tempt> Sweet. [15:36:15] <Atomdrache> Meaning that if it had a hard drive an an OS, something cool would happen. [15:36:19] <cmihai> Well, you don't need VGA anyway, you don't have a monitor. [15:36:20] <Tempt> Digital UNIX [15:36:24] <cmihai> It uses separate signal cables. [15:36:25] <Tempt> or OpenVMS [15:36:42] <Atomdrache> Yeah, I'll just use a serial console. Hell, my SS630MP doesn't even have a video card. [15:36:43] <Tempt> I really want a nice Alpha. [15:36:51] <Tempt> I used to have ES47s at work. [15:36:54] <Tempt> I liked those. [15:36:54] <cmihai> Tempt: to run OpenVMS right? [15:37:03] <cmihai> There are some DL20's that don't run VMS... only Tru64 [15:37:05] <cmihai> Not DEC [15:37:06] <Tempt> cmihai: In this case, yes, to play with VMS [15:37:07] <cmihai> EV7 CPUs [15:37:16] <cmihai> From Cray arrays and such [15:37:25] <Atomdrache> Tempt: Some guy I was talking to has a couple of AlphaStation 200s sitting around and he won't let me have one for under $500. [15:37:35] <Atomdrache> This came up when he said he wanted my Blade 100. [15:37:39] <Tempt> I figure if I learn VMS properly I can be a *real* OS bigot. Just like an Amiga or OS/2 user! [15:37:45] <Atomdrache> (why he wanted THAT of all the Sun hardware in here is beyond me.) [15:37:51] <Tempt> Atomdrache: What a wanker [15:38:01] <Atomdrache> Well...he probably knows that they go for that on eBay. [15:38:08] <Tempt> Atomdrache: A lot of those alphastations only run NT [15:38:12] <Atomdrache> Really? [15:38:15] <Atomdrache> That's crap. [15:38:16] <Tempt> yes. [15:38:22] <Atomdrache> I'm glad I got hold of some big iron, then. [15:38:35] <Tempt> I got VMS up and running on my Multia but it exploded before I got very far [15:38:36] <flyingparchment> or linux ;-) [15:39:00] <WickedWicky> werent you able to run MILO and run Linux on those NT-only thingies? [15:39:09] <flyingparchment> yes [15:39:09] <Tempt> WickedWicky: Why bother? [15:39:09] <Atomdrache> Linux on an Alpha is very much a case of pearls before swine. [15:39:16] <flyingparchment> probably only useful if you happen to have one around though [15:39:17] <Tempt> WickedWicky: Might as well just run Windows, it'll be more stable. [15:39:18] <ofu> what happens when a zpool log device runs full? how can I see if it is full? [15:39:21] <cmihai> You can run Win NT 4 on those Alpha :P [15:39:30] <flyingparchment> win2k beta too [15:39:35] <timeless> and mips [15:39:40] *** sparc-kly has quit IRC [15:39:49] <timeless> technnically only up to nt5b2, no? [15:39:59] * timeless seems to recall alpha was dropped before 2k-b3 [15:40:09] <cmihai> Something along those lines. [15:40:28] <flyingparchment> is there some difference between 2k and NT5 other than the marketing name? [15:40:45] <cmihai> Solaris vs SunOS [15:40:52] <Gekkko[PDA]> why'd they drop Alpha support >.> [15:41:02] <cmihai> DEC died [15:41:04] <timeless> no sales? [15:41:05] <cmihai> Remember? [15:41:09] <cmihai> And nobody used it ffs [15:41:13] <cmihai> The Alpha biggots ran VMS [15:41:16] <Gekkko[PDA]> lol [15:41:18] <cmihai> And they weren't much to begin with [15:41:21] <timeless> support costs money [15:41:23] <cmihai> No software either. [15:41:26] <Tempt> Alpha. [15:41:27] <timeless> time, verification, testing [15:41:28] <cmihai> Win NT on Alpha ran.. Windows. [15:41:29] <cmihai> Nothing else. [15:41:32] <Atomdrache> Sad that DEC went out that way. [15:41:35] <Tempt> A truly glorious architecture. [15:41:41] * timeless nods [15:41:46] <Tempt> I remember fondly the performance from my Alphas. [15:42:07] <Tempt> It was embarrassing how much they outstripped my SPARCs, given it was a "dead architecture" by thenl [15:42:17] <Atomdrache> Just like it's sad that HP, which bought the consumer crap peddler which bought DEC, used to make really awesome spectrum analyzers (hey, maybe they still do) and PA-RISC workstations now makes the kind of crap Compaq made. [15:42:20] <Tempt> Full of speedy performance. [15:42:42] <Tempt> Atomdrache: All the test equipment was spun off to Agilent. [15:42:53] <Atomdrache> Okay, then there's nothing good left in HP. [15:43:09] <jpdrawneek> pretty much [15:43:15] <Atomdrache> Because my having seen "Hewlett Packard" on good test equipment is the only reason I respected HP. [15:43:25] <Atomdrache> And that HP Visualize workstation sitting behind my roommate's couch. [15:43:38] <WickedWicky> HP makes nice laptops though [15:43:41] <cmihai> Remember FX!32 [15:43:56] <cmihai> Alphas were running x86 code faster then x86 :P [15:44:03] <Atomdrache> It's like what I'd think of Sun if they stopped making SPARC machines and decided to discontinue Solaris and replace it with, say, some version of Red Hat Linux. [15:44:04] <jpdrawneek> yep [15:44:33] <Atomdrache> (omitting any mention of SGI) [15:44:43] <Doc> hmm.. i thought that was still internal information at this stage? [15:44:54] <Doc> ohh.. wait.. you said "if". my mistake [15:44:59] <oxygene> heh [15:45:06] <Tempt> I used an HP workstation for a while. [15:45:09] <Tempt> Very zippy. [15:45:13] <oxygene> Doc: sorry, ian leaked some of that [15:45:14] <trochej> Ehh, free software spoils managers [15:45:18] <Tempt> I thought it'd be a dog but it wasn't too bad. [15:46:02] <cmihai> Nah, HP-UX is OK I guess. PA-RISC is fine too. Sure, it's no Alpha or SPARC, but it's not x86 either [15:46:15] <cmihai> Can't say I fancy the "new" Itanium deal though. [15:46:27] <Tempt> I like having architecture diversity [15:46:43] <Tempt> PA-RISC and Alpha and MIPS and SPARC [15:46:44] <cmihai> Well, if it's one thing that's good about Itanium... [15:46:45] <jpdrawneek> ya stops thing getting boring [15:46:50] <Tempt> All good friends to the UNIX community [15:46:51] <cmihai> Is the virtualization. [15:46:57] <cmihai> Get yourself an Integrity server... [15:46:59] <Tempt> not to mention POWER [15:47:03] <oxygene> Tempt: and ARM is the step child :-/ [15:47:04] <Tempt> Another fine CPU. [15:47:06] <cmihai> And you can run OpenVMS, HP-UX and Linux at the same time. [15:47:07] <Atomdrache> I heard Itanium is good for doing a lot of floating-point calculations and reducing your winter heating bills. [15:47:37] <jpdrawneek> only if you write your code for it [15:47:39] <cmihai> add vPars/nPars with HP-UX and you're set. [15:47:58] <jpdrawneek> scientific guys like itanium [15:48:02] <Tempt> I heard Itanium was good for making people remember Alpha and POWER [15:48:17] <cmihai> Ehem [15:48:21] <cmihai> Power is very much alive. [15:48:25] <cmihai> More alive then SPARC I'd say. [15:48:30] <cmihai> Well, apart from the new T2 deal [15:48:35] <Tempt> Oh, POWER is looking shit-hot. [15:48:39] <Atomdrache> IBM uses it in their awesome crazy big iron systems, yes? [15:48:50] <Atomdrache> Are there any POWER workstations? I'd try one. [15:48:51] <Tempt> IBM uses it in 1RU machines too. [15:48:52] <cmihai> Yeah [15:48:56] <Tempt> Intellistation. [15:49:06] <Atomdrache> How does POWER compare to SPARC? [15:49:07] <cmihai> Check out power 6 [15:49:17] *** Gekkko[PDA] has quit IRC [15:49:19] <cmihai> They got the thing to 6Ghz [15:49:23] <jpdrawneek> power 6 has to be compiled for [15:49:43] <Tempt> They're fast. Like, really fast. [15:49:45] <cmihai> The mass production stuff is 4.7Ghz [15:49:45] <jpdrawneek> its only 4.7Ghz [15:49:56] <cmihai> 6 Ghz is in the labs though. [15:50:02] <Tempt> POWER6 is looking like it'll be eating SPARC for breakfast for a while. [15:50:10] <jpdrawneek> we had 10Ghz random chips in labs [15:50:13] <Tempt> For grunt-computing. [15:50:17] <oxygene> cmihai: 6ghz in lab conditions are possible with pentium 4s.. [15:50:28] <Tempt> Still, you can't run Solaris on POWER6, so no fun there. [15:50:33] <cmihai> oxygene: so? [15:50:43] <jpdrawneek> Power 6 need a recompile to get its performace [15:50:44] <cmihai> You'll see 6Ghz Power 6 or Power 7 soon. [15:50:49] <cmihai> Hence AIX 6.1 [15:50:55] <cmihai> With the whole Solaris stolen features [15:51:02] <jpdrawneek> just better :( [15:51:09] <cmihai> You know, virtualization, dynamic tracing and all that [15:51:26] <oxygene> jpdrawneek: how? [15:51:32] <jpdrawneek> with zones that do everything you want + more :( [15:51:53] <jpdrawneek> zone + vmware = ibm zones [15:52:12] <Tempt> Hey. [15:52:14] <jpdrawneek> you can move live zones from box to box [15:52:19] <cmihai> It's virtual partitions.... [15:52:24] <Tempt> The IBM LPAR+WPAR is looking pretty shiny [15:52:32] <cmihai> LPARS ftw [15:52:33] <Tempt> Live migration; that's not a stolen Sun feature. [15:52:39] *** triplah_ has quit IRC [15:52:54] <Tempt> cmihai: I gather even the lowest-end POWER machines do LPARs these days? [15:52:56] <jpdrawneek> well zones + the stuff we want [15:53:10] <cmihai> Sure, Sun has Logical Domains (LDOMS) on UltraSPARC T1... [15:53:13] <cmihai> but.. meh [15:53:19] <jpdrawneek> yep its standard - but thats what you get for a 5k box [15:53:27] <cmihai> Tempt: yeah, they do. Everything newer then Power 4 [15:53:35] <Tempt> That's the one. [15:53:43] <Tempt> I'm really, really thinking about buying one. [15:53:54] <jpdrawneek> yes - then you look at the price [15:53:57] <cmihai> Yeah, but they can get a bit expensive. [15:54:04] <Tempt> p505 is $US3700 list [15:54:23] *** Vanuatoo has joined #opensolaris [15:54:30] <Tempt> That's not that expensive [15:54:34] <Tempt> Same as a V210 [15:54:40] <cmihai> Still, if you want AIX 6.1 and LPARS you want Power 4 minimum. [15:54:49] <Tempt> Actually, a little less than a V210 of similar grunt [15:54:55] <cmihai> Tempt: cheapest I've seen is abou 1000$ [15:54:59] *** karrotx has quit IRC [15:55:02] <jpdrawneek> which one? [15:55:04] <Tempt> cmihai: I'm talking new price. [15:55:13] <cmihai> I'm talking sh :-) [15:55:25] <cmihai> Still, they're awesome machines. [15:56:17] <Atomdrache> Suppose I could find a POWER machine for cheap on eBay, or is that an old Sun thing? [15:56:24] <Tempt> Well [15:56:32] <Atomdrache> I sure as hell can't get old Alphas as cheap as old Suns. [15:56:33] <Tempt> I've been watching ebay for a POWER4 or better machine of late [15:56:37] <Tempt> and there's been sweet fuck all. [15:56:38] <cmihai> Same [15:56:50] <cmihai> You can get Power 3 / AIX 5.3 for as little as 300$ [15:56:51] <Tempt> Hence getting a quote on a p505 [15:56:54] <Tempt> Just buy a new one. [15:56:54] <cmihai> Even 150 [15:57:04] <jpdrawneek> ibm have a long life [15:57:07] <cmihai> But you really need to shell out 1000$ or more for anything that runs AIX 6.1 [15:57:13] <Atomdrache> Eh...that seems a bit much when I can get a completely riced out Ultra 2 for $100. [15:57:18] <cmihai> Tempt: a new one? For _home_ use? :-) [15:57:41] <Tempt> Atomdrache: Sadly, a 1.9Ghz p505 (that's lowest end) would probably cane 8 Ultra-2s without warming up. [15:57:46] <Tempt> cmihai: Yes. [15:57:59] <cmihai> Eh, I guess I can find a few Power 4 we can retire so I can put 6.1 [15:58:03] <Atomdrache> How would that compare to, say, a fully loaded Ultra 80? [15:58:09] <cmihai> Oh please :-) [15:58:45] <Tempt> Atomdrache: that single CPU machine would take a loaded Blade-1000 quite easily. [15:59:03] <cmihai> He's not kidding :-) [15:59:08] <Atomdrache> I'm thinking along the lines of things that are cheap because they're old. [15:59:10] <Tempt> We're talking about a seriously speeeeedy box. [15:59:14] <Atomdrache> Anything older than that? [15:59:34] <cmihai> Power 4 aren't exactly considered old you know. [15:59:36] <Tempt> Atomdrache: If you were local you could take my J50 for a while. It's got 8 x 604e and 3gb of RAM [15:59:36] *** MattAFC is now known as MattMan [15:59:48] <Atomdrache> Holy flaming crap O_o [16:00:05] <cmihai> Just add a storage and punish it with Oracle [16:00:07] <Tempt> and even with those comparitively sluggish 200Mhz CPUs, it slams the transactions through with Oracle. [16:00:10] <Tempt> I did [16:00:13] <Tempt> My DBAs were amazed [16:00:24] <Tempt> They grew a lot of respect for POWER very quickly. [16:00:52] <cmihai> Tempt: we used to be full Power/AIX before HP moved in... some of the ancient IBM systems still perform better then new HP ones. [16:00:58] <cmihai> I mean, same database and all. [16:01:04] <Atomdrache> Would you consider POWER to be an improvement over SPARC machines from the same time? [16:01:10] <Tempt> Yes. [16:01:25] <cmihai> Yeah, but they're way more expensive :-) [16:02:01] <Tempt> I'm not so sure. [16:02:23] <Tempt> I'd really, really love to compare benchmarks on a T2000 compared to the same money spent on Series p [16:02:24] <cmihai> If you're thinking in terms of sockets... they are [16:02:40] <cmihai> Tempt: eh, well, you're thinking in terms of performance :P [16:03:47] <cmihai> Heh, cool, now we have "Smoking will cause a slow and painful death" packs [16:03:47] <Atomdrache> Ooh, I actually see a couple 333MHz workstations on eBay for arond $100. [16:03:56] <Atomdrache> I should get one of those eventually. [16:03:57] <cmihai> Those are Power 3 [16:03:58] <cmihai> They're good. [16:04:21] <Atomdrache> 604e. Good processor? [16:04:24] <cmihai> Will get you up and running with AIX 5.3 and Oracle. Well, will run the latest AIX to date.. apart from 6.1 beta... but now hardware virtualization for you. [16:04:45] <cmihai> 604e you say? [16:04:54] <cmihai> I guess 200Mhz [16:04:55] <Atomdrache> That's what the eBay description says. [16:05:00] <cmihai> Link? [16:05:35] <Atomdrache> It's on the first page if you search for "IBM workstation". I'd give you the link but that'd involve copying the URL between machines, and I haven't got any suitable software for that. [16:05:42] <Atomdrache> I'd type it by hand, but...enormous script-generated URL. [16:05:45] <cmihai> Synaptics [16:05:47] <cmihai> Use that [16:05:52] <cmihai> you can move your mouse and keyboard too [16:06:02] <cmihai> Like a TwinView (Xinerma) but between systems [16:06:16] <cmihai> erm.. [16:06:20] <Atomdrache> Will that have any complaints about different versions of Solaris and different types of framebuffer? [16:06:26] <cmihai> Shit [16:06:29] <cmihai> Synergy [16:06:36] <renihs> ya i was wondering already :p [16:06:43] <renihs> synaptics touchstuff :p [16:06:59] <cmihai> Sorry, no sleep, 8 espressos and 12 cans of energy drink :P [16:07:05] *** hohum_ has joined #OpenSolaris [16:07:06] *** Megaf|Sleep has quit IRC [16:07:15] <cmihai> http://synergy2.sourceforge.net/ [16:07:18] <renihs> cmihai, :p [16:07:23] <cmihai> Trust me, this is COOL [16:07:33] <cmihai> If you use more then one keyboard, you need this. [16:07:36] <Atomdrache> I'll check it out. Hard to compile at all? [16:07:38] <Atomdrache> And yeah, I do. [16:07:53] <cmihai> blastwave has it [16:07:59] <cmihai> Works with Windows and Linux too [16:08:02] <Atomdrache> Oh, okay. Excellent. [16:08:05] <cmihai> Clipboard rocks also [16:08:10] *** hohum- has joined #OpenSolaris [16:08:12] *** Megaf has joined #OpenSolaris [16:08:16] <cmihai> I mean, I can copy/paste from windows to Solaris... ffs :-) [16:08:30] <Megaf> # sh NVIDIA-Solaris-x86-100.14.11.run [16:08:32] <Megaf> /usr/bin/tar: Read 2048 bytes from - [16:08:33] <Atomdrache> Huh...can't find Synergy on Blastwave. [16:08:34] <Megaf> ./install: line 29: UID: readonly variable [16:08:44] <Atomdrache> Wait, there it is. [16:08:49] <cmihai> Megaf: is it a broken download? [16:09:01] *** hohum_ has quit IRC [16:09:06] <cmihai> PS: Megaf OpenSolaris SXCE / SXDE already has NVIDIA drivers (SXCE 69 ones are broken though) [16:09:07] <Megaf> i dont think so [16:09:17] <cmihai> Megaf: make sure, works here. [16:09:31] <Megaf> how/ [16:09:32] <cmihai> If not, edit and see line 29 (duh) [16:09:45] <Megaf> wow [16:09:46] <cmihai> Broken == no GL [16:09:48] <Megaf> glxgears [16:10:03] <cmihai> /usr/X11/demo/glxgears [16:10:09] <cmihai> If that's what you're looking for. [16:10:36] <cmihai> Megaf: once you're done, grab Quake 1/2/3 http://unixsadm.blogspot.com/2007/08/gaming-on-solaris.html and have fun :P [16:10:37] <jpdrawneek> also nvidia drivers which i think you are trying to install are broken - cap on res 1280x1024 or something [16:10:45] <Megaf> $ glx [16:10:46] <Megaf> glxdemo glxgears glxheads glxinfo [16:10:47] <cmihai> jpdrawneek: that's a lie :P [16:10:47] <Megaf> nice [16:10:48] *** hohum- is now known as hohum_ [16:10:53] *** RaD|Tz has joined #opensolaris [16:10:57] <Megaf> cmihai: thanks [16:11:09] <cmihai> jpdrawneek: doing 2560x1024 here just fine :P. [16:11:21] <jpdrawneek> ok - just seen lots or rants about it [16:11:39] <jpdrawneek> and the statement from nvidia thats its a know bug [16:12:08] <WickedWicky> does 1440x900 just fine on my laptop [16:12:14] *** stevel has joined #opensolaris [16:12:15] *** ChanServ sets mode: +o stevel [16:12:18] <Tempt> cmihai: I've got a sneaky suspicion that AIX might be getting a big lift in the industry sometime soon. [16:12:19] <jpdrawneek> is this the first or second 100 series driver? [16:12:32] <cmihai> Tempt: I've got my money on it too. [16:13:01] <Tempt> cmihai: You can see why I'm thinking about buying a machine. [16:14:40] <jpdrawneek> ok lost track of driver revision - think the issues was in 100.14.09ish [16:16:46] <StylusEater_Work> are sparc stations now using DVI instead of traditional serial? [16:17:31] <jpdrawneek> if you take the card out you can use serial port [16:18:00] <jpdrawneek> had to do that when all i had was a cheap mointor [16:18:41] *** jpdrawneek has quit IRC [16:18:53] <Atomdrache> I just installed synergy on two of my Sun boxes. I can't find where to run it, and I can't find a man page. [16:19:03] <Atomdrache> Is there some special command or location I should know about? [16:19:36] *** sommerfeld has joined #opensolaris [16:19:37] *** ChanServ sets mode: +o sommerfeld [16:19:40] <Megaf> cmihai: i dont have the pak0.pk [16:19:46] <Atomdrache> Oh. I guess it's "synergys". [16:20:12] <cmihai> Megaf: download it, follow the link, it's the other package. [16:20:19] <cmihai> That's all you need to play online and with bots. [16:20:38] *** vmlemon has joined #opensolaris [16:21:49] <Megaf> can i install this software on OpenSolaris/ http://aria2.sourceforge.net/ [16:23:41] <quasi> no [16:24:37] <Atomdrache> Oh, sweet. [16:24:39] <Atomdrache> This rocks :D [16:24:51] <StylusEater_Work> what does? [16:24:55] <Atomdrache> Synergy. [16:25:15] <Atomdrache> It just weirds me out to move a mouse from a sun4u machine with SXCE to a sun4m with Solaris 9 and an eight-bit framebuffer. [16:25:33] <Atomdrache> It's creepy. And it rocks. [16:25:42] <Tempt> Try moving from Windows to Solaris [16:26:27] <StylusEater_Work> interesting [16:27:58] *** cmihai has quit IRC [16:28:55] <solar-star> Help! Solaris reboots all the time [16:29:04] <solar-star> Installed OSS [16:29:11] <Tempt> That's the good thing about Suns. They boot quickly. [16:29:51] <renihs> heh :p [16:29:57] <renihs> if you disable everything :p [16:29:59] <solar-star> Can I prevent Solaris from rebooting [16:30:18] <Atomdrache> Useless reply: Yes, by fixing it. [16:30:25] <Atomdrache> (I've got nothin'.) [16:30:26] <Tempt> Atomdrache++ [16:30:48] <Atomdrache> Any interesting errors in /var/adm/messages? [16:31:10] <solar-star> I do not come to that place [16:31:14] <Tempt> and in /var/crash [16:31:29] <solar-star> Immediatly after log-in. Solaris reboots [16:31:30] <Atomdrache> Also try logging in with command-line mode. [16:31:43] <solar-star> That is what I am doing [16:31:48] <solar-star> command-line [16:31:57] <solar-star> I recently installled OSS [16:32:07] <Atomdrache> Hmm...try booting into single-user mode? [16:32:18] <solar-star> Did not try yet [16:32:27] <Atomdrache> (stop-A and boot -s if you're using a SPARC) [16:32:46] <solar-star> amd64 [16:33:00] <oxygene> then edit the grub entry and add "-s" to the kernel line [16:33:10] <Atomdrache> Oh. There's probably a boot menu if you mash the right function key while the machine's--what he said. [16:33:29] <oxygene> that function key is "e" (as in "e"dit) [16:33:36] <solar-star> Ok [16:33:39] *** smellyfish has joined #opensolaris [16:33:42] <oxygene> documented on-screen [16:33:51] <solar-star> I already have an entry with -s [16:34:02] <Atomdrache> Then use that one. [16:34:07] <solar-star> It is just now booting ... [16:34:41] *** Giaco has quit IRC [16:34:42] <solar-star> The last thing I see is something with dumping ... [16:34:47] <WickedWicky> if the reboots started after installing oss then -s prolly wont do a lot of help either since it can be a kernel module giving him an attitude, no? [16:34:50] *** EchoBinary has joined #opensolaris [16:35:06] *** EchoBinary has quit IRC [16:35:20] <Atomdrache> Eh, maybe. I don't know about that stuff. [16:35:22] *** Tpenta has joined #opensolaris [16:35:30] <Atomdrache> Though it might not load all the modules in single-user mode. [16:35:33] <WickedWicky> try to boot from the installation cd, choose single user shell [16:35:34] <WickedWicky> cd /a [16:35:38] <WickedWicky> chroot . /bin/bash [16:35:45] <WickedWicky> pkgrm <oss package> [16:35:48] <WickedWicky> and reboot [16:35:50] <solar-star> So I am in. [16:36:00] <solar-star> as a single user [16:36:04] <Megaf> nice site http://www.blastwave.org/ =) [16:36:11] <WickedWicky> reverse your last actions [16:36:24] <solar-star> with pkgrm ? [16:36:33] <quasi> WickedWicky: instead of chroot, you could just -R /a [16:36:42] <WickedWicky> really? cool [16:37:25] <quasi> works for the pkg and patch tools [16:38:19] <WickedWicky> nice [16:39:07] <oxygene> has the advantage of not failing if the cd and on-disk libc disagree on kernel interfaces [16:39:14] *** Jondice has quit IRC [16:39:21] * WickedWicky takes notes [16:40:11] *** Belgar has joined #opensolaris [16:41:11] *** alanc_away has quit IRC [16:42:29] *** Shiv__ has joined #opensolaris [16:43:10] <Shiv__> Hi Laca, tomww [16:43:20] <tomww> hi Shiv__ :-) [16:44:26] <Shiv__> tomww: Not sure on which mailing lists you are. I sent an update mail about the event on desktop discuss... [16:45:38] <tomww> Shiv__: I've read your mail with the 35 spec files ... [16:46:33] <Shiv__> tomww: Yes that was the one. It was nice to see you and laca online on saturday. [16:47:13] <tomww> It was very nice at least for me too, and I think it was the same for laca. [16:48:13] <Shiv__> tomww: But we had considerable things happening there, too many minor hiccups coming in, so ability to collaborate online was limited. [16:48:20] <Shiv__> tomww: so I left co-ordinating on IRC to Ananth. [16:48:29] *** xushi_ has quit IRC [16:49:31] <tomww> I can imagine, that this has been hard work for the one already knowing how to use pkgtool/pkgbuild/spec-files ... [16:50:19] <solar-star> Thanks for your all help [16:50:30] <solar-star> pkgrm did stabilize the system [16:51:21] <Atomdrache> Excellent. [16:51:29] <solar-star> OSS fails while installing [16:51:44] <solar-star> and then the system is somehow buggy [16:51:49] <Atomdrache> (Also, OSS is crap and I wish there were an alternative besides Sun's native sound system.) [16:52:12] <solar-star> any attempt to reinstall fails [16:52:26] <Atomdrache> (well, I hear the new version isn't, but...what fun I've had with it.) [16:52:37] <Atomdrache> Details on how it fails? [16:52:53] <solar-star> The system reboots all the time [16:53:01] <Atomdrache> I mean the reinstall. [16:53:12] <Atomdrache> Does the installation program say anything's wrong? [16:54:07] <solar-star> I think due to the incomplete installation the kernel is somehow crashed [16:54:35] <solar-star> Maybe I should compile the package myself? [16:54:52] <solar-star> What new version did you mention? [16:55:20] *** Jondice has joined #opensolaris [16:55:23] <Atomdrache> Whichever one is new enough that it'll let multiple programs access the same sound device at once. [16:55:30] <Atomdrache> There may be more than one such version by now :3 [16:55:45] <solar-star> Is there an alternative to OSS? [16:56:00] <Atomdrache> On Solaris? No, unless you're okay with the Sun sound system. [16:56:19] <solar-star> How to attach the latter to my device? [16:56:31] <Atomdrache> Install Solaris, usually. [16:56:56] <solar-star> Is the driver called 'audiohd'? [16:56:57] <Atomdrache> And if that doesn't work, err...can't help you. I would be utterly confuzzled. [16:57:04] <Atomdrache> I don't know. [16:57:23] <tomww> Shiv__: I'll have a look on the spec files created. Any plans to feed them into spec-files-extra? [16:57:43] <solar-star> I attached 'audiohd' to '"pci1002,437b"' [16:58:13] <solar-star> The system said: Successfully installed, but failed to attach [16:58:34] <solar-star> So I found out that some codecs could not be computed [16:59:09] <solar-star> And this was the reason I tried OSS [17:00:57] <laca> hi Shiv__, yes i saw your mail [17:00:59] <laca> good stuff [17:15:28] *** Shiv_ has joined #opensolaris [17:15:55] *** Shiv_ has quit IRC [17:22:50] *** Belgar has left #opensolaris [17:26:26] *** Shiv__ has quit IRC [17:30:17] *** trs81 has quit IRC [17:30:24] *** trs81 has joined #opensolaris [17:30:40] <axisys> ippool -A -m 300 does not add a new pool and exit with `1' . ippool -l does not show new pool [17:30:59] <axisys> it is on sol 10 u3, am I experiencing a bug? [17:31:34] <quasi> axisys: no, you're experiencing a wrong channel moement - this is #opensolaris ;) [17:32:03] <nikinana> sol 10 u3...... [17:32:37] <nikinana> sol 10 u4 will release this month [17:33:50] <quasi> and yet neither of those are == opensolaris [17:35:33] <kaiwai> nikinana: best go check out #solaris [17:37:01] <Tempt> quasi: Don't be a ratbag. [17:37:18] <quasi> Tempt: I'm just teasing axisys [17:37:19] <Tempt> If you don't have a solution, say ... nothing. [17:37:50] <Tempt> Oh, hang on. [17:38:03] <Tempt> Giving shit to axisys is completely legit [17:38:08] <Tempt> axisys: Go ask somewhere else! ;) [17:38:18] *** tsoome has joined #opensolaris [17:38:36] <tomww> #define opensolaris friendlychannel [17:39:15] <Tempt> s/friendly/rowdy/ [17:39:39] <Tempt> axisys: Tried adding -v to see if it spits anything more useful out? [17:40:06] <quasi> tomww: it is already set to that - 'get lost', 'rtfm' or '/ban' would have been the answer in a !friendlychannel ;) [17:41:04] <WickedWicky> will you still be friendly to me when i tell you I removed kernel32.exe from my freshly installed SXCE installation? WHAT TO DO?! WHAT TO DO?! ACK ACK [17:41:04] <axisys> quasi: someone having a pms moment? sorry did not mean to burst your bubbles [17:41:18] <WickedWicky> PMS is a myth dude [17:41:28] <WickedWicky> it's just an excuse we bought for them being a b*tch [17:41:40] <axisys> WickedWicky: hehe [17:41:42] <quasi> axisys: well, most times you don't bother reading man pages ... [17:42:20] <axisys> quasi: now you are talking .. according to man I could add a new pool with that command btw [17:43:58] <axisys> Tempt: `-v' does not give anything [17:44:04] <Tempt> crap. [17:44:09] <Tempt> That's no fun. [17:44:12] <Tempt> truss it? [17:45:51] <axisys> Tempt: http://rafb.net/p/aoWIk485.html [17:46:56] <quasi> and you're not missing the original enabling of pools and boot? [17:47:07] <Tempt> OH NOES [17:47:13] <Tempt> I've spent too much time looking at truss today. [17:47:20] <axisys> Tempt: hehe [17:47:25] <Tempt> My eyeballs will fall out if I look at any more [17:47:28] *** sparc-kly has joined #opensolaris [17:47:28] *** ChanServ sets mode: +o sparc-kly [17:47:31] <quasi> and this bit of truss looks pretty much useless [17:47:33] <axisys> quasi: ippool -l shows the original pools still there [17:47:53] <axisys> let me see if I can find any bug in sunsolve [17:53:04] *** Erwann has quit IRC [17:53:43] *** comay has joined #opensolaris [17:54:49] <axisys> i were able to add a pool like this `ippool -A -dv -m 300 -o ipf -t pool' [17:55:06] *** calumb has joined #opensolaris [17:57:10] *** dlynes has joined #opensolaris [18:00:37] <CIA-17> seb: 6479886 mipagent and related kernel interfaces should be removed (fix unref) [18:00:38] *** dlynes has joined #opensolaris [18:01:05] <axisys> so the following two command worked... `ippool -A -dv -m 300 -o ipf -t pool' and `ippool -a -m 300 -o ipf -i 1.2.3.4/32' - but removing the entry failed with this message `remove_pool:SIOCLOOKUPDELNODE: No such file or directory' [18:01:06] *** dlynes has quit IRC [18:02:23] <axisys> i used this command to remove `ippool -r -m 300 -o ipf -i 1.2.3.4/32' [18:02:50] <axisys> according to sunsolve bug http://sunsolve.sun.com/search/document.do?assetkey=1-1-5038684-1 there is problem with the ippool code [18:05:42] *** dlynes has joined #opensolaris [18:05:43] *** Jondice has quit IRC [18:05:48] *** Jondice_ver_2 has joined #opensolaris [18:06:21] *** reidms has quit IRC [18:07:48] <Tempt> axisys: Is there a patch? [18:10:55] *** Megaf has quit IRC [18:11:20] *** gnut has joined #opensolaris [18:11:23] <gnut> hi all [18:12:19] <axisys> Tempt: there are quite a few bugs.. looking for it [18:12:50] <gnut> I'm using dtrace to figure out what's making firefox-bin so slow. is it normal for firefox-bin to be in LCK (according to prstat -m) 75% of the time? [18:13:52] <asyd> your home dir is on nfs ? [18:15:18] *** stevel has quit IRC [18:17:44] <timeless> gnut: erm [18:17:54] <timeless> firefox has on average 6-10 threads [18:18:01] <timeless> plus 10 java threads [18:18:09] * flowolf reboot [18:18:11] *** flowolf has quit IRC [18:18:20] *** stevel has joined #opensolaris [18:18:21] <timeless> so if you're doing something that's counting all threads, then yes, you're doing something really silly [18:18:21] *** ChanServ sets mode: +o stevel [18:18:25] <ofu> *lol* how do I disable caching of ZFS? [18:18:27] *** stevel has left #opensolaris [18:18:34] <ofu> my "disk" is too fast [18:18:48] <aruiz> ofu, zfs set [18:18:49] <asyd> hmm i don't remember the exact definition of user lock, mean it waits for another user process? [18:19:12] <timeless> it could be x communication... [18:19:17] <gnut> timeless: hmm.. okay. [18:19:21] <gnut> timeless: I just did prstat -m [18:19:26] <gnut> let me try it with -mL [18:19:31] *** mikefut has quit IRC [18:19:32] <ofu> zfs set? What should I set? [18:19:37] <Tempt> firefox is just dead slow on Solaris [18:19:42] <Tempt> Use opera, it's roaring fast. [18:19:43] <timeless> really? [18:19:48] <asyd> ofu: why you want that? [18:19:52] <timeless> do you guys have actual use cases? [18:19:53] <Tempt> timeless: By comparison, yes. [18:19:54] <gnut> firefox used to be fast when i was using b65 [18:20:00] <gnut> now on b69 it's much slower [18:20:00] * timeless hasn't had problems w/ firefox on solaris [18:20:08] <Tempt> Okay. [18:20:09] * aruiz has [18:20:11] <gnut> BUT it is much much faster (usable) on another machine of mine [18:20:23] <Tempt> Firefox is dead slow on my host running Solaris 10 on SPARC [18:20:23] <ofu> asyd: because my iRAM is too fast [18:20:31] <aruiz> timeless, on SPARC is slooooow :) [18:20:36] <gnut> prstat -mL isn't the right thing. it shows that each thread is in LCK almost 95% of the time [18:20:42] <Tempt> Opera on SPARC is blazing fast. [18:20:53] <Tempt> Hence, my suggestion to use Opera. [18:20:57] <aruiz> Tempt, is there Opera/SPARC/Solaris version? [18:21:02] <Tempt> Yes, there is. [18:21:09] <aruiz> cool [18:21:12] *** Megaf has joined #OpenSolaris [18:21:13] <Tempt> And they've got a static linked version now so you don't have to mess with qt versions [18:21:30] <timeless> do they actually have dynamic versions? [18:21:38] <Tempt> I was using Opera/SPARC over SSGD today at work because it was faster than running Opera on Windows. [18:21:42] <Tempt> timeless: Yes, they do. [18:21:52] <timeless> ssgd? [18:21:57] <gnut> do they have opera/solaris/x86? [18:22:02] <gnut> i will go see [18:22:02] <timeless> yes [18:22:05] <asyd> ofu: I know you can decrease the kernel memory used by arc (the zfs cache) but I don't think you can disable it, and anyway it's a bad idea i think [18:22:11] <aruiz> my friends at Oslo aren't going to believe this [18:22:14] <gnut> does it have ads? [18:22:22] <timeless> opera doesn't have ads <period> [18:22:23] *** Megaf has quit IRC [18:22:31] *** Megaf has joined #OpenSolaris [18:22:33] <timeless> you've been out of the world too long [18:22:43] <gnut> timeless: hmm.. yeah. out of the opera world. [18:22:57] <aruiz> 1200 Kb/s wee! [18:23:27] <Tempt> Opera lost the ads .. what ... 4 years ago? [18:23:36] <Cyrille> at least three [18:23:41] <Tempt> timeless: secure global desktop [18:23:48] <ofu> asyd: I can do 115MByte/sec, no matter how many ios that are [18:23:52] <gnut> okay... downloading [18:23:53] <Tempt> basically running it over remote X from home to work [18:24:02] <timeless> tempt; an rdp competitor i suppose [18:24:12] <Tempt> timeless: Not really. [18:24:13] <Cyrille> more than that [18:24:20] <timeless> more how? [18:24:27] <Tempt> timeless: SSGD handles RDP along with X and ssh and telnet and 3270 [18:24:45] <Tempt> timeless: Head over to sun.com and have a look - I believe there is still a demo server to look at. [18:24:46] <Cyrille> it can aggregate access to various applications provided by different platforms into a single desktop [18:24:51] * timeless can't remember ever needing/understanding 3270 [18:24:57] <Tempt> And run the whole thing over https [18:25:16] <Cyrille> that's an implementation detail ;-) [18:25:19] <timeless> istrc at least an activex control mstsc rdp client [18:25:35] <nikinana> does gcc 4.0 have x86 version run on nv? [18:27:03] <aruiz> ugh [18:27:06] <aruiz> this is ugly [18:27:13] <aruiz> no antialiasing [18:27:14] <aruiz> :( [18:27:54] *** czubin has joined #opensolaris [18:28:16] <ofu> wow, no more performance penalties with zfs and nfs-exports [18:28:24] <ofu> *evillol* [18:29:31] *** RaD|Tz has quit IRC [18:29:48] <RElling> ofu: test results? [18:30:03] <ofu> I will blog about it after dinner [18:30:12] <RElling> thanks [18:30:59] <ofu> http://www2.fuckner.net/~ofu/performance-log-device-nfs [18:31:15] <gnut> i'm downloading opera as we speak, but does it support flash/java as easily? or do I have to move the flash plugins into some directory? [18:31:16] <ofu> i only use one disk as primary... [18:31:40] <ofu> c1d0 is a samsung hd501lj sata, c3d0 my 2gbyte iram [18:32:08] <asyd> iram? [18:32:21] <RElling> it is more fun taking a gun to a knife fight :-) [18:32:34] <ofu> http://www.gigabyte.com.tw/Products/Storage/Products_Overview.aspx?ProductID=2180&ProductName=GC-RAMDISK [18:32:36] <holcomb> mmmm slog [18:32:59] <holcomb> how much is one of those cards? [18:33:23] <ofu> i payed 170E for it, including shipping hk -> de (without RAM) [18:33:49] <holcomb> nice [18:34:05] <ofu> can I snapshot the slog-device (for backup-purposes)? [18:34:24] *** Yamazaki-kun has joined #opensolaris [18:34:31] <RElling> snapshot should commit the slogs, so they aren't needed [18:34:41] *** lisppaste3 has quit IRC [18:34:52] <RElling> you should try a create-intensive test: tar -x on lots of little files [18:35:00] <sommerfeld> ofu: so that critter uses PCI for power and a SATA port for all I/O so no special drivers needed? [18:35:12] <ofu> no, I got the "dvd-sized" one [18:35:16] <gnut> aruiz: were you referring to opera? [18:35:23] *** Yamazaki-kun has quit IRC [18:35:31] <ofu> it uses a normal 5 1/4 power connector and a sata-port [18:35:35] *** Yamazaki-kun has joined #opensolaris [18:35:42] <aruiz> gnut, yes, but it was just a matter of the fonts chosen by the css, it works pretty well [18:35:46] *** Yamazaki-kun has joined #opensolaris [18:35:52] <ofu> 2. c3d0 <GIGABYTE-2C3BB6329781C921-0001-2.00GB> [18:36:03] <ofu> thats what format is telling me [18:36:12] <asyd> funny [18:36:28] <RElling> interesting that the slog was generating so much I/O... need to track the highwater mark [18:36:49] <WickedWicky> http://news.bbc.co.uk/2/hi/uk_news/england/wear/6944026.stm [18:37:06] <WickedWicky> how much espresso do you have to drink to OD? [18:37:13] <gnut> aruiz: yeah. my antialiasing on that works, too. [18:37:25] <gnut> well, it seems like opera works, but the slowness is still there [18:37:32] <gnut> so maybe its not firefox/opera, but Xorg? [18:37:36] <aruiz> :) [18:37:44] <asyd> WickedWicky: lol [18:37:45] <aruiz> gnut, you can dtrace xorg too [18:37:46] <aruiz> :P [18:37:47] <gnut> prstat shows my cpu usage soar up to around 30%-40% [18:37:47] <Yamazaki-kun> WickedWicky: I've had 2 doubles and three cans of Diet Coke in one day and got somewhat twitchy. [18:37:48] <g4lt-mordant> WickedWicky, more than you can befor puking [18:37:57] <gnut> aruiz: i've been playing around with dtrace a lot lately [18:38:06] <aruiz> gnut, you might want to poke on #dtrace [18:38:20] <gnut> aruiz: thakns for the tip [18:38:42] <Tempt> 'night all [18:39:26] <kaiwai> hmm, firefox always sucks, no matter what platform it runs on [18:39:35] <timeless> oh, g, thanks [18:39:37] <WickedWicky> so now girls found another way to be EMO and cranky... first there was PMS.. now there is espresso OD [18:39:42] <timeless> how about constructive criticism? [18:39:58] <LeftWing> timeless: kaiwai is never constructive, no matter what software he's talking about [18:39:59] <WickedWicky> I like firefox, especialy the 2.x series [18:40:08] <WickedWicky> what makes it sucky are the 0984309430943094 plugins people install to render IE pages [18:40:26] <kaiwai> OD'ed on caffeine - after 7 cups? geeze, how whimpy [18:40:44] <kaiwai> LeftWing: explain why it leaks memory like a sieve [18:40:45] <timeless> wickedwicky: *shrug*, I install noscript and get back to work [18:40:50] <WickedWicky> she should go to portugal or croatia and drink coffee there, she'll cry her guts out [18:41:06] *** yippi has joined #opensolaris [18:41:17] <kaiwai> I would have 5 cups of triple strength coffee when i was at work [18:41:39] <RElling> I used to inject it directly into my cranium [18:41:54] <LeftWing> kaiwai: Seems to be running OK here. [18:42:21] <kaiwai> I've got 2.0.0.6 - sluggish to say the least [18:42:32] <timeless> how do you use it? [18:42:38] <timeless> 1000 tabs, or just 600? [18:42:39] <LeftWing> Sluggish and memory leaks are not the same thing. [18:43:43] <WickedWicky> well, memory leaks I cant defend.. there are plenty in FF [18:43:44] <LeftWing> 2.* seems to explode less than 1.5* did. [18:43:44] <timeless> kaiwai: seriouusly? [18:43:44] <WickedWicky> that's not really hard to acomplish though.. [18:43:46] <kaiwai> timeless: yeap [18:43:56] <timeless> kaiwai: i was looking for a number.. [18:44:11] <kaiwai> LeftWing: when the memory usage just keeps growing at a faster rate than my waist line - its pretty freaky going ons [18:44:14] <richlowe> timeless: good luck [18:44:18] <kaiwai> normally I have 10 tabs open [18:44:21] <timeless> wickedwicky: do you use java and flash, or just java? [18:44:22] <LeftWing> Oh no, another useless analogy. [18:44:26] *** RaD|Tz has joined #opensolaris [18:44:35] <WickedWicky> timeless: both [18:44:42] <WickedWicky> that is: java and flash [18:44:48] <kaiwai> LeftWing: aww, what me to call you a waaaaaahbulance? [18:44:51] <LeftWing> timeless: I would like to use Java, but it frequently does not go well. Flash isn't much better, unfortunately. [18:44:57] * kaiwai rings whine-one-one [18:45:14] <timeless> waistlines don't grow that fast [18:45:24] <timeless> so if ff is growing slightly faster than that, it's in good shape [18:45:29] <timeless> :) [18:45:34] *** aruiz has quit IRC [18:45:34] <WickedWicky> crap [18:45:41] <WickedWicky> and of course the dutch radio goes beserk on it [18:45:43] <delewis> you know, there's a reason IRC serves have channel ban functionality. [18:45:48] <delewis> servers, rather. [18:45:51] <WickedWicky> "And now.. coffee addiction? how bad is it and how do you recognize it?" [18:45:59] <LeftWing> delewis: Unfortunately nobody seems up to it. [18:46:00] * WickedWicky starts to cry [18:46:09] <delewis> LeftWing: unfortunately. [18:46:22] <timeless> wickedwicky/leftwing: so, do you guys collect cores when ff crashes? [18:46:35] <LeftWing> delewis: So, instead, we have the Waaah-bulance. Which I honestly though had more M's in it. [18:46:36] <WickedWicky> timeless: honestly, no [18:46:44] <timeless> wickedwicky: coreadm [18:46:44] <WickedWicky> but I have to tell you also, I never had it coredump [18:46:45] <LeftWing> timeless: Would you like me to? [18:46:54] <WickedWicky> it just gets very very slow, so I kill -9 it [18:46:55] <WickedWicky> and restart [18:46:56] <timeless> leftwing: if you want to complain, yes [18:47:02] <LeftWing> timeless: Fair cop. :) [18:47:05] *** calumb has quit IRC [18:47:15] <LeftWing> I usually pstack them and then throw them away [18:47:17] *** nostoi has joined #opensolaris [18:47:18] <WickedWicky> I am not complaining, somebody asked how firefox was [18:47:23] <WickedWicky> so I gave my opinion [18:47:25] <WickedWicky> jeez [18:47:44] <WickedWicky> I even said how it's not firefox itself but the plugins [18:47:57] <LeftWing> WickedWicky: I don't think timeless is attacking you. =P [18:47:59] *** gnut has quit IRC [18:48:06] *** ichigo has quit IRC [18:48:19] <WickedWicky> oh [18:48:20] <WickedWicky> ok. [18:48:21] <WickedWicky> sorry :P [18:48:29] * WickedWicky can be a tad impuslive [18:48:34] <WickedWicky> impulsive too [18:49:12] <g4lt-mordant> WickedWicky, perhaps you DO need to cut back on the caffeine... ;P [18:49:17] <WickedWicky> lol [18:49:33] <WickedWicky> maybe I didnt have enough today. ;-) [18:50:07] * WickedWicky is running memtest for 6 hours straight at the moment [18:50:42] <WickedWicky> I just hope one of the DIMMS was not sitting correctly in its bank... testing the DIMMS individualy doesnt give any errors [18:51:13] <oxygene> heh, pentium4 overclocked to 8.3ghz.. so much for the "power6 at 6ghz (in the lab)" talk [18:51:26] <WickedWicky> when I had the 3 DIMMS in their banks on the mainboard I got errors from the 2048.48MB range towards 2400MB [18:51:30] *** Cyrille has quit IRC [18:51:35] <WickedWicky> serious? link? [18:53:58] *** mexcellent has joined #opensolaris [18:54:20] *** mexcellent has quit IRC [18:54:20] <oxygene> WickedWicky: http://www.nordichardware.com/news,6678.html [18:54:40] *** slowhog has joined #opensolaris [18:54:49] <kaiwai> WickedWicky: I've gotta stop going near your machine - I kill dimms :) [18:56:48] <sommerfeld> oxygene: use of liquid nitrogen kinda makes it out of reach for production systems. [18:57:55] <WickedWicky> didnt they do something similar with cooling a CPU to -34 before boot up, with freon? [18:57:56] <oxygene> sommerfeld: well, that's lab conditions [18:58:50] <Stric> Ah. TSM client 5.4.1.2 supports ZFS with ACLs.. [18:59:14] *** peteh has quit IRC [19:01:07] *** brendang has joined #opensolaris [19:01:40] *** damienc has quit IRC [19:02:31] <richlowe> mornin' brendang. [19:04:04] <ofu> creating 100000 dirs takes 160 seconds :( [19:04:53] <brendang> G'Day [19:05:08] <mustang> wow, it lives. hi brendang ;) [19:07:50] *** bnitz has left #opensolaris [19:07:57] <moazamraja> re [19:10:05] <g4lt-mordant> ofu then don't create 10K dirs ;P [19:10:56] <WickedWicky> what's adviceable? 3 DIMMS on myu mainboard gives me Errors in a certain range, testing every DIMM independantly gives nothing but OK. Just call it a wrong seated module or would you recommended to test all banks? [19:11:23] <Stric> WickedWicky: re-seat them and see if it goes away [19:11:56] <sommerfeld> wickedwicky: 3 out of how many DIMMs? [19:12:04] <WickedWicky> 3 [19:12:12] <WickedWicky> no its one dimm that gives an error [19:12:30] <WickedWicky> 1/3 gives an error, testing all 3 DIMMS individualy doesnt give any error [19:12:40] <sommerfeld> try all possible 2/3 combinations [19:12:52] <g4lt-mordant> is it the socket that might be bad? [19:12:59] <sommerfeld> that, too [19:13:24] *** gdamore has joined #opensolaris [19:13:38] <WickedWicky> okies [19:14:25] <sommerfeld> once had to perform that exercise on an 8-DIMM system (turned out, after much dimm swapping, that 6 of the 8 DIMMs I bought were bad in different ways) [19:15:38] *** nikinana has quit IRC [19:15:55] *** nikinana has joined #opensolaris [19:15:58] <ofu> reminds me of my first encounter with ibm ddys disks: 13 of 13 disks dead [19:16:58] *** jfndi has joined #opensolaris [19:17:08] <WickedWicky> 6/8? that's gotta hurt [19:18:12] *** Jondice_ver_2 has quit IRC [19:18:15] *** bondolo has joined #opensolaris [19:18:28] *** SYS64738 has joined #opensolaris [19:18:33] <SYS64738> hi [19:18:36] <sommerfeld> Yup. after an hour flailing around I got out a sharpie and put numbers on each DIMM [19:18:57] <sommerfeld> one DIMM was completely toxic (if it was in the system, system didn't come up) [19:19:13] <sommerfeld> others spewed ECC errors at various rates [19:20:14] *** Jondice has joined #opensolaris [19:20:53] <SYS64738> what nic is e1000g ? [19:21:20] *** mikefut has joined #opensolaris [19:21:25] <wesolows> It's, um, e1000g. [19:21:26] <ofu> SYS64738: Intel [19:21:31] <wesolows> Intel 8254x generally. [19:21:41] <SYS64738> is it supported ? [19:21:42] <sommerfeld> SYS64738: e1000g drives many intel gigabit chips [19:21:45] <wesolows> Of course. [19:22:01] <SYS64738> doenst work [19:22:22] <wesolows> Well, with that bug report in hand I'll have it fixed in a jiffy. [19:22:31] <WickedWicky> try prtconf -v | less [19:22:34] <WickedWicky> look for "Ether" [19:23:12] <WickedWicky> note the pci address, (like 8086.1092 or pci1095.2010 , whatever) and add a line to /etc/driver_aliasees [19:23:41] <SYS64738> ifconfig -a shows it to me [19:23:44] <ofu> oh, btw: those Intel gigabit-nics, that are part of the ich9-chipset, are they supported? [19:23:44] <WickedWicky> oh [19:23:48] <WickedWicky> so the driver IS working [19:24:15] * WickedWicky goes back in the times where he used to work for an internet service desk [19:24:16] <ofu> I got pci8086,294c and it wont work on nv70, even after update_drv [19:24:18] <SYS64738> but nwam service doesn't start [19:24:19] <WickedWicky> what exactly isnt working? [19:24:44] <WickedWicky> did you bring the interface up after nwam was started? [19:24:57] <WickedWicky> and do you want to use DHCP? [19:24:58] <SYS64738> nwam doesn't start [19:25:10] <SYS64738> I set the ip with ifconfig [19:25:22] <SYS64738> I reach to ping others pc's [19:25:25] <WickedWicky> I think nwam does DHCP configuration [19:25:26] *** yippi has quit IRC [19:25:30] <SYS64738> but service like sshd doens't start [19:25:39] <ofu> svcs -x [19:25:43] <WickedWicky> (I am not sure though) [19:26:07] <SYS64738> now I plumbed and set the ip to the bge0 [19:26:11] <WickedWicky> what happens when you try to start sshd manualy? surely something is written to the log [19:26:27] <SYS64738> I take a look [19:27:45] <richlowe> WickedWicky: I think you can tell the phase0 bits that you'd rather it not, but by default it does. [19:28:13] <WickedWicky> in reply to nwam + doing dhcp? [19:28:19] <richlowe> Yeah. [19:28:22] <WickedWicky> ah [19:29:02] <SYS64738> I followed the link in svcs -vx and:" [19:29:02] <SYS64738> bash-3.00# cat /etc/nwam/11p [19:29:02] <SYS64738> e1000g0 172.16.0.241 255.255.255.0 [19:29:05] <WickedWicky> when I want to use static addresses I do: svcadm disable network/physical:nwam && svcadm enable network/physical:default [19:30:07] <WickedWicky> SYS64738: I am afraid I used nwam too view to help you with this one, sorry :s [19:30:26] <SYS64738> ok [19:30:45] <WickedWicky> did you try svcadm enable ssh ? [19:30:56] <SYS64738> yes [19:31:01] <WickedWicky> i? [19:31:10] <SYS64738> nothing [19:31:22] <SYS64738> only seting up the ip to the bge0 ssh starts [19:31:22] <WickedWicky> nothing in the log either? [19:31:50] <SYS64738> I am searching, where must I look ? [19:32:17] <ofu> /var/svc/log or /var/adm/messages [19:32:23] <WickedWicky> there [19:32:44] <WickedWicky> I am not on a solaris machine atm so I cant look for you [19:32:50] <SYS64738> thanks [19:32:56] *** anilg has joined #opensolaris [19:33:15] <WickedWicky> or opensolaris machine for that matter [19:34:12] *** anilg has quit IRC [19:35:04] <SYS64738> the only error is this: [ID 652011 daemon.warning] svc:/network/physical:default: Method "/lib/svc/method/net-physical" failed with exit status 96. [19:35:09] <SYS64738> it's a warning not an error [19:35:26] <SYS64738> this happens when I try to use the e1000g0 [19:36:08] <SYS64738> but /var/svc/log/network-physical\:default.log is empty [19:36:11] <ofu> svcadm clear/restart network/physical doesnt work? [19:36:22] *** Chihan has quit IRC [19:36:23] <ofu> i got this problem on vmware [19:36:24] *** migi has quit IRC [19:36:43] <ofu> ifconfig plumb up and svcadm restart network/physical might work [19:37:01] <SYS64738> now I am at home [19:37:10] <SYS64738> tomorrow I'll try, thanks [19:37:53] *** Bartman007 has joined #opensolaris [19:37:54] *** mikefut has quit IRC [19:38:31] <SYS64738> can I set another ip of the same subnet of bge0 to the e1000g0 ? [19:40:40] *** MattMan has quit IRC [19:42:37] *** migi has joined #opensolaris [19:43:24] *** yippi has joined #opensolaris [19:45:31] *** jfndi has left #opensolaris [19:54:13] *** Snake007uk has quit IRC [19:55:53] <SYS64738> ofu you are right [19:56:06] <SYS64738> I must umplump and re plumb it [19:56:16] *** esproul has joined #opensolaris [19:56:19] <SYS64738> the problem is that I need at boot [19:57:16] <holcomb> hmmm [19:57:29] <holcomb> my e1000g seems to have stopped accepting packets [19:59:13] <SYS64738> I hate all this [19:59:29] <SYS64738> I think that I'go to run for an hour [19:59:31] <SYS64738> bye [19:59:46] *** SYS64738 has quit IRC [20:02:08] *** migi has quit IRC [20:02:16] *** Teltariat has joined #opensolaris [20:02:37] *** nostoi has quit IRC [20:03:02] *** Megaf is now known as Megaf|Away [20:03:08] <Teltariat> Hello folks, quick question: my dtlogin is failing due to bad automatic configuration on the part of Solaris' getconfig. I think I can fix it via the rules in xorg.cfg, but I need to grab my devices vendor id [20:03:15] <Teltariat> Which utility can I use to do that? [20:03:16] <ofu> so, sys64378 is away now... perhaps he should use something like rc.local [20:05:49] <Teltariat> Or a better question is how can I tell dtlogin to use an existing xorg.conf, rather than insisting on using its own getconfig utility to try to automatically configure? [20:05:54] *** Bartman007 has quit IRC [20:09:38] *** Tekni has joined #opensolaris [20:11:00] *** tombhadAC has joined #opensolaris [20:13:12] *** kFuQ has quit IRC [20:13:56] *** millhouse has joined #opensolaris [20:13:57] *** scarab1 has joined #opensolaris [20:14:50] <scarab1> hey [20:14:50] *** dosiu has quit IRC [20:14:58] *** millhouse has left #opensolaris [20:15:08] *** scarab1 has quit IRC [20:15:24] *** millhouse has joined #opensolaris [20:17:16] <millhouse> hey [20:17:29] <millhouse> i'm curious how ZFS compares with XFS [20:17:46] <Teltariat> 2 different things [20:17:57] <Teltariat> XFS is just a filesystem, whereas ZFS is much much more. [20:18:05] <millhouse> throughput wise [20:18:11] <Teltariat> oh, :D [20:18:40] <Teltariat> I was about to say that I'm not really qualified to speak on ZFS's behalf as I'm still very much a Solaris poopyheaded noob. [20:18:46] <millhouse> me too [20:19:09] *** catena has joined #opensolaris [20:19:12] <millhouse> but I've also got a Gentoo/XFS file server and was curious if there was any doc on how the two FS' compared (performance wise) [20:19:30] <Teltariat> There are articles online with speed comparisons [20:19:31] <richlowe> don't think so, but you could run filebench and find out. :) [20:19:38] <catena> when i run uninstaller of netbeans 5.5.1 i get "engine is missing (/InstallShield/Universal/common/Gen1): run installation from source media" [20:19:45] *** Samy has joined #opensolaris [20:19:56] <sommerfeld> zfs performance is (a) generally very good and (b) still improving [20:19:56] <Teltariat> I remember seeing performance reports, perhaps try searching Digg or OSNews [20:20:09] <richlowe> Teltariat: from people who actually know how to do such things, or users? [20:20:13] <richlowe> (yay benchmarks) [20:20:15] <Samy> Google "fefe scalability" [20:20:20] <richlowe> you either get bogus ones from users, or bogus ones from engineers. :) [20:20:20] <Teltariat> richlowe: :E [20:20:25] <Samy> He updated his old scalability tests for FS benchmarks. [20:20:32] <Teltariat> fefe ftw [20:20:40] <Samy> He does have a presentation up with figures showing ZFS, etc... [20:20:53] <Samy> Teltariat: I didn't like his code personally :-P [20:21:06] <Teltariat> Samy: He still undeniably rocks. [20:21:11] <Samy> I was trying to run his scalability tests for Solaris, since there are NO figures for this (excluding the ZFS things). [20:21:22] <Samy> Teltariat: Yes, he did something very useful :-) [20:21:39] * Teltariat wails within the dark pits of dtlogin/Xorg Hell [20:22:13] <Teltariat> Alas, my i810 driver keeps crashing X on my 11/06 install. Alas, alas, woe is me! [20:22:21] <sommerfeld> but, well the reason i like zfs is that when a disk goes insane, it doesn't fly to pieces [20:23:24] <Teltariat> I like it better that when a disk goes insane, there are others there to restrain it within an industrial-strength straitjacket. Yay RAID5. (Have to go try ZFS and RAIDZ) [20:24:11] *** Bart_M has quit IRC [20:26:00] *** catena has quit IRC [20:29:22] *** bubbva has quit IRC [20:29:23] *** linma has quit IRC [20:29:23] *** yippi has quit IRC [20:29:23] *** gdamore has quit IRC [20:29:42] *** bubbva has joined #opensolaris [20:29:43] *** yippi has joined #opensolaris [20:29:44] *** linma has joined #opensolaris [20:36:10] *** migi has joined #opensolaris [20:37:03] *** perlmonk has quit IRC [20:39:23] *** AtomicPunk has joined #opensolaris [20:48:05] *** wamty has joined #opensolaris [20:48:14] <wamty> hat do you think will happen with SCO now? stock dropped to 50 cents lol [20:49:07] *** Jondice has quit IRC [20:50:31] *** mikefut has joined #opensolaris [20:50:33] *** Jondice has joined #opensolaris [20:51:54] *** RaD|Tz has quit IRC [20:53:12] <axisys> anyone know of a 64bit perl pkg for sol 10 ? [20:53:21] <axisys> not in sunfreeware or perl site [20:53:56] <seanmcg> there a reason for needing a 64bit perl ? [20:54:26] <richlowe> seanmcg: depends on who you ask. [20:54:46] <axisys> seanmcg: oracle 10 64 bit needs 64bit perl [20:54:47] <coffman> svc:/network/rpc/bind:default (RPC bindings) State: maintenance since 13. August 2007 20:36:34 CEST Reason: Start method failed repeatedly, last exited with status 1. [20:54:51] <coffman> gar [20:55:07] <coffman> bind refuses to start for some reason [20:55:11] <coffman> cant figure why [20:55:15] <seanmcg> axisys: really ? how is it using it ? [20:55:57] <axisys> seanmcg: oracel developer finding the perl api for oracle libabry needs to be 64 bit [20:57:34] *** Tekni has quit IRC [20:57:56] <axisys> CSKperl is also 32 bit [20:58:02] <millhouse> is there a way i can get bonnie/bonnie++ pre-packaged for solaris? [20:58:18] <coffman> some one can help me with that rpc-bind problem? [20:58:25] <coffman> i cant find any logs etc [20:58:37] <coffman> nothing that says me something [20:58:48] <seanmcg> axisys: gotya thanks. [20:59:16] <axisys> so I guess there is no 64bit perl pkg [20:59:27] <WickedWicky> "Unfortunately, Oracle 10g is distributed as 64-bit libraries only and [20:59:28] <WickedWicky> the perl version from sunfreeware is a 32-bit executable. If you try to [20:59:28] <WickedWicky> link you get complains about 'wrong ELF class: ELFCLASS64'. The only way [20:59:28] <WickedWicky> to resolve this I found was to recompile perl in 64 bit mode from [20:59:28] <WickedWicky> sources. This is pretty much straight forward, just make sure you [20:59:28] <WickedWicky> include full 64-bit support. [20:59:30] <WickedWicky> " [20:59:36] <WickedWicky> http://coding.derkeiler.com/Archive/Perl/perl.dbi.users/2004-04/0122.html [21:00:04] <timeless> i built perl64 myself [21:00:07] <timeless> it wasn't too painful [21:00:23] <timeless> the suckier one is building sleepcat or something [21:00:25] <axisys> timeless: i am doing it now [21:01:24] <timeless> heh [21:02:23] * timeless grumbles and considers installing python [21:02:30] <coffman> rpcbind fails to start, and i cant see why [21:06:50] *** Mdx4 has joined #opensolaris [21:06:58] *** Fullmoon has joined #opensolaris [21:07:46] *** sunmaster has joined #opensolaris [21:11:20] *** nrubsig has joined #opensolaris [21:11:20] *** ChanServ sets mode: +o nrubsig [21:11:38] * coffman kicks rpc [21:11:38] * nrubsig looks for the person who created his misery... [21:11:40] <axisys> coffman: what does this say ? /var/svc/log/network-rpc-bind:default.log [21:11:54] <nrubsig> !seen Tpenta [21:11:57] <Drone> Tpenta is currently online in #opensolaris and last spoke on Mon 13 Aug 2007 08:17 GMT, saying 'brendan was saying that's progressing pretty well'. [21:11:57] *** boyd has quit IRC [21:12:00] * e^ipi hands nrubsig a mirror [21:12:08] <e^ipi> there is nothing but what you make of situations [21:12:21] <axisys> coffman: try to start it manually /lib/svc/method/rpc-bind start and see what it is complainign about [21:12:22] <e^ipi> if you choose to be miserable, recognize that it's your choice [21:12:27] <coffman> axisys: [ Aug 13 20:36:34 Executing start method ("/lib/svc/method/rpc-bind start") ] [21:12:28] <WickedWicky> wisely spoken, e^ipi [21:12:32] <coffman> rpcbind failed with 1. [21:12:36] <coffman> thats it [21:12:39] <nrubsig> e^ipi: Yeah, but this time it's all Tpenta's fault... really... [21:12:43] *** kaiwai has quit IRC [21:12:48] <nrubsig> e^ipi: They're chasing me... [21:12:53] *** oninoshiko has joined #opensolaris [21:12:56] <axisys> /lib/svc/method/rpc-bind start <-- and see what it is complaining about [21:12:59] * WickedWicky calls 911 [21:13:12] * nrubsig lives in fear of the next email [21:13:26] <nrubsig> and the next [21:13:27] <coffman> axisys: again, "rpcbind failed with 1." [21:13:28] <nrubsig> and the next [21:13:29] <nrubsig> next [21:13:31] <nrubsig> next [21:13:46] <nrubsig> coffman: man firewall ? Or rpcbind leftover process ? [21:13:55] <axisys> coffman: svcs -d rpc/bind [21:13:59] <nrubsig> coffman: did you test the issue with lsof ? [21:14:06] <axisys> are all the service online there? [21:14:22] *** Tekni has joined #opensolaris [21:14:23] *** nrubsig sets mode: +o Tpenta [21:14:26] <coffman> axisys: yes [21:14:32] <oninoshiko> ohh... svn_69 [21:14:54] <coffman> axisys: would not ask, if they werent [21:15:12] <axisys> coffman: sh -x /lib/svc/method/rpc-bind start [21:15:19] <axisys> see if that debug tells u anything [21:16:06] * oninoshiko wonders what changed, other the it's sexyness [21:16:30] <coffman> http://pastebin.ca/655947 [21:16:35] <PerterB> seeing as all the method does is basically run rpcbind, it would be better to truss that directly and look for clues as to why it exits [21:16:42] <axisys> coffman: also check linma nrubsig suggeted.. see if ps -ef | grep ipmon shows anything [21:16:49] *** dlynes_laptop has joined #opensolaris [21:17:05] <axisys> coffman: or what PerterB said [21:17:10] <nrubsig> !summon Tpenta [21:17:40] <e^ipi> roland, you need to chill out my friend [21:17:55] <axisys> coffman: did u find anything with ps -ef | grep ipm [21:17:58] <moazamraja> !summon HalleyBarry [21:17:58] <e^ipi> have a cup of tea, watch some birds [21:18:01] *** dlynes_laptop has quit IRC [21:18:06] <coffman> "ps -ef | grep ipmon" shows nothing [21:18:41] <coffman> realy strange, all worked fine until this evening, it broke at runtime [21:18:46] * nrubsig watches the birds in the fridge [21:18:53] <oninoshiko> o.0? [21:18:54] <axisys> coffman: ok.. so truss may help u find some issue.. also check the messages file.. if something pops up there [21:18:55] <nrubsig> e^ipi: this is so depressing. [21:19:23] <e^ipi> i'm not sure what "it" is, but just don't let it get to you [21:19:24] <nrubsig> e^ipi: and I hate "dried dead mole"-tea [21:19:47] <e^ipi> it's probably not worth giving yourself a stroke over [21:19:55] * nrubsig twitches [21:20:02] <oninoshiko> nrubsig, keeping pinguins? [21:20:03] <nrubsig> grrrwwwwaaaaaahhhh [21:20:06] <PerterB> huhuh huhuh... he said "stroke over" [21:20:23] <nrubsig> oninoshiko: no, dead birds from the supermarket [21:20:24] <coffman> demsg only craps me with kerberos errors, its also down cause of rpc-bind [21:21:04] <nrubsig> coffman: maybe the new toiditoidi/RPC-worm ? [21:21:06] <oninoshiko> oh... i suppose that makes sense too... [21:21:38] <PerterB> coffman: pastebin the truss output... [21:21:52] * nrubsig giggles and hopes coffman doesn't read "toidi" backwards... [21:22:06] * nrubsig hides from coffman's wrath [21:22:34] * oninoshiko giggles [21:23:02] <nrubsig> ekoj-terror!! [21:23:14] *** Tekni has quit IRC [21:24:05] <coffman> http://pastebin.ca/655958 [21:24:15] <nrubsig> !summon --mode=now_or_I_ask_your_wife_to_send_you_here Tpenta [21:24:54] <coffman> nrubsig: well, someday... you cant hide from that easy... and since im in german too, it gets easier for me [21:25:30] * nrubsig hides deeper [21:25:56] <Teltariat> scary [21:26:19] <Teltariat> So Xorg works, but the dtlogin doesn't. [21:26:27] <Teltariat> Anyone have any suggestions? [21:27:13] *** dos000 has joined #opensolaris [21:27:15] <Teltariat> When called directly as /usr/X11/bin/Xorg, damn thing runs, and runs fine, but when /etc/init.d/dtlogin is issued, its a complete wash [21:27:16] <nrubsig> Teltariat: blocked port, nscd not working, DNS not working etc. [21:27:25] *** Jondice has quit IRC [21:27:28] <dos000> anyone here tried running solaris on ibm x336 ? [21:27:33] <nrubsig> Teltariat: and what does /var/dt/Xerrors say ? [21:27:46] <dos000> this is becoming one hell of a drive [21:27:49] <ofu> RElling: http://oliver.fuckner.net/index.php?/archives/33-Gigabyte-GC-Ramdisk.html [21:28:00] <Teltariat> nrubsig: Xerror says that i810 is crashing. The weird thing is that dtlogin and Xorg don't seem to both be using /etc/X11/xorg.conf [21:28:19] <nrubsig> Teltariat: ask alanc [21:28:22] <PerterB> coffman: has your /etc/netconfig gone missing? rpcbind probably needs it for doing TLI sockets... [21:28:37] <Teltariat> alanc: ping... ? *sheepish grin* [21:29:00] * nrubsig is away for 15mins [21:29:34] <coffman> PerterB: its there, its fine [21:29:51] <coffman> i allready did sys-unconfig btw [21:31:22] <axisys> timeless: I downloaded the CSKperlSrc .. do u use -Duse64bit with Configure to compile 64 bit perl? [21:31:28] <PerterB> and world readable? rpcbind seems to mess with its privileges before trying to open it, then the open fails with EACCES which implies it doesn't have permission [21:31:47] <timeless> sounds right [21:31:54] <timeless> although i was most concerned about 64bit files [21:32:06] <timeless> because i had a very big database :( [21:32:21] <timeless> basically there are ~4 64bit flags, read the docs about them and pick properly [21:32:40] <timeless> i'm about to head home or something. i could try to find my solaris package, it's on my stage zone [21:32:47] <timeless> (but not now/today) [21:33:28] * timeless cursers [21:33:42] <timeless> what's the redo keystroke in /usr/bin/vi ? [21:34:04] <PerterB> z [21:34:34] <PerterB> uh, 'u' (not sure where the z came from) [21:34:51] <coffman> PerterB: well, with 777 permission it still fails [21:34:59] *** Jondice has joined #opensolaris [21:35:21] <PerterB> weird [21:37:10] *** Jondice has quit IRC [21:37:29] *** Jondice has joined #opensolaris [21:39:19] *** Mdx4 has quit IRC [21:39:39] *** Jondice has quit IRC [21:39:46] <Teltariat> General question: Xorg, when called directly, will use /etc/X11/xorg.conf; but dtlogin uses what? [21:40:42] <axisys> timeless: I get this error while trying to compile perl http://rafb.net/p/U62tA368.html [21:41:01] *** Bartman007 has joined #opensolaris [21:41:43] <timeless> remember: i said something about needing to bbuild a second package? [21:41:45] <timeless> db is it :) [21:41:55] *** estibi has joined #opensolaris [21:41:59] <palowoda> Teltariat: /etc/X11/.xorg.conf [21:42:18] *** Mdx4 has joined #opensolaris [21:42:35] <axisys> timeless: aitght let me play with them a little [21:42:45] <timeless> BerkeleyDB-4.4 -> db-4.4.20 [21:42:49] <Teltariat> palowoda: thank you, I will try that [21:43:03] <axisys> timeless: ok [21:43:06] <timeless> i can stick that up somewhere if you really want it [21:43:07] <axisys> timeless: is that a pkg? [21:43:14] <timeless> yes [21:43:21] <timeless> OSSOdb-4.4.20-sol10: package datastream [21:43:25] <axisys> timeless: u built or sun? [21:43:34] <timeless> osso is my entity [21:43:52] <timeless> it'd say SUNW if it was sun :) [21:43:57] <axisys> timeless: gotcha.. i like your clue will help me.. [21:44:02] <nrubsig> WTF is "osso" ? [21:44:10] <nrubsig> "osso" the cookie monster ? [21:44:15] <axisys> timeless: well they have CSK the new one [21:44:18] <timeless> good question [21:44:28] <axisys> timeless: so did not knwo if this is a third type [21:44:35] <timeless> axisys: fair enough [21:45:02] <timeless> https://engineering.purdue.edu/ECE/Seminars/2006-07Seminars/03.21.07Abdullah.pdf [21:45:12] <timeless> is one of the few urls i ever find that actually says what osso is [21:45:25] <timeless> lists.xiph.org/pipermail/foms/2007-January/000102.html is an html url that does [21:46:07] <timeless> it's clearly not technically a secret, but it's mentioned so rarely, that i'd rather not be blamed for spelling it out :) [21:46:43] <Teltariat> palowoda: dtlogin is still explicitly starting X with "-depth 24", and this is whats causing problems. .xorg.conf doesn't seem to be overriding it; any suggestions? [21:48:37] <palowoda> Teltariat: What is wrong with that color depth? [21:48:47] <Teltariat> My machine seems unable to run at that depth [21:48:54] <palowoda> Why? [21:49:00] <Teltariat> i810 crashes at that depth. This is Solaris 11/06 [21:49:14] <palowoda> Does that latest opensolaris build work? [21:49:26] <Teltariat> I have not tried the latest OpenSolaris build\ [21:49:35] <palowoda> OK give it a try. [21:49:39] <Teltariat> I was trying to get 11/06 working before I did [21:49:42] <Teltariat> heh :) [21:49:48] <palowoda> Mistake. [21:50:42] <Teltariat> Never BFUed before; I'll give it a shot now [21:52:27] *** yarihm has joined #OpenSolaris [21:52:36] <Teltariat> Thanks anyway, palowoda. [21:52:47] *** KermitTheFragger has quit IRC [21:52:50] *** wamty has quit IRC [21:53:02] * nrubsig puts a cup full with milk outside his house to make Tpenta purr ... [21:53:18] <palowoda> Teltariat: Why don't you just try build 69 and see if the video comes up first. [21:53:40] <Teltariat> palowoda: the video and everything worked fine during the installation phase [21:53:57] <Teltariat> I installed within the graphical environment; everything worked perfectly then. [21:53:58] <palowoda> Which build? [21:54:05] <Teltariat> Same, Solaris 11/06 [21:54:08] <palowoda> Bah. [21:54:41] <Teltariat> I take it you don't like Mr. 11/06. [21:55:02] <palowoda> It's a lousy topic on this channel. [21:55:08] <Teltariat> I got that DVD in the mail from Sun as part of a promotion [21:55:36] <Teltariat> Well then SX 69 it is. [21:58:52] *** seanmcg has quit IRC [22:00:48] * nrubsig summons Tpenta [22:01:59] <quasi> nrubsig: don't disturb him - with luck he'd be in the middle of adding dtrace probes to ksh ;) [22:03:40] <brendang> it's 6am Tpenta time. since he is an early bird, he might be online soon... [22:03:44] <nrubsig> quasi: I guess he is more likely caught in some horrible customer escalation with 10 super-important meta-managers standing behind him, all armed with axes and the most important manager tries to threaten Tpenta to burn all contracts with Sun if Tpenta can't fix the current problems ASAP [22:03:58] <coffman> another truss output, the first could be wrong. http://pastebin.ca/656024 [22:04:50] <quasi> nrubsig: does sun even do that sort of escalation? [22:05:33] <sommerfeld> quasi: they are shockingly common [22:06:10] <brendang> coffman: open("/etc/netconfig", O_RDONLY) Err#13 EACCES [file_dac_search] [22:06:17] <brendang> coffman: is your /etc 700? [22:06:39] <quasi> sommerfeld: honestly I would have been very surpised if that wasn't the way of most escalations [22:06:51] *** bengtf__ has joined #opensolaris [22:07:06] <coffman> drwxr-xr-x 92 root sys 4,5K Aug 13 21:59 etc [22:07:09] *** stevel has joined #opensolaris [22:07:09] *** ChanServ sets mode: +o stevel [22:07:43] *** nostoi has joined #opensolaris [22:09:21] <sommerfeld> coffman: ls -ld / ? [22:10:29] *** yippi has quit IRC [22:11:03] <brendang> coffman: file_dac_search normally means the user doesn't have execute perms. [22:11:07] <richlowe> stevel: so, you survived, huh? :) [22:11:49] *** mikefut has quit IRC [22:13:38] <stevel> richlowe: yeah - but with a helluva sore throat [22:14:38] <coffman> sommerfeld: oh hell, permissions where wrong [22:14:48] <coffman> on / [22:14:49] <sommerfeld> on / ? [22:14:53] <coffman> thats so bad [22:14:54] <coffman> yes [22:14:56] <brendang> :) [22:15:23] <coffman> i had 750 on / [22:15:33] *** Jondice has joined #opensolaris [22:15:49] <coffman> and owner $myacc group root [22:15:51] <brendang> coffman: ahh well, almost a chance to pull out DTrace... [22:16:13] <coffman> brendang: still got that issue with the mem leak :P [22:17:16] <coffman> strange thing, i did do some permission stuff today, but it was a total path but with -R [22:17:31] <richlowe> stevel: sounds like an excuse for (cheap) whisky and lemon. [22:17:44] <richlowe> stevel: though do not defile good whisky with additives. [22:17:54] * stevel was more inclined to just drink the good whisky straight [22:18:01] * coffman blames a symlink in his home [22:18:11] * coffman wonders where that one is [22:18:30] <brendang> coffman: yes - we had determined the application, just needed to catch a translated ustack() to see why it was leaking [22:19:11] <coffman> any how i can list only symlinks in dir/subdirs ? [22:20:50] <stevel> find <dir> -type l [22:21:21] *** whaq_ has quit IRC [22:21:32] *** whaq has joined #opensolaris [22:21:53] *** movement has quit IRC [22:22:32] *** movement has joined #opensolaris [22:22:49] *** sunmaster has quit IRC [22:24:02] *** movement has quit IRC [22:24:24] *** movement has joined #opensolaris [22:25:24] *** movement has quit IRC [22:25:48] *** movement has joined #opensolaris [22:26:28] *** migi has quit IRC [22:26:31] *** migi_ has joined #opensolaris [22:26:40] *** migi_ is now known as migi [22:28:00] * coffman feels dirty [22:28:18] <coffman> that was some rm -rf / style error [22:28:26] <coffman> but i realy did not do it [22:28:29] <coffman> hmpf [22:30:19] * oninoshiko doesn't like wiskey... now some fine rum... thats a whole nother matter [22:31:49] <WickedWicky> wiskey, a cigar and some chocolate [22:32:07] * oninoshiko wonders why whiskey seems associate with more temperate climates, and rum with more tropical... "is there a reason other then historical origin?" [22:32:15] <oninoshiko> associated* [22:32:26] <WickedWicky> oninoshiko: basicly we dont have sugar cane here :P [22:32:40] <sommerfeld> source of sugars for fermentation [22:32:55] <sommerfeld> (or starches..) [22:33:07] <coffman> gar, have to get up early and i need to stay sober [22:33:16] * coffman teaches kids this week [22:33:19] <WickedWicky> fine, I'll drink this caipirinha by myself then [22:33:24] <oninoshiko> not exactly what i ment -_- [22:34:23] <coffman> what a pitty that i missed the ccc-camp this year [22:35:32] <oninoshiko> i mean i understand the difference between malt and molasses... [22:36:02] * coffman wonders if there is a marketing budget for sun externals that do speaches about solaris [22:36:55] <oninoshiko> but its like whisky "worms you up" while rum seems more ok drink in the middle of a heat wave... [22:37:13] * oninoshiko wonders back to work [22:37:44] <WickedWicky> in the middle of a heat wave you eat vodka-melon [22:38:11] *** Tekni has joined #opensolaris [22:38:13] <oninoshiko> vodka-melon? [22:38:16] <WickedWicky> yes [22:38:20] <WickedWicky> get a water melon [22:38:23] <WickedWicky> drill a hole in it [22:38:43] <oninoshiko> add cheap vodka? [22:38:44] <WickedWicky> put a bottle of vodka top under, in the hole and let the vodka drip into the melon [22:38:58] <WickedWicky> let it rest for an hour of 4, then eat the melon [22:39:16] <WickedWicky> well yes, using a 20 euro vodka is kind of a waste [22:39:24] <coffman> pah [22:39:34] <axisys> timeless: i were able to compile perl 64 bit.. i used thedb library that came with sun companion cd [22:39:39] *** migi has quit IRC [22:39:40] <oninoshiko> hey, sounds like a plan to me [22:39:45] <timeless> cool [22:40:02] <axisys> timeless: for for the major clue [22:40:08] <timeless> :) [22:40:13] <axisys> s/^ for/thanks/ [22:40:38] <timeless> it's bad when i auto correct things like that as i read .. [22:41:32] <axisys> heh [22:42:26] * oninoshiko knew what you ment [22:44:25] *** Mdx4 has quit IRC [22:47:08] *** migi has joined #opensolaris [22:49:18] <millhouse> I'm having issues with ZFS... If I'm deleting a large set of files (in this case a few thousand directories and files) the system as a whole seems to almost stop [22:49:21] <millhouse> is this normal? [22:50:06] <elektronkind> stevel, are you about? [22:50:10] *** smg_ has joined #opensolaris [22:50:22] <RElling> millhouse: is the disk busy? [22:50:24] <stevel> elektronkind: yup [22:51:08] *** sparc-kly has quit IRC [22:51:09] <elektronkind> stevel: heya :) The cr.o.o site updates its users who have Contributor status on the half hour, yes? [22:51:24] <stevel> elektronkind: i believe so [22:51:54] <stevel> every 20 minutes [22:52:08] <stevel> i.e.: 14:00, 14:20, 14:40 [22:52:16] <elektronkind> ok... I'll pop off an email to the swamp admins then. I'm not showing up as of yet [22:53:15] <axisys> timeless: i did use this sh Configure -Duse64bitint [22:53:23] <axisys> but still it says 32 bit [22:53:25] <axisys> hmm [22:53:45] <richlowe> stevel: hey, on that note... [22:53:50] <timeless> 64bitint doesn't relate to 64bitpointer [22:53:55] <timeless> it just means 64bitint [22:54:09] <timeless> and in fact that alone would be 32bit :) [22:54:34] <axisys> timeless: i see so I should use the 64bitall option [22:54:47] <timeless> would be simpler :) [22:55:00] <timeless> i'm pretty sure you're back to needing to build the other package :) [22:55:21] <pjd-> millhouse: How much memory ZFS consumes? One FreeBSD guy has theory that metaslab fragmentation can kill machine taking all the memory. [22:55:34] <stevel> richlowe: hrm? [22:55:40] <richlowe> stevel: see msg. [22:55:43] <stevel> oh wait. PM. got it [22:56:02] <richlowe> like I'd admit to being that stupid in public ;) [22:56:11] <stevel> heh [22:56:24] <elektronkind> pjd-: we've had to set a hard limit on ARC cache size. On our s10 boxes, ARC will grow and grow and eventually cause processes to swap out :/ [22:56:45] <elektronkind> pjd-: we put an autosizing init script up on the genunix site that does that [22:57:08] *** StylusEater_Work has left #opensolaris [22:57:59] <pjd-> elektronkind: ARC limits don't work on metaslab. The test case is to create huge number of files and delete every second file. [22:58:17] <timeless> sp3000: maemo.2006-11-05 brings back memories :) [22:58:21] <timeless> hrm, wrong window [22:59:00] <millhouse> it's got 8GB on the system.. the disk was busy deleting stuff, but nothing otherwise [22:59:44] <sommerfeld> there's been a bunch of work done on zfs memory usage relatively recently (post s10u3) [22:59:57] *** Yamazaki-kun has quit IRC [23:02:02] <millhouse> i've been trying to benchmark ZFS (iscsi setup) against a gentoo/linux box running XFS. The gentoo/XFS server is what we have now and Solaris/iscsi/zfs is what we want to go to... but it seems like XFS is still faster than ZFS [23:02:06] <millhouse> am I doing something wrong? [23:02:38] <jmcp> yeah [23:03:04] <pjd-> sommerfeld: I see not limits whatsoever in the current metaslab code ... [23:03:05] <jmcp> if you want benchmarking help you'll need to go into some detail about what you're doing, and you should do that on the zfs-discuss at opensolaris dot org mailing list [23:03:08] *** alanc_away has joined #opensolaris [23:03:09] <Bartman007> millhouse: yes, you haven't tried sudden powerfailure yet, then see which server actually retains your data :-) [23:03:43] <sommerfeld> pjd-: so clealy there's still some more work to do.. [23:04:15] <sommerfeld> millhouse: number one thing to check: make sure you're not running a DEBUG build of solaris [23:04:23] <WickedWicky> yea.. XFS is so damn fast cause it caches everything and doesnt commit to disk very often... when you have a power fail with XFS you can go buy kleenex and cry [23:04:44] <millhouse> Actually... We had that with solaris/zfs and lost the pool (unrecoverable). The problem though, was the Linux distro we were using on the iscsi target machines. [23:05:23] <millhouse> on the head node i'm running solaris 10 (whatever the latest dvd they're shipping). And on the targets I'm running OpenSolaris b68 (express, not developer) [23:05:36] <WickedWicky> keeping everything in cache is cool, performance wise.. but i prefer data intergrity over 1GB/s [23:06:01] <millhouse> how can I check if I'm running the DEBUG build of solaris? [23:06:12] <WickedWicky> it tells you on boot I think.. DEBUG enabled [23:06:18] <jmcp> yup [23:06:31] <WickedWicky> it does for me at least [23:06:31] <WickedWicky> ah [23:06:41] <millhouse> I think partially, the gentoo/xfs server has the advantage of being a DAS solution, whereas Solaris/ZFS/iscsi uses iscsi [23:06:52] <millhouse> okay, i don't think I am then [23:07:16] <RElling> the vdevs are iscsi clients? [23:07:24] <jmcp> millhouse: I'm fairly sure that whatever you get on the OpenSolaris dvd is not a debug build [23:07:37] <millhouse> could running Solaris 10 on the head node and OpenSolaris on the back-end nodes be a possible source of slowness? [23:07:45] <millhouse> I didn't think so [23:08:08] <millhouse> yes [23:09:01] <millhouse> OpenSolaris is on the end "target" machines, and I use ZFS to share itself via iscsi. Then I use Solaris' iscsiadm suite on the head node to add them (static-config) and then create a "master" zpool/zfs on it [23:09:43] <WickedWicky> I think what RElling is asking is whether the zpool consists out of iscsi targets or sad [23:10:16] *** Bart_M has joined #opensolaris [23:10:53] *** kFuQ has joined #opensolaris [23:11:48] <millhouse> the zpool consists only of iscsi targets [23:12:00] <WickedWicky> that won't be very fast then... will it [23:12:33] <millhouse> well, it's proven itself (I think). We've put it against windows and every time it comes in at just under twice as fast [23:13:02] *** SYS64738 has joined #opensolaris [23:13:27] <jmcp> you're still going to be hamstrung by the abomination which is iscsi [23:13:30] <jmcp> :) [23:14:01] <millhouse> and the speed differential between XFS is noticeable (at least while running synthetic benchmark tests), but it's still pretty good... And the trade off is: it'll outgrow XFS many times over [23:14:10] <jmcp> that's for sure [23:14:20] <jmcp> keep trying the sudden powerloss scenario too [23:14:32] <WickedWicky> (not on live data) [23:14:33] <brendang> millhouse: were they opensource benchmark tests? (eg, filebench, bonnie, etc?) [23:14:53] <millhouse> WickedWicky: I've also been fighting against marketing... It's hard to say that I've got a "good" solution when 10 marketing companies out there boast 2x-10x better the performance [23:15:11] <millhouse> ...I'm pretty sure they're boasting about potential and theoretical performance...but that doesn't translate... [23:15:26] <WickedWicky> try working for a telco and say you have a proposal to change things when they're doing the same for years and years :P [23:15:30] <WickedWicky> Iknow your pain :P [23:15:39] <millhouse> jmcp: i'd like to, but not now... right now that gentoo/xfs box has our most productive office, literally, "by the balls" [23:15:48] <jmcp> ah [23:15:53] <millhouse> if it died... that site would not have work... [23:15:58] <jmcp> sorry you've got that limitation [23:16:05] <brendang> millhouse: exactly. marketing often says whatever suits them. try filebench - it is quite configurable. [23:16:15] <WickedWicky> it'd give you the perfect "I told you so" though [23:16:19] <jmcp> can you setup a parallel system to benchmark it with? [23:16:30] <millhouse> later though... and this friday is slated to be the day we start the 2-3 day implementation/migration process from XFS/DAS to ZFS/iSCSI [23:16:46] <WickedWicky> ZFS should be faster since it does striping, no? [23:17:27] <WickedWicky> ah no, caching [23:17:53] <millhouse> I did some tests (a home-brewed test that creates 5000*26 directories and then copies 2 1KB .tiff files into each directory) [23:18:16] <millhouse> ZFS/iSCSI came in around 13m56s [23:18:32] <millhouse> for raidz, mirror, and raid-0/span config [23:18:41] <jmcp> how about tests of larger files, such as 10Mb? [23:18:58] <millhouse> redhat/lvm+ext3/iSCSI came in around 40s quicker and Gentoo/XFS (DAS) came in at 8m13s [23:19:26] <millhouse> well, this test came as a result of a job we might get where the data will literally come to us similar to that [23:19:37] <jmcp> ah [23:19:45] <oninoshiko> we are just DIEING to go to ZFS/iSCSI here... we have been bit to many times (read ever) by "client disables our monitoring and has drives fail" [23:19:55] <millhouse> and small files are ZFS' downfall I believe [23:20:12] <jmcp> millhouse: benchmarks which are as close to your actual expected operating conditions are incredibly better than made up ones [23:20:19] <millhouse> oninoshiko: have you been playing with ZFS/ISCSI any??? [23:20:23] <oninoshiko> yes [23:21:02] <millhouse> My first thought when everyone came back and said "So, Solaris/ZFS is slower.." was "true... but it'll still be here when we need 100+TB" [23:21:15] <oninoshiko> ZFS is working within specified parameters, iSCSI is not (yet) [23:21:29] <millhouse> oninoshiko: may I inquire as to your setup and results? [23:21:36] <millhouse> how do you mean? [23:21:57] *** vmlemon has quit IRC [23:22:00] <oninoshiko> our big desire is to have HBAs booting from the iSCSI targets [23:22:47] <oninoshiko> I'm not sure why yet, but the HBAs cannot seem to see the target consistantly [23:23:38] <millhouse> I noticed some issues with targets being seen and then remapped to a different vdev [23:24:16] <millhouse> have you been able to get any performance data on it? [23:24:48] <sle> on a slightly related note, anybody have any experience with using ZFS as just a volume manager for UFS? [23:24:51] <oninoshiko> im going to be grabbing a wireshark capture of a successful boot, and adding it to my thread on the topic [23:25:15] <oninoshiko> sle: no, but i did see one of the sun peoples blog talking about it [23:25:28] *** Mdx4 has joined #opensolaris [23:25:33] <oninoshiko> dont recalls who's though... you'll have to google around [23:25:36] <sle> i think i read that blog post too [23:25:41] <sle> but it didn't give any performance numbers [23:26:10] <sle> i'm wondering if it's something i should really investigate [23:26:17] <millhouse> yeah... I've seen the throughput numbers vary quite a bit.. A conservative, safe, guess is that we're getting around 20-30MB/sec right now [23:26:25] <sle> or just toss into the 'neat-idea-but-not-gonna-work' bin [23:26:36] <millhouse> I blame windows clients for not being better, but that doesn't go very far...heh [23:27:38] <sle> what kind of networking are you using for iSCSI? [23:27:59] <sle> (speed-wise) [23:28:38] <millhouse> our systems are SuperMicro's (Intel 1000 NICs), we're using NetGear for our switch (it supports 10G and "daisy chaning" to multiple switches) [23:28:45] <millhouse> gbit [23:28:57] <sle> oh, that's pretty good [23:29:07] <millhouse> and we haven't tested it yet, but we're going to put 10G on the head node to talk to the outside world [23:29:37] <sle> just single GbE lines? [23:29:39] <millhouse> yeah... it just seems...sluggish at times [23:29:56] <millhouse> yeah, but my boss just told me he wants to go for aggregate bonding so we'll have 2Gbit [23:29:56] <sle> or aggregated? [23:30:03] <sle> cool [23:30:06] <oninoshiko> we are using an isolated GigE network, Std MTU (ATM) [23:30:10] <millhouse> yeah, hopefully we'll see a perf gain.. [23:30:15] <sle> we're just about to do that too [23:30:23] <sle> to give our LDoms some breathing room [23:30:30] <sle> 2.5" SAS drives are way too expensive :/ [23:30:37] <millhouse> yeah, we are too. The head node is physically connected to the LAN (external) switch and also to the iSCSI internal switch [23:30:44] <millhouse> all the targets only see the iscsi lan [23:30:46] <oninoshiko> i want a full set of tests before i go to 9K frames [23:30:59] <millhouse> the target systems have 4GB RAM and the head node has 8GB [23:31:18] <millhouse> we're using 7200 SATA drives [23:31:43] <millhouse> with LSI MegaRAID adapters to do hardware raid-5 on the targets [23:33:10] <sle> not using RAID-Z? [23:33:20] <millhouse> i'm using raid-z on the head node [23:33:48] <SYS64738> millhouse, what server do you have ? [23:33:52] <oninoshiko> im using raid-z for my target pool [23:34:02] <millhouse> what do you mean? [23:35:34] * oninoshiko wants Q3 *cries* [23:36:00] <millhouse> the head node is managing all the iscsi vdev's via a raid-z config [23:37:07] <coffman> even if i repeat my self, where is b70? [23:38:58] <oninoshiko> coffman: i beleave you have to build ON 70 yourself [23:39:14] <sommerfeld> not escaped yet [23:39:38] <elektronkind> wondering one of the Sun campuses in a hospital gown, mumbling to itself and conversing with the concrete planters [23:39:49] <WickedWicky> I'm gonna call it the day [23:39:52] <WickedWicky> see ya all tomorrow [23:40:19] * coffman raises his fist [23:40:28] <coffman> damm sxde [23:41:00] <millhouse> later [23:41:29] *** yippi has joined #opensolaris [23:41:46] *** reidms has joined #opensolaris [23:41:54] <Pietro_S> coffman: from the plans about 70b incoming big changes with being sxde and vacantions I thouught it will have big delay [23:42:04] *** linux_user400354 has joined #opensolaris [23:43:26] <coffman> yeah [23:43:53] <elektronkind> the more likely situation is that b70 and s10 7/07 ran off to Vegas to get married and they're both sitting on a beach in Fiji sipping from cononut drinks. [23:44:11] <elektronkind> er coconut [23:44:31] <coffman> people allways tell that germans have to much vacantions, but americans are even worse since they go in vacantions with the whole company [23:44:33] <coffman> sucks [23:44:40] <oninoshiko> FTR, they do monitor this channel, and are using your plight as a form of entertainment [23:45:27] <elektronkind> coffman: be prepared for what's going to happen 1.5 weeks from now. A respectable chunk of silicon valley workers are taking the week off to go to Burning Man [23:45:32] <wesolows> Eh, I don't take vacation. Then again, I don't consider myself an American. [23:45:53] <oninoshiko> i am an american, and never have time to vacation [23:45:55] <wesolows> At least there was no Sun July 4 shutdown this year. [23:46:25] <axisys> with sun studio 12 can I use -m64 for 64 bit? [23:46:55] <elektronkind> axisys: yes. with studio 10 and 11 too [23:47:02] <elektronkind> er wait [23:47:07] *** nostoi has quit IRC [23:47:18] <elektronkind> oh... m64 in 12, and -xarch=$foo in 10 and 11 [23:47:23] <wesolows> I don't think that's true. SS12 only. [23:47:54] <timeless> o New -xarch Flags For x86 Development [23:47:55] <timeless> The -xarch option now supports the following new flags [23:47:55] <timeless> for development on the x86 platform: amd64a, [23:48:05] <elektronkind> sorry, I was daydreaming of vacationing in fiji [23:48:27] <wesolows> Vacation is overrated. Retirement is what you want. [23:48:42] *** Megaf|Away is now known as Megaf [23:48:50] <elektronkind> which reminds me, I saw a edu job posting a while ago... the University of Fiji was looking for a Director of IT (seriously) [23:49:01] <wesolows> Ugh, Director of IT. [23:49:02] <oninoshiko> millhouse: as far as my results go, the iSCSI is a little sluggish, but im hoping i can make some tweaks to help with that (9k packets come to mind) [23:49:21] *** cypromis has joined #opensolaris [23:49:22] <oninoshiko> and the output of bonnie is to long to post [23:49:47] <wesolows> The Director gets all the shit that rolls down the hill, then gets it again as the wave action splashes it back up. [23:50:51] <coffman> fijis is the place where the money goes of all that bad companys? has to be a nice place [23:51:02] <oninoshiko> ofcourse my point of reference was a 3ware 8006-2LP in a RAID-1, which was not exactly the definition of "zippy" [23:51:35] <millhouse> okay, it's comforting to know that I'm not alone in this. I don't know Solaris or TCP/IP well enough to know where and how to tweak it. [23:52:23] <millhouse> I played with jumbo frames once with almost no change in performance, but I didn't spend much time with it. [23:53:05] <oninoshiko> the trick with Jumbo frames is you have to have it enabled on all the devices on the network, or you can actually lose performance [23:53:14] <millhouse> yeah [23:53:57] *** smg_ is now known as seanmcg [23:53:58] <millhouse> this whole thing has been a little daunting.. I went from being a college grad to being hired to be the "Linux guy" and promptly handed the task of developing and implementing a scalable, enterprise storage solution [23:54:13] *** migi has quit IRC [23:54:19] <oninoshiko> i also am looking at the possibility of using iscsi over SCTP/IP [23:54:25] <oninoshiko> welcome to my world :p [23:55:09] <millhouse> nice [23:55:29] <oninoshiko> actually i still have classes yet to finish [23:55:37] <millhouse> nice [23:55:45] <millhouse> i take it u need security? [23:57:06] <oninoshiko> not too much... we plan on running it on an isolated network in production too [23:58:22] <millhouse> I followed the ZFS send/receive steps to first send the initial backup and then incremental ones afterwards... My question is, is there anyway I can "merge" the incremental backups back into the initial? [23:59:52] *** derchris has joined #opensolaris [23:59:55] *** hohum_ has quit IRC