July 18, 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
July 18, 2007 [00:00:33] <CIA-26> zx151605: 6473224 ncpunode unused in sun4v kernel [00:00:34] <CIA-26> ja97890: 6577516 getsockname() on AF_UNIX returns all zeros in the socket address structure [00:00:57] *** tomww has joined #opensolaris [00:05:08] <coffman> stevel: speedstep? what cpu it has? intel core 2 duo or single core? [00:05:21] <stevel> coffman: single core [00:07:19] <coffman> hmm, any arrive date for the new installer? [00:07:48] <oxygene> coffman: more interesting: ETA for the new installer incl. new features (esp. zfs related) ;) [00:08:23] *** obsethryl has quit IRC [00:08:57] <coffman> oxygene: the demo seem to do it [00:09:25] <stevel> coffman: build 70 is what i hear [00:10:21] <coffman> oh i was not wrong... [00:11:15] <asyd> \_o< [00:11:25] <quasi> evenin asyd [00:13:06] <asyd> just a little question about zfs, most of slides use a hirearchical blocks data, with an uberblock as root block, but I'm not sure to understand if one uberblock can handle more than file, of one directory with multiple blocks pointers which handles multiples files [00:13:49] *** KermitTheFragger has quit IRC [00:14:38] <sommerfeld> the uberblock is the root of a tree of indirect blocks which point to all allocated blocks in the pool [00:15:09] <asyd> well, I understand this point [00:16:04] <sommerfeld> okay, then can you try restating the question? [00:17:13] *** m0le has joined #opensolaris [00:18:33] <sommerfeld> changes to the pool are grouped into transaction groups; updates to the tree of blocks are clustered together; at the end of the transaction group the last thing written are several copies of the uberblock [00:18:44] <asyd> hmmm well, I think my main problem with zfs is I don't understand well the relation between ZPL and DMU, specially with directories [00:20:29] <asyd> I can easily understand a file is destribe by uberblock and a lot of blocks pointers/data regarding the size of the file, I just wondering about how a directory is 'translated' in blocks pointers/datas [00:20:56] <alanc> MindDrive: I don't think we've changed xterm in Solaris in a long time [00:20:58] <asyd> s/d a/d how a/ [00:21:23] <boyd> Morning, all [00:21:48] <coffman> isnt xterm xsun? [00:22:13] <boyd> Anyone built wireshark on Sol10? Or better, got a pointer to some binary packages that don't require installing the universe (I'm looking at you, blastwave) [00:22:14] <sommerfeld> asyd: at a high level directories in ZFS are pretty much like directories in other unix filesystems -- data structures which map from file name to file number [00:22:36] <alanc> coffman: huh? Xsun is an X server, xterm is a client that emulates a terminal [00:22:46] <sommerfeld> asyd: and those data structures are stored in what are, to the lower levels, just specially marked files [00:25:36] <sommerfeld> if i'm not mistaken, at the DMU layer, a filesystem is just an array of numbered file objects. one of them is the root directory, and the rest are files and other directories. [00:25:53] <sommerfeld> ZPL and its parts actually gives meaning to the directory content [00:26:13] <asyd> ok [00:30:00] <asyd> btw, I gave a talk about OpenSolaris few days ago at rmll(.info), just after Alan Cox, and I'm happy because there are so many people interest by opensolaris nowadays :p [00:30:10] <boyd> In a way, that's very similar to the way that UFS can be thought of. An array of numbered file objects (inodes + blocks) some of which happen to be a directory [00:30:30] <sommerfeld> yup. zfs adds another dimension (multiple filesystems per pool) [00:30:42] <asyd> not trivial to understand :/ [00:31:42] <boyd> In fact, we could do away with directories entirely if humans would get with the program and just be happy with file numbers :) [00:32:46] <sommerfeld> actually, two dimensions if you consider snapshots [00:33:45] *** LuckyLuk1 has joined #opensolaris [00:34:32] <asyd> well, just a little another question about zfs. When you rollback a fs, zfs 'scrub' (i.e. parse the whole blocks used by the file system) to rollback (replace the newer by the older) blocks pointers ? [00:34:45] <asyd> (just to be sure) [00:35:24] <asyd> and the blocks data used by the filesystem which are not used by the snapshot are mark as free [00:37:04] <asyd> (yeah I know, my english is worst than usal) [00:40:34] <sommerfeld> asyd: something like that [00:41:11] *** Pietro_S has quit IRC [00:41:54] <asyd> ok [00:42:32] <sommerfeld> looks like the "meat" is in dsl_dataset_rollback_sync() [00:43:17] *** vmlemon_ has joined #opensolaris [00:43:55] *** vmlemon has quit IRC [00:43:56] *** vmlemon_ is now known as vmlemon [00:45:38] <asyd> sommerfeld: whaouh :p [00:45:51] <asyd> I need another life [00:46:53] <asyd> I think the useful line is "/* Free blkptrs that we gave birth to */" [00:47:44] <sommerfeld> yup. free everything newer than the snapshot [00:48:25] <asyd> and a little question.. about compression [00:48:38] <asyd> the blocksize is the same for a whole dataset, right ? [00:49:12] <asyd> but when compression is enabled, the blocksize is compute before or after the compression? [00:49:21] <sommerfeld> blocksize goes before compression [00:49:28] *** tfb has quit IRC [00:49:40] *** LuckyLuke has quit IRC [00:49:49] *** SirFunk has joined #opensolaris [00:49:50] <asyd> ok [00:50:20] <sommerfeld> so, conceptually, each file is chopped into <blocksize> chunks, then each chunk is compressed independently into a smaller blocksize [00:50:51] <sommerfeld> if the compression algorithm fails to make it meaningfully smaller, the block is written uncompressed. [00:51:50] <asyd> ahh, the compression is made before the 'block layer' ? [00:53:54] <asyd> hmm just few seconds to find the zfs architecture [00:56:22] *** jpdrawneek has quit IRC [00:56:28] <asyd> in http://www.opensolaris.org/os/community/zfs/source/ [00:56:40] <asyd> as far as I understand, the compression is made at the ZIO layer [00:56:52] <coffman> whats a normal io nr for intrstat on a usb device? [00:57:01] <asyd> but I thought the ZIO layer was not aware about 'file', but just block [00:57:17] <coffman> i got cpu0 about 3500 [00:59:36] <richlowe> asyd: compression doesn't need to be aware of 'files'. [00:59:42] <richlowe> it just needs a chunk of data to try and make smaller. [01:00:45] *** stevel has quit IRC [01:00:49] <asyd> well, that the point I have difficulties to understand. If there is a layer (DMU ?) which try to translate a file into a multiple chunk of (fixed) size of data [01:01:11] *** curlyman has joined #opensolaris [01:01:27] <asyd> where is the interest to compress these data while a 128k block of data consume the same diskspace for 120k of real data ? [01:01:38] *** hohum has quit IRC [01:04:44] <asyd> there are few things that are really (at least for my poor brain) to understad in zfs [01:05:21] <asyd> but I guess that the reason why it rocks, because it is a real fs/vm, not a toy :p [01:06:48] <sommerfeld> the reason it rocks is that it started from a blank sheet of paper, putting functionality where it made most sense rather than working on a preconceived notion about what layers made sense [01:07:06] *** nachox has joined #opensolaris [01:07:18] <nachox> evening fellas [01:07:28] <asyd> hehe [01:08:34] <asyd> thats "sux" the solaris internals 2nd edition doesn't cover zfs, but I can understand [01:09:04] <nachox> the book was likely written before zfs was released in solaris [01:09:09] <sommerfeld> and zfs is still a fast moving target [01:10:33] *** BadKarma has quit IRC [01:11:21] <asyd> anyway, thanks for your answers, time to sleep [01:11:24] <nachox> the ufs chapter is really cool though [01:11:34] <asyd> indeed [01:15:41] <asyd> & [01:18:00] *** vmlemon has quit IRC [01:18:08] *** aghaster has joined #opensolaris [01:23:38] <aghaster> I've made a mistake when installing opensolaris, I've allocated 512mb for the swap space [01:23:48] <timeless> ? [01:23:55] <timeless> you can add more swap later [01:23:56] <wesolows> how do you know that's a mistake? [01:23:59] <aghaster> opensolaris.org says building opensolaris requires a minimum of 1gig, recommended 2gigs [01:24:12] <timeless> man swap [01:24:15] <timeless> or something like that [01:24:16] *** pfa3rh has quit IRC [01:24:28] <timeless> you can swap to a file if you do it right [01:24:29] <aghaster> yeah, but if I want more swap I need to resize partitions. How do I do that? [01:24:41] <timeless> no, you don't need to do that [01:24:41] <aghaster> If I'd like to avoid adding a swap file [01:24:50] <aghaster> just to make it "clean" [01:24:51] <timeless> oh, well, yes, if you want to do that... [01:24:53] *** dgtl__ has joined #OpenSolaris [01:25:32] <aghaster> what tools are available to resize solaris partitions? [01:25:50] <wesolows> partitions, slices, or filesystems? [01:26:11] <dgtl__> Can anyone tell me how Solaris names drives? for instance, in Linux, the first SATA hard disk would be /dev/sda, second would be /dev/sdb, etc [01:26:17] <aghaster> hum... shrink my main partition where solaris is installed and grow my swap part [01:26:38] <aghaster> dgtl: open a terminal and enter format -l [01:26:41] <wesolows> you can resize partitions with fdisk, you can resize slices with format, but if there are filesystems on them, you'll want to resize those or else you'll have to recreate them. And that we don't really have. [01:27:12] <aghaster> dgtl: it is nothing related to the same command on windows, it will list hard drives and partitions [01:27:42] *** alfism has joined #opensolaris [01:27:49] <sommerfeld> wesolows: well, you can grow a filesystem if you change where it ends [01:27:50] <wesolows> dgtl__: It depends on whether multipathing is active. Normally it's /dev/dsk/cXtXdXsX [01:28:12] <wesolows> sommerfeld: I think that supports only increasing the size, though, right? [01:28:15] <dgtl__> wesolows, thank you very much. I'll give that a shot [01:28:29] <sommerfeld> yes. though Jim Carlson has code which will shrink UFS .. [01:28:32] <dgtl__> aghaster, thanks for the reply, but these are disks taht aren't part of a filesystem yet [01:28:32] <wesolows> dgtl__: SATA and ATA disks normally lack the 't' portion. [01:28:39] <aghaster> ok [01:28:47] <dgtl__> wesolows, ah, that's what I was doing wrong. thanks. [01:28:52] *** alfism has quit IRC [01:29:09] <richlowe> Huh, we really say 2G to swap to? lint used to hit /tmp (and core) pretty hard, but it's behaving a lot better now, depending on real memory, it may survive. [01:29:12] <richlowe> may not be pretty though. [01:29:31] <aghaster> wesolows: hum... I cannot do that while the partitions are mounted, I guess I can do that from the install disk? [01:29:34] <nachox> richlowe, storage is cheap [01:29:37] <wesolows> richlowe: I would hope we say you need a certain amount of memory + swap since there should be no need for any swap if I have, say, 16GB. [01:30:13] <wesolows> aghaster: You cannot resize a mounted filesystem, no. Nor will you be unable to umount your root filesystem while the system is running. So yes, you'll need to boot from alternate media to do any of this. [01:30:41] <aghaster> okay [01:31:00] <sommerfeld> actually, growfs can work on a live non-root ufs filesystem; it uses lockfs to lock out changes [01:31:26] <sommerfeld> (tried using it on the root partition and was sad..) [01:31:34] <dgtl__> I'm assuming the sX and pX are slice and partition numbers? [01:31:40] <wesolows> dgtl__: correct [01:31:43] <dgtl__> thanks [01:33:23] *** derchris has quit IRC [01:33:39] *** derchris has joined #opensolaris [01:37:02] *** curlyman has left #opensolaris [01:39:02] <alanc> hmm, I checked in a change to X.Org's libX11 which broke KDE - wonder if I get GNOME bonus points for that? [01:39:04] <dgtl__> so, I would assume that port one on a SATA controller would be c1d0 and then port 2 would be c1d1, but that doesn't seem the case. am I incorrect? [01:40:23] <wesolows> Usually SATA shows up as a separate controller for each port. [01:40:32] <wesolows> I'm not certain of the reason for this. [01:40:47] <dgtl__> aha [01:40:53] <dgtl__> this makes a lot more sense now, thank you [01:41:08] <wesolows> There are exceptions, though - the ones in Thumper, for example, show up as one controller per controller, with multiple ports on each one. [01:42:03] <nachox> alanc, so much for binary compatibility guarantee :P [01:42:31] <alanc> nachox: binary compatibility guarantee doesn't mean we'll never have bugs - just that we'll fix them when we do [01:42:51] <alanc> besides - my bug didn't break KDE on Solaris, only on UnixWare [01:43:05] <alanc> which is why it took 9 months for anyone to notice [01:43:53] <nachox> alanc, you just earnt some linux bonus points, join the war on sco! [01:44:28] <alanc> yeah, but that just is digging me out of the hole I got when I helped the SCO guy put the UnixWare & OpenServer support into X11R6.9 [01:45:34] <dgtl__> i'm having a difficult time locating these drives in my system. a Coworker mentioned that he used to know of a command a few years ago that would scan the device tree and recreate the appropriate nodes in /dev/dsk, but he can't remember what it was, or if it even exists in OpenSolaris. Any ideas? [01:46:49] <sommerfeld> devfsadm -v [01:47:04] *** EchoBinary has joined #opensolaris [01:47:05] <dgtl__> thanks [01:47:16] <sommerfeld> read the man page [01:47:43] <dgtl__> doing so right now. thanks :) [01:48:20] *** nrubsig has joined #opensolaris [01:48:21] *** ChanServ sets mode: +o nrubsig [01:48:23] <sommerfeld> (there's also an option to nuke links for devices that aren't there any more) [01:48:52] *** danv12 has joined #opensolaris [01:49:21] <nrubsig> alanc: Is it true ? Sun declared SCO/OS as "obsolete" and removed support from X.org tree ? It's now a featured article on slashdot.int ... [01:49:58] <coffman> fun! [01:50:32] * nrubsig watches alanc falling off his chair and then starting to throw the said chair into nrubsig's direction as punishment for bad jokes. [01:50:58] * nrubsig stares at coffman [01:51:10] <alanc> hasn't most of the world declared SCO obsolete already? [01:51:39] <CSFrost> most except those who are still using it, and they are not even aware of it [01:51:43] <coffman> alanc: pizza hut uses it [01:52:02] <nrubsig> alanc: Nope, they may still have some grown-up Ex-Atari users who are now CEOs somewhere in a cage... [01:52:36] <alanc> yeah, I seem to remember SCO powering a large portion of the worlds fast food cash registers in the past [01:53:20] <nrubsig> alanc: another market runied by linux [01:53:27] <nrubsig> er [01:53:29] <nrubsig> runied [01:53:34] * nrubsig checks his beer [01:53:39] <nrubsig> 7% [01:53:45] <nrubsig> *hicks* [01:54:01] <coffman> pretty much like netware in his field, except that netware does not suck that hard [01:54:20] <coffman> nrubsig: which sort of beer? [01:54:57] <CSFrost> alanc, if I remember correctly, a few financial organizations still have a foot left in SCO also [01:55:12] <nrubsig> coffman: flens [01:55:13] *** Trident has joined #opensolaris [01:55:28] <nrubsig> coffman: the remains of by briday [01:55:32] <nrubsig> broithday [01:55:33] <alanc> that's almost as scary as realizing how many ATMs are still running OS/2, or worse, migrating from OS/2 to Windows [01:55:35] <nrubsig> *hicks* [01:55:40] <nrubsig> birthday [01:56:27] <nrubsig> alanc: yeah, today I tried to get a train ticket... the machine didn't work: Blue screeen [01:56:32] <CSFrost> alanc, we were just joking about it earlier about a little bug that would turn the cameras into webcams :-P [01:56:33] <coffman> nrubsig: yours? when and how old your are now? [01:56:59] <nrubsig> coffman: 5.7.1975 [01:57:22] *** dvorak has quit IRC [01:57:41] <coffman> nrubsig: thats sad :P [01:57:44] <CSFrost> I am pretty sure atm's in Japan are on windows or some type of linux.. judging by the long pauses between hitting buttons :-P [01:57:46] <coffman> but flens is nice [01:57:48] <nachox> i've seen an sco unix running as a printer server [01:58:49] <coffman> pretty much every account printer in the banks run on os/2 [01:59:22] <coffman> i think i have some ecom station stuff laying around... [01:59:27] <CSFrost> except banks are going to windows for their entire front end.. it's so sad it's kind of funny :-P [01:59:48] <nrubsig> coffman: oh yes, sad story... give me a tree, rope, chair and a rotten fish that I can finish myself off... [01:59:59] <coffman> nrubsig: :P [02:00:51] <coffman> nrubsig: nah, we need you for our distro, some one has to do the scripting [02:01:10] <coffman> CSFrost: not all, i know a couple that go for linux in germany [02:01:29] <CSFrost> "We also have stuff that has been going to the list but isn't appearing on the forum." [02:01:38] <CSFrost> what's so bad about that? [02:01:39] <aghaster> I know nobody that goes for SkyOS anywhere. lol. [02:01:57] <CSFrost> coffman, well, anything is still better than windows :-P especially in a bank [02:02:26] *** dvorak has joined #opensolaris [02:03:46] <coffman> i know why to much what is going on in the german flight controll... [02:03:58] <coffman> *\why\way [02:04:46] <CSFrost> "Sir, what do I do, I am at 30,000 meters and I have a blue screen" [02:04:55] <CSFrost> "look down instead of up" [02:04:58] <CSFrost> "I am looking down" [02:05:15] <CSFrost> doh :-P [02:05:46] <coffman> nah, more like the things in the controll systems, where the backup system lack parts that are digital in the main and such [02:05:51] <coffman> pretty scarry [02:06:23] <CSFrost> all systems are lacking something, it's part of the new flight standards. [02:08:21] *** bubbva has quit IRC [02:08:22] *** m0le has quit IRC [02:08:34] <sommerfeld> friend of mine from college is a flight instructor. he let loose with a rant on his blog today about instructors who let their students solo too early. [02:09:58] <CSFrost> with windows, your always solo :-), or is that SoL? [02:10:11] <CSFrost> sorry, couldn't help it.. [02:10:37] <coffman> after my school i think i will apply for flight controll [02:10:44] <coffman> maybe i can make it [02:13:44] <CSFrost> I'm thinking fly fishing... [02:18:43] *** Kprofthreat has joined #opensolaris [02:19:29] <Tempt> Morning all! [02:19:42] <Tempt> Oh, and OS/2 was the traditional platform for ATMs. [02:19:55] <Tempt> Now replaced by Windows. [02:21:13] <CSFrost> so we can theoretically rig up a atm card to just start spitting out cash? :-P [02:21:39] <wesolows> No, you can never get Windows to do what you want. [02:21:46] <nrubsig> heh [02:21:49] <nrubsig> wesolows: +1 [02:21:54] <wesolows> It never works right, but never in a way that helps you, either. [02:22:09] <CSFrost> hehe, then we make it do the opposite! right? [02:22:18] <CSFrost> ahh.. right right [02:22:27] <nrubsig> CSFrost: there is always a 3rd way to do it wrong [02:22:52] <jbk> heh a friend of mine was just lamenting (does support for M$) how much of a PITA it is apparently to walk poeple through starting a kernel debugger on windows [02:22:53] <CSFrost> I am sure with trial and error it would be possible.. [02:22:57] <nrubsig> CSFrost: and the univerise is infinite and therefore the numbers of bugs in Windows is infinite, too [02:23:06] *** Fish has quit IRC [02:23:37] <alanc> who needs to exploit Windows bugs when people deploy ATM's with the default passwords still in place, and the ATM owners manuals with those passwords available on the net? [02:23:41] <CSFrost> nrubsig, so then just not try to do that.. and sooner or later it will happen? :-) [02:23:42] <alanc> http://blog.wired.com/27bstroke6/2007/07/atm-reprogrammi.html [02:24:21] <coffman> i like voting machines [02:24:22] <nrubsig> CSFrost: what will happen ? Extinction of mankind ? Sure... [02:24:47] <wesolows> coffman: Why? If you tampered with them so that a good candidate actually won, everyone would know it's fraudulent. [02:24:56] <nrubsig> heh [02:25:33] <nrubsig> wesolows: noone will notice it when the "good canidate" is from the republican party [02:25:49] <coffman> wesolows: they are shit, but why not exploit it then? [02:26:11] <CSFrost> No No.. your all wrong [02:26:14] <wesolows> nrubsig: I guess my point was that the only candidates considered "viable" (that is, are from the 2 largest parties) are invariably bad. [02:26:17] <CSFrost> there is no good candidate. [02:26:22] <coffman> if they are coming up here more i will be happy to have some fun with them [02:26:32] <CSFrost> You'd need to make it write in someone.... [02:26:33] <Trident> CSFrost: Just for your info as you tried to help with my strange network problem with snv_62 some weeks ago. It was: http://bugs.opensolaris.org/view_bug.do?bug_id=6545032 that bit me. [02:26:47] <CSFrost> like.. Robin Williams.. [02:27:07] <wesolows> CSFrost: Oh...I've heard he holds "San Francisco values." [02:28:58] <FrostCS> wesolows, (other terminal froze) I was referring to the movie he was in.. where he ran for president and won, because the machines were rigged. [02:29:16] <wesolows> well, that shows how many movies I see [02:29:38] <FrostCS> yea, I was recommended to watch it several times before I actually did.. [02:29:45] <FrostCS> though, can't even remember the name of it [02:30:44] <FrostCS> Trident, so your up to current, and it is working now? [02:31:09] <sommerfeld> http://www.imdb.com/title/tt0483726/ "Man of the Year" ? [02:31:27] * FrostCS honestly would be lucky to remember what happened yesterday.. let alone some weeks ago.. [02:31:38] <Tempt> What day was yesterday? [02:31:43] <FrostCS> that sounds about right, sommerfeld [02:31:48] <FrostCS> Tempt, Friday. [02:31:59] <aghaster> should I use the "format" tool to resize my partitions? [02:32:21] <nachox> monday, july 16th [02:32:26] [02:32:47] <sommerfeld> yes, "format" is the tool for that. [02:33:02] <aghaster> proper for shrinking without destroying data? [02:33:59] <sommerfeld> aghaster: uh, no, it just cuts the end off of the slice. it will destroy a filesystem [02:34:14] <sommerfeld> the command to write the changes is "label" in the partition> menu [02:34:26] <aghaster> thanks for telling me. how do I shrink/grow partitions without destroying the filesystems? [02:34:31] <nachox> i thought growfs could only grow the fs, i doubt you can shrink ufs [02:34:46] <aghaster> crap, you mean I can't shrink it? [02:35:20] <FrostCS> you can move the data to another, then arrange, and then move it back [02:35:21] <nachox> aghaster, you can always backup, repartition and restore the backup [02:35:26] *** CSFrost has quit IRC [02:35:35] <aghaster> meh, this is a fresh install [02:35:43] *** CSFrost has joined #opensolaris [02:35:46] <FrostCS> although nachox's way sounds more professional by his choice of words lol [02:35:57] <aghaster> this means I need to reformat just to get more swap space other than from a swapfile [02:36:16] <nachox> what's wrong with swap files? [02:36:25] <Trident> FrostCS: Havn't had the time to upgrade yet, it's not a production machine after all. But the description fits just right, both the actual manifestation of the problem and that there are multiple NCEs when not working. And the workaround works, now I am only seeing the problem a few times a week instead of several times a day. [02:36:53] <aghaster> just a habit. do they work as good as a swap partition? [02:37:13] <FrostCS> Trident, you could run a live update on it right? [02:37:14] *** baijiutong has quit IRC [02:37:23] <CSFrost> or just start from scratch again.. :-) [02:37:25] *** nrubsig has quit IRC [02:37:28] <FrostCS> live upgrade.. whichever wording they use now.. [02:37:59] <FrostCS> yay, other irc is back up [02:38:12] <aghaster> I'll go for a swap file, then [02:39:37] <Tempt> You cannot shrink UFS. You can shrink VxFS, if you're using that. [02:40:18] <Tempt> Of course, using VxFS probably means using VxVM, and if you're using VxVM the first thing to shrink is your patience [02:40:32] <nachox> hehe [02:41:12] * nachox started reading the svm documentation last night and it is actually surprised it makes a lot of sense [02:45:25] <aghaster> why does opensolaris install UFS by default, instead of ZFS? [02:46:06] <CSFrost> because the new installer is still in alpha stages [02:46:27] <wesolows> The actual answer is "it doesn't; opensolaris isn't a distribution and doesn't have an installer" [02:46:51] <CSFrost> hehe, there I go assuming he meant SXCE :-P [02:47:00] <nachox> opensolaris is just code and a community (various actually) around it [02:47:03] <sommerfeld> as well as "because ZFS boot and install isn't done yet" [02:47:22] <Tempt> nachox: SVM is pretty simple stuff. It does actually make sense and is useful. [02:47:29] <CSFrost> though doesn't nexenta or belenix have zfs on the installer? [02:47:51] <Tempt> Because it doesn't really matter? [02:48:28] <nachox> Tempt, is it reliable too? [02:48:51] <richlowe> CSFrost: I think so. [02:48:52] <Tempt> I'd have more faith in SVM for mirroring rootdisks than ZFS. [02:49:03] <Tempt> Because SVM/SDS has been around a looooooooong time doing just that. [02:49:27] <wesolows> It's an interesting tradeoff between dramatically better reliability features and a much longer track record. [02:49:41] <Tempt> Track records count. [02:49:49] <wesolows> I just said that they do. [02:50:02] <wesolows> But I don't share your overwhelming preference for them. [02:50:03] <Tempt> There are still some tricks you can pull in SVM that you can't quite pull in ZFS yet. [02:50:23] <Tempt> I don't think I'd use SVM for anything other than simple mirroring anymore though. [02:50:29] <nachox> having to build a 2 stats databases in the root mirrors really sounds like a big bad hack but it makes sense [02:59:00] <coffman> the new installer is coming in nv70 [02:59:13] <coffman> so will it be zfs root default on x86 then? [02:59:45] <richlowe> Dwarf Caiman isn't all of "The new installer" [03:00:16] <richlowe> just a differently stripped-down hack for DE [03:00:58] <coffman> kay, so the backend remains the old? [03:02:05] <coffman> based on svr5 packages? [03:02:55] * nachox can almost see a complaint about svr5 packages on the way [03:03:22] <wesolows> Someone did an svr5? That's cool. [03:03:54] <coffman> nachox: nah, im fine with it [03:03:59] <richlowe> wesolows: I think SCO claim something of there's to be SysVr5 [03:04:04] <richlowe> but I'm guessing a typo. [03:04:13] <richlowe> I'm also yet to see a compelling reason to move away from v.4 packages. [03:04:22] <richlowe> never mind why someone would think Caiman would be the project to do it. [03:04:47] <wesolows> richlowe: Well, the compelling reason is that it's old, and not hip, modern, trendy, and cool. [03:05:02] <coffman> richlowe: its not about the packages in general its only about the installer [03:05:04] <nachox> s/5/4/; #ops [03:05:05] <richlowe> it's also a package. [03:05:08] <wesolows> richlowe: Same reason everything in SFW has to be "updated". [03:05:23] <richlowe> is there a package format that isn't pretty much a tarball in a wig? [03:05:29] <richlowe> (or cpio archive, or ar, or $ARCHIVER_OF_CHOICE) [03:05:40] <wesolows> It doesn't really matter whether something is broken, see, it just matters that the kids today are comfortable and familiar with it from "other systems." [03:05:42] <coffman> building the database, adding every package one by one is just not the "fast" [03:06:05] <nachox> richlowe, slackware's? :) [03:06:21] <wesolows> coffman: Very true. But the database is not an interface, so one could change the way that works without changing the package format itself. [03:06:59] <coffman> wesolows: right [03:07:18] <nachox> iirc someone did say here that some engineers tried to use a db for that and it wasnt cool [03:07:28] <wesolows> nachox: I've heard that as well. [03:07:31] <coffman> wesolows: i was thinking of a lives system where you images with bin diff or such [03:07:34] <wesolows> I think Bart was involved in it somehow. [03:07:39] <coffman> *have [03:07:47] <dlg> maybe someone could profile and see where its slow [03:08:20] <wesolows> I guess I just don't see the problem with the speed of installing packages. [03:08:21] *** slowhog has quit IRC [03:08:32] <wesolows> It if takes an hour, well, ok. So what? I only do it once. [03:08:49] <nachox> usability gap [03:08:57] <coffman> or putting a complete system on the disk and then remove the packages... [03:09:01] <nachox> in linux packages install in seconds :P [03:09:09] <wesolows> It's a lot more important to me how well the system works *after* I'm done installing it. [03:09:31] <wesolows> nachox: Because, of course, the kernel provides the install_package syscall. [03:09:44] <wesolows> See, if we'd just implement install_package(2), we'd be set. [03:09:51] <nachox> it doesnt? it really should! [03:10:01] *** seanmcg has quit IRC [03:10:21] *** seanmcg has joined #opensolaris [03:10:22] <wesolows> But those Linux guys, they're smart and they patented install_package(2), so they'll forever have a "usability gap" ahead of us. [03:10:34] <coffman> sure, i dont complain that much about the speed, i would more like to see zfs root which would make lu a much more easier [03:10:44] <wesolows> coffman: Now you're talking. [03:10:47] <nachox> this conversation is testing the known limits of sarcasm [03:11:21] <coffman> and i would like to have a way updating my system without having to download a whole dvd [03:11:38] <wesolows> nachox: It may be, but at least we don't have to contend with the missile gap and the mineshaft gap. [03:11:41] <richlowe> that's not the packaging format either. [03:11:54] <wesolows> If we had those gaps, Linux could steal ou rprecious bodily fluids! [03:11:57] <nachox> now, that would be usefull but it's not related to the package format [03:12:15] <coffman> sure :P i was just asking if the backend changed :P [03:12:16] <richlowe> hell, pkgadd will even do so via http (and https?) [03:12:49] <coffman> so, how about zfs root, does it come in nv70 with caiman? [03:12:50] <nachox> why do you need https when packages have digital signatures themselves [03:13:29] <coffman> nachox: wasting computing time :D [03:13:49] <wesolows> coffman: See http://www.opensolaris.org/os/project/caiman/Dwarf/ [03:14:40] <wesolows> coffman: Short answer is no. [03:14:54] <nachox> see? a lot less typing [03:14:57] <wesolows> In fact, looking at that page, I don't see how dwarf caiman adds any value whatsoever [03:15:23] <wesolows> But I tend toward the cynical, so I'm probably shortchanging them. [03:15:24] <richlowe> wesolows: depending on how you read it, it makes DE do the right thing finally [03:15:30] <nachox> wesolows, to start with it doesnt need n jvms to run [03:15:33] <richlowe> wesolows: (ie, toast your data and make you regret having anything to do with it) [03:15:57] *** LuckyLuke has joined #opensolaris [03:15:59] <wesolows> richlowe: Sure. That's a nice simplification, but I don't see how it makes anyone happier. [03:16:08] <wesolows> Also, SXDE is pointless and stupid anyway. [03:16:17] <wesolows> Who cares if it upgrades Netbeans?! [03:16:23] <nachox> i agree there [03:16:29] <wesolows> Why would I want netbeans on my system in the first place? [03:16:32] <wesolows> sigh [03:16:34] <wesolows> back to work [03:18:12] <nachox> everything is a marketing move, the netbeans guys needed better stats so it is now included in sxde [03:18:42] <coffman> nachox: its included in studio... [03:18:59] <coffman> gar shity dwarf, no zfs, no need [03:19:02] <nachox> again, another marketing move :) [03:19:03] <CSFrost> I dont think DE is going to help the Caiman guys any.. [03:19:15] <richlowe> see, the sarcasm missed. [03:19:20] <CSFrost> since they get to slack off and only update it every 6 months with DE... :-P [03:19:27] <richlowe> wesolows: the "... toast your data ..." was intended to imply the pointlessness of DE [03:19:50] <richlowe> though, given the hell it causes, perhaps making it more violently destructive on that end will help close the violent-rage gap [03:19:55] <richlowe> we're all about closing vaguely defined gaps, right? [03:20:26] <wesolows> richlowe: Well, I guess. But I don't store valuable data on workstations, so the idea of a workstation install toasting all my data doesn't necessarily mean that workstation install is pointless. [03:20:45] *** mega has quit IRC [03:20:51] <wesolows> richlowe: Dude, I close the violent rage gap all by myself. [03:21:15] <nachox> all those engineers working in packaging and testing sxde when they could be doing something more usefull like.... anything... [03:21:28] <wesolows> nachox: yeah, it makes you wonder, eh? [03:21:56] <coffman> *sigh* [03:24:33] <g4lt-U60> nachox, would you prefer they were packaging and testing project indiana? ;P [03:24:50] <nachox> speaking about the dwarf page, shouldnt it be better to stack milestones instead of queing them? this way we need to go to the bottom of the page to see the actual status of the project [03:25:40] *** mega has joined #opensolaris [03:26:50] <coffman> g4lt-U60: i would prefer that they would remove the last binary blobs and remove stupid depends, so that making a distro would be easy [03:27:07] <coffman> depends like wbem and such [03:27:40] <nachox> g4lt-U60, i'm sure there are more productive things to do, for example they could teach wesolows there is a bright side in life and not everything is sarcasm [03:27:52] *** Tpenta has quit IRC [03:27:52] <richlowe> nachox: where's the fun in that? [03:28:21] *** Tpenta has joined #opensolaris [03:28:38] <nachox> richlowe, it'd save a lot of money in broken sarcasm-meters [03:29:35] *** ChanServ sets mode: +o Tpenta [03:30:01] <g4lt-U60> coffman, the dwarf caiman project reminds me a lot of one of my ex cow-orkers, she spent so much time fussing with design documents she never actually coded [03:30:49] *** LuckyLuk1 has quit IRC [03:33:49] *** yongsun has joined #opensolaris [03:45:30] <comay> g4lt-U60: have you been following the project? you know they have a mailing list, right? [03:45:41] <g4lt-U60> yep [03:45:53] <g4lt-U60> that's why I said what I said [03:46:20] <comay> well, given that the code is behind the firewall (since much of it is still encumbered), it seems [03:46:35] <comay> they've done a fairly decent job keeping folks apraised of their progress [03:46:38] *** halton has left #opensolaris [03:46:48] <richlowe> It's somewhat funny that everything is open *other* than the code. [03:46:50] <comay> i thought they had a code drop or two [03:46:54] <richlowe> comay: no. [03:47:11] <comay> well, i'm sure parts of the code could be made open but it wouldn't be [03:47:38] <richlowe> they've dropped do-nothing GUI demos. [03:47:41] <coffman> who should i talk with about merchandising stuff like dvds stickers ? [03:47:46] <comay> buildable since the backend is still pfinstall [03:48:08] <coffman> i would like to have such stuff since ill do some speeches on opensolaris soon [03:48:33] <richlowe> comay: maybe. [03:48:34] <nachox> "and this is what our new installer will look like!!" [03:49:45] <alanc> coffman: someone from the marketing community like Sara or Teresa I think [03:51:18] <nachox> marketing people do not visit this channel because they fear being beaten to death? [03:52:03] *** halton has joined #opensolaris [03:52:37] <coffman> i think benr is also in marketing [03:52:42] <CSFrost> beaten? hrm... [03:52:51] <CSFrost> benr doesn't work for Sun. [03:54:00] <coffman> CSFrost: its about opensolaris community marketing, not sun marketing [03:54:21] <g4lt-U60> gman's in marketing now ;P [03:54:28] *** mega has quit IRC [03:54:43] <CSFrost> where oh where does the other side start.. and this side begin.. [03:54:57] <nachox> indiana != marketing [03:55:14] <CSFrost> of course it does [03:55:21] <CSFrost> linux = solaris marketing [03:55:27] <CSFrost> goblins = cows [03:55:34] <CSFrost> dentist = lawyers [03:55:37] <g4lt-U60> no, he got moved to marketing just before the PI announcement [03:55:38] <CSFrost> me = good at math [04:01:15] <Tempt> CSFrost: You been breathing too many traffic fumes lately? [04:03:08] <coffman> i think its realy about getting people that use linux over to the solaris camp, but the right ones, and they dont care about a fancy installer or if the gnu tools are in /usr/bin [04:04:34] <CSFrost> Tempt, probably, been reading too much political agenda [04:09:02] <richlowe> I doubt anyone cares about it only being the "right" ones. [04:09:12] <richlowe> and I'd bet that no two people would ever agree on what who the "right" ones were. [04:09:41] *** Yamazaki-kun has joined #opensolaris [04:11:57] <nachox> the right people are always the ones willing to pay for support contracts [04:12:38] <coffman> nachox: really? [04:12:45] *** mega has joined #opensolaris [04:13:13] <coffman> i think the right people are the people willing to work on opensolaris [04:13:32] <nachox> only if you want to keep your software company proffitable and running, yes [04:15:40] <jamesd_> has anyone heard of a solaris 10u3 bug where the t2000 + automounter crashes the system [04:15:59] *** LuckyLuk1 has joined #opensolaris [04:19:13] <coffman> jamesd_: hey there, was it an accident you gave me op? :P [04:19:54] <jamesd_> coffman, no... you were alive, and i was going off line soon, and figured we needed at least one live op, and i thought you could handle the job [04:20:40] <coffman> jamesd_: thx for the trust [04:21:05] <jamesd_> coffman, no problem [04:25:01] <jamesd_> hmm no sun engineer is interested in a nice t2000 + automounter bug? the thing dropped a nice core and everything... [04:25:56] <jamesd_> debugging crash dump vmcore.0 (64-bit) from iusza3200 [04:25:56] <jamesd_> operating system: 5.10 Generic_118833-36 (sun4v) [04:25:56] <jamesd_> panic message: [04:25:56] <jamesd_> BAD TRAP: type=31 rp=2a1032437c0 addr=18 mmu_fsr=0 occurred in module "genunix" due to a NULL pointer dereference [04:25:56] <jamesd_> dump content: kernel pages only [04:26:24] *** bondolo has quit IRC [04:27:32] <nachox> maybe it's too late at night? [04:28:56] <jamesd_> i guess.... just trying to debug it now.. since i can't access this awesome resounrce during the day... we have firewalls from hell at work [04:29:10] *** yongsun|wfh has joined #opensolaris [04:30:19] <nachox> as long as you can have the core available when there are other engineers online... [04:30:23] <jamesd_> i just hope i dianosed it right... i found a solaris bug that was fixed in a later version than what we are running, but its strange that the automounter triggered a bug. [04:30:37] *** LuckyLuke has quit IRC [04:30:51] <jamesd_> all i have right now is stack traces... and that status message. [04:30:52] <nachox> jamesd_, btw, i thought you didnt have a job [04:31:05] <jamesd_> i just got it 3 weeks ago... started 7/2 [04:31:18] <nachox> congrats :) [04:31:56] <jamesd_> i get to baby sit 140 servers, and god knows how many idiot application programmers that think its a great idea to store busy log files into /tmp [04:32:19] <nachox> 140 servers? nice [04:32:23] <jamesd_> then bitch when it takes down the server and there is no copy of the log file... [04:32:40] <jamesd_> nachox, yeah the unix team has 140 servers mostly fujitsu boxes [04:33:36] <nachox> proposed solution, dont use tempfs for /tmp, colateral effect decreased performance [04:33:55] <nachox> jamesd_, but youre not the only admin i assume [04:34:22] <sommerfeld> what's worse is when junior engineer fills /var/tmp (filling the root partition), apologizes, says he won't do it again, and then fills the root partition a new way [04:34:47] <jamesd_> nachox, i know that, we bitch at them... but they dont listen... of course they bitch when we take the boxes down twice a year for patching with a months notice... but think nothing about rebooting the box because of there idiot decision at a moments notice [04:34:59] <nachox> sommerfeld, a semisenior admin forgetting to use quotas? [04:35:00] <jamesd_> nachox, i'm part of a team of 5 admins [04:35:25] *** bondolo has joined #opensolaris [04:35:26] <jamesd_> sommerfeld, have you heard anything of a bug where the automounter takes down a t2000? [04:36:28] <jamesd_> nachox, quotas are outlawed at this place... they allready tried to get that past operations. [04:37:37] <nachox> ouch, quotas are actually usefull [04:39:30] <jamesd_> nachox, i know... but they are unfriendly to line of business ( they are employees of the same company, but since they are a different division, they by contract are treated like customers, so they can bitch like an outside customer) or some crap that says quotas are outlawed... [04:40:37] *** spackest has joined #opensolaris [04:40:38] <jamesd_> but hey... if they want to pay me in the high xx,xxx.00 i will reboot there boxes when they go crazy and follow the crazy rules, just don't blame the BS on me. [04:41:08] <spackest> so, trying to get apache2 working on solaris 10 (build 68) using mod_perl, using packages from blastwave [04:41:22] <nachox> jamesd_, i guess you cant make them use sunrays right? they need the processing power [04:41:29] <spackest> I finally got it to start, and now it core dumps like five seconds after it starts [04:41:48] <jamesd_> nachox, been talking up the sunrays.. most people have laptops for desktops... [04:42:23] <coffman> spackest: you might want to consider to ask in #blastwave ... [04:42:30] *** yongsun|wfh has quit IRC [04:42:54] <jamesd_> but of course they are wedded to MS line of crapware... even the unix guys have to use MS mail bloatware... because of encryption and constant callendar events... [04:43:18] <coffman> spackest: but go on, wich apache, blastwaves? [04:43:33] <spackest> coffman: yes, apache2 from blastwave [04:43:39] <jamesd_> and the answer to any documenting need is excel, and heat tickets... [04:44:09] <nachox> jamesd_, well, sun has a connector... but yes, most people default to windows, it's already installed in the laptops anyway [04:44:46] *** Yamazaki-kun has quit IRC [04:44:59] <nachox> pray your managers do not find out about ms visio.... [04:45:00] <sommerfeld> jamesd_: nope [04:45:02] <jamesd_> you have just got to love... filling out a excell spreadsheet with data that should of been grabbed and exported using a shell script, box type, cpu count and speed, ram amount, and disks inventory. [04:45:07] <sommerfeld> nachox: quotas not an option in this case [04:46:47] <nachox> jamesd_, god bless sed and awk [04:47:01] *** Dink has quit IRC [04:47:54] <jamesd_> nachox, exactly... i wrote scipts to dig out the info, but then i had to sit and copy and paste, it because another person on my team had gone out and did 1/2 of it by hand... so i just couldn't recreate the damm thing. [04:48:39] <richlowe> export to csv, append, import? [04:48:45] <nachox> nod [04:48:56] <jamesd_> nachox, i'm still new didn't want to step on too many toes... and make the other guy look bad.. [04:49:23] <nachox> jamesd_, IT workers do not have feelings [04:49:31] <nachox> :P [04:50:28] <spackest> anyone use apache2 with mod_perl(2) in solaris 10? [04:50:38] <spackest> I am up for whatever, it doesn't have to be blastwave [04:50:47] <spackest> generally #blastwave is pretty dead [04:50:51] <spackest> but I did post there [04:50:56] <jamesd_> nachox, yeah right, as it was i sat there and helped figure out that a disk model name "Sun 9G" was a 9GB drive, and told him that slice two is the full disk size... he had no clue about either fact. and he has only been on the unix team for a year, and 99% of our boxes are solaris 8 [04:52:31] <spackest> when is that indiana distro coming out? cause I gotta tell ya, the linux to solaris path is a tough one [04:52:38] <nachox> jamesd_, stuff like that makes me feel i can get a job managing solaris boxes :P [04:52:39] <coffman> spackest: you might consider using cooltools http://cooltools.sunsource.net/coolstack/index.html its an optimized stack for solaris 10 with apache perl mysql php etc [04:52:54] <spackest> and for whatever reason I really, really want solaris 10 to work for me :) [04:53:00] <spackest> coffman: thanks, will take a look [04:53:36] <jamesd_> nachox, me too... of course this guy also didn't know how to do global search and replace in vi, and didn't feel the need to research the subject when making 20 changes to the file, yet claimed to know "sed" [04:54:58] <nachox> i dont blame them, i dont like vi either.... i do know sed quite a lot though [04:55:29] <nachox> i have to admin that you need to learn vi because it's the standard editor everywhere [04:55:50] <spackest> coffman: so that still wouldn't get me mod_perl, is that right? [04:56:01] <spackest> coffman: I have apache2 running without mod_perl [04:56:08] <spackest> just not with :) [04:56:28] <nachox> *admit [04:57:05] <coffman> spackest: the package ships with mod perl [04:57:14] <jamesd_> nachox, this place only editor on 99.9% of the boxes running solaris 8 is vi, there isn't any other editor availible, after a year of maintaining solaris boxes using vi, you have to learn how to do it. [04:57:21] <Tempt> coolstack includes mod_perl, just checked. [04:58:26] <nachox> jamesd_, yep, no matter what unix youre using, the only thing you can be sure of is that there is vi [04:59:19] <spackest> hmm, they list php [04:59:20] <nachox> that doesnt mean i have to like it, i just need to know how to use it... [04:59:41] <spackest> and CSKperl, which I guess includes mod_perl? [04:59:55] <jamesd_> nachox, and after a year, you would at least figure out search and replace... because its quite common for a change to effect 20 or more lines. [05:00:25] <nachox> or you'd just use sed [05:01:00] <jamesd_> but if you can use sed you can use :%s/this/that/g [05:01:52] <nachox> how did that guy get a job managing solaris anyway? [05:01:55] <jamesd_> or in this case :%s/user/user, user2/ [05:01:58] *** cmang_ has joined #opensolaris [05:02:03] <jamesd_> nachox, no idea.... [05:02:23] *** sommerfeld has quit IRC [05:02:43] <coffman> spackest: " [05:02:52] *** rbrown has joined #opensolaris [05:03:00] <coffman> spackest: " Apache httpd is built with MPM pre-fork and modules for LDAP, Cache, , PHP, [05:03:03] <coffman> SSL, mod-deflate and Perl. [05:03:14] <spackest> ahh, cool [05:03:17] <coffman> thats out of the doc [05:03:19] <nachox> jamesd_, these days all job interviews include some basic shell scripting problem you need to solve, i should know since i did today :P [05:03:38] <Tempt> The question is, why fork whole processes instead of just running up a nice stack of threads? [05:04:08] <nachox> to not share the address space? [05:04:22] <nachox> :) [05:04:37] <jamesd_> nachox, i know about a month ago i did the same... but this places favorite example is why is this script bad cd location/x ; rm * why is this bad... [05:04:58] <jamesd_> actually it was rm -rf * [05:05:18] <nachox> if location/x doesnt exist i suggest you start looking for another job [05:05:55] <jamesd_> yeap.. that is the correct answer in this case it was the nfs server was down. so the cd command failed.. [05:06:11] <jamesd_> but course it was restored via tape backup [05:07:21] <coffman> wildcards are nothing to play with [05:07:42] <richlowe> zsh's extended globbing sure is. [05:07:53] <richlowe> pretty much the only way to understand it, is to see how much of your filesystem is left afterward. [05:07:56] <nachox> cd location/x && rm -rf * is safe [05:10:41] <Tempt> Heh. [05:10:52] <Tempt> I've seen someone rm -rf / a box they were about to scrap. [05:10:59] <Tempt> Too bad they didn't unmount the NFS shares first. [05:11:16] <nachox> Tempt, if they run solaris 10, it doesnt matter [05:11:46] <Tempt> Well, this wasn't even Solaris. [05:11:51] <nachox> and if the root has nobody access to the shares it doesnt matter either [05:12:34] *** jdavis has quit IRC [05:16:40] *** baijiutong has joined #opensolaris [05:19:59] *** JWheeler has quit IRC [05:24:10] *** alanc is now known as alanc-away [05:25:00] <nachox> solaris 10 u4 will be released next month, nice [05:25:01] *** spackest has quit IRC [05:32:53] <Tempt> Indeed, very nice. [05:32:54] *** bondolo has quit IRC [05:36:11] *** boyd_ has joined #opensolaris [05:37:08] *** Fullmoon has quit IRC [05:37:47] *** nachox has quit IRC [05:38:27] *** kaiwai has joined #opensolaris [05:41:55] *** boyd has quit IRC [05:42:04] *** boyd_ is now known as boyd [05:42:07] *** mega has quit IRC [05:43:14] <boyd> richlowe: Ha! [05:44:29] <kaiwai> anyone got the email address of Sophia Li? [05:46:49] *** fedorared has quit IRC [05:50:38] *** aghaster has quit IRC [06:32:30] *** Peanut_ has joined #opensolaris [06:38:26] <theRealballchalk> no who's she? [06:38:53] <theRealballchalk> how yall doing? [06:39:02] <theRealballchalk> i've been dormant for a while [06:39:03] <theRealballchalk> haha [06:39:57] <bda> Are you just now coming out of your chrysalis for the summer? [06:44:07] <theRealballchalk> yea been quite busy with school if u ask me [06:44:07] *** lloy0076 has joined #opensolaris [06:47:48] *** slowhog has joined #opensolaris [06:47:50] *** Peanut has quit IRC [06:53:14] *** Jondice has quit IRC [06:58:16] *** Jondice has joined #opensolaris [07:10:46] *** delewis has joined #opensolaris [07:22:27] <bda> Wawa closed until 5am for a "technology upgrade". If that doesn't mean they're getting Lucky Dragon-style nanotech replicators, I'm going to be pretty pissed. [07:23:34] *** boyd_ has joined #opensolaris [07:30:14] <g4lt-U60> yes, but do you want nekkid chix0rz hopping out of the faucet? [07:30:28] *** comay has quit IRC [07:36:27] *** boyd has quit IRC [07:36:47] <hile_> WaWa closes? [07:36:54] <hile_> that's fucked up.... [07:37:44] <bda> Yeah, I really wanted a chicken sammich. :< [07:39:47] *** jcea has joined #opensolaris [07:41:09] *** LuckyLuke has joined #opensolaris [07:47:37] <_mary_kate_> is wget in S10 not capable of transferring large files? [07:47:42] *** monzie has joined #opensolaris [07:47:59] <delewis> that's a limitation of HTTP/FTP. [07:48:00] <delewis> not wget. [07:48:12] <delewis> you don't transfer > 2GB files over HTTP or FTP. period. [07:48:17] <_mary_kate_> delewis: no it's not, i've transferred large files fine in the past [07:48:26] <dlg> delewis: it can so do it [07:48:37] <dlg> content size is represented by a string on the wire [07:48:58] <dlg> also, optional [07:49:40] <delewis> sure, the RFC doesn't say anything about it, but almost all of the HTTP servers and clients that I've seen don't support large files. [07:49:51] <delewis> example: try creating a > 2GB flash archive, and using HTTP/FTP. [07:50:04] <delewis> the installer will barf and laugh at you in a very subtle manner. [07:50:31] <_mary_kate_> yes, there are quite a few broken clients. (like wget, apparently :) [07:50:33] <dlg> apache on a platform with sane types for offset_t is fine on the sender side [07:50:51] <dlg> the client i used is fine with it [07:53:39] *** gdamore has joined #opensolaris [07:53:51] <gdamore> hi * [07:53:56] <richlowe> from memory, the installer is fine with large flars [07:54:02] <richlowe> finding something to serve it, however, can be more irritating. [07:54:25] <_mary_kate_> oh well, compiled my own lftp and it works fine :) [07:57:12] *** LuckyLuk1 has quit IRC [08:00:04] *** theRealballchalk has quit IRC [08:00:38] *** boyd has joined #opensolaris [08:04:44] *** Gropi has joined #opensolaris [08:08:03] *** mihaic has joined #OpenSolaris [08:08:42] *** sahafeez has joined #opensolaris [08:08:54] *** mihaic is now known as cmihai [08:12:58] *** boyd_ has quit IRC [08:17:25] *** dunc has quit IRC [08:18:33] *** Cyrille has joined #opensolaris [08:19:12] *** danv12 has quit IRC [08:25:01] *** slowhog has left #opensolaris [08:26:48] <e^ipi> I wonder how one breaks in to consulting [08:26:54] <e^ipi> because those guys make a lot of money [08:27:42] <g4lt-U60> get fired [08:27:49] *** danv12 has joined #opensolaris [08:28:23] <bda> ha [08:29:52] <e^ipi> fired from what, i sling coffee for a living [08:30:02] <e^ipi> which is like unemployment, except i make barely enough to pay rent [08:30:42] <lloy0076> "sling coffee" [08:30:46] <lloy0076> You're a Barista. [08:31:14] <lloy0076> Or are you a sommelier? [08:31:21] <gdamore> e^ipi: you need to have some industry experience. you can contract, but you have to get it somehow. [08:31:28] *** estibi has quit IRC [08:31:37] <bda> Sommelizin' some phat rhymes. [08:32:04] <gdamore> btw, Sun is going to start hiring college interns soon.... pay attention if you want an internship. :-) [08:32:28] <lloy0076> heh [08:32:29] <e^ipi> the campus ambassador thing, or ? [08:32:42] <lloy0076> I got into the IT game because I said to a well-known Samba developer that I'd do work for nothing for him. [08:33:07] <lloy0076> He took me up on it and then when he saw I could actually do stuff, he started paying me too. [08:37:30] <bda> I took a buddy's jr admin job at the local HSD when he went off ot play in the bubble. [08:38:02] *** LuckyLuk1 has joined #opensolaris [08:42:15] <Tempt> lloy0076: A sommelier deals in wine, not *$. [08:42:46] <e^ipi> you can have a coffee sommelier [08:42:57] <e^ipi> but not nearly as common as a wine sommelier [08:43:04] <gdamore> real engineering intern. [08:43:09] <Tempt> Bloody neologisms [08:44:24] <e^ipi> gdamore: presumably this would require relocation to california for the summer? [08:44:35] <e^ipi> not that I'm opposed to that actually, it would be fun [08:44:48] <e^ipi> but being a canadian citizen it may cause some headaches, INS wise [08:45:32] <Tempt> Got a peecee here that one of the co-workers installed Solaris x86 10, 11/06 on. [08:45:33] <gdamore> oh, it could. [08:45:50] <Tempt> I think he screwed the graphics config because it's booting into a charming screen of corruption. [08:45:53] <gdamore> i think they are US only positions.... but I"m not sure [08:45:57] <Tempt> How did I boot it single-user? [08:46:04] <Tempt> do I [08:46:17] <Tempt> (hint: Just Googling didn't work for me here) [08:47:06] <_mary_kate_> Tempt: press 'e' at the grub screen, scroll to the 'kernel' line, press 'e' again, go to the end, append '-s', press enter, press 'b' [08:47:36] <_mary_kate_> (all hail x86 newboot, because typing 'b -s' was just too easy) [08:47:39] <Tempt> Still tries to start X [08:47:50] <Tempt> That's the answer googling gave me. [08:49:09] <_mary_kate_> hm [08:49:14] <_mary_kate_> oh! not the kernel line. the 'module' line. [08:50:06] <_mary_kate_> no, that's wrong. i t should be kernel. this works for me: kernel /boot/multiboot -s -B console=ttya [08:52:13] <Tempt> Fixed it anyway. [08:52:18] <Tempt> booted failsafe [08:52:20] <Tempt> cd /usr [08:52:22] <Tempt> mkdir splat [08:52:25] <Tempt> mv openwin dt splat [08:52:28] <Tempt> hah. [08:52:30] <_mary_kate_> heh [08:53:15] <Tempt> Man, this used to be easier ... [08:53:28] <seanmcg> Tempt: if you booted failsafe then /usr is just memory, you men /a/usr right ?-) [08:53:35] <seanmcg> s/men/mean/ [08:53:42] <Tempt> Oh, I chrooted to /a first [08:53:47] <seanmcg> ahh. [08:54:31] <seanmcg> _mary_kate_: multiboot ? That went away ages ago.. its kernel$ /platform/i86pc/kernel/$ISADIR/unix now... [08:54:36] <_mary_kate_> seanmcg: he said S10 [08:54:40] *** LuckyLuke has quit IRC [08:54:56] <seanmcg> I sould really scroll up... apols folks. [08:55:19] <Tempt> This is what happens when the desktop guys get into your media packs. [08:55:55] <_mary_kate_> the failsafe thing on S10 is fun. the patch that added grub to 3/05 forget to add the failsafe entry. [08:56:08] <Tempt> Yes, I remember that. [08:56:09] <_mary_kate_> break your boot archive, spend an hour trying to get a shell :) [08:56:17] <Tempt> It was also the patch that broke the system typically ;) [08:56:41] <Tempt> I think that was the point at which my eternal hatred for grub was cermented. [08:58:10] <e^ipi> since solaris doesn't use multiboot anymore, it's pointless, but I've considered making a multiboot compliant bootloader under a less restrictive license [08:58:50] <_mary_kate_> e^ipi: i thought the new 'unix' was a multiboot binary [08:59:29] <_mary_kate_> please write a multiboot loader that gives back the old 'boot:' prompt :) [09:02:40] *** triplah_ has joined #opensolaris [09:05:34] <e^ipi> actually I was just more concerned about getting it off the GPL and on to something like BSD [09:06:09] *** Symmetria has joined #opensolaris [09:06:31] *** g4lt-lappy has joined #opensolaris [09:06:56] <Tempt> What does it matter? [09:07:02] <Tempt> License-wise? [09:08:19] <e^ipi> *shrug* I imagine there might be some company or another weary of using GRUB because of the GPL [09:08:32] <dlg> yeah [09:09:03] *** jwest- has joined #opensolaris [09:09:32] <e^ipi> *lery [09:10:39] *** CSFrost has quit IRC [09:10:40] <Tempt> Bring back DCA [09:10:55] *** sahafeez has quit IRC [09:12:28] <e^ipi> as a standard, multiboot isn't that bad to work with ( x86 stupidness aside ) [09:12:41] <e^ipi> unfortunate is that the only implementation of it is under a poisonous license [09:13:25] <Tempt> Well, it did come from the Hurd project - you can't expect anything other than GPL [09:14:37] *** phonohead has joined #opensolaris [09:14:55] <phonohead> hey! would i configure NexentaOS as a normal Ubuntu distro? [09:15:16] <e^ipi> no, I don't expect the FSF to produce anything licensed under anything better than the GPL [09:15:37] <e^ipi> just saying, it'd be nice if the standard were implemented under a license that people could actually work with [09:16:24] <kaiwai> "but teh GPL is trooo licence!" [09:16:32] <e^ipi> shut up kaiwai [09:16:39] <e^ipi> phonohead: how do you mean? [09:21:10] *** mikefut has joined #opensolaris [09:21:59] <coffman> hmm [09:22:22] <coffman> how would i get the sound keys on my type 6 keyboard to work in xfce? [09:23:09] <quasi> coffman: solder them to the controls on your speaker [09:24:24] <coffman> yeah nice [09:26:12] *** JWheeler has joined #opensolaris [09:27:12] <Tempt> coffman: read brendans blog [09:27:15] <Tempt> Or... [09:27:20] <Tempt> read the manpage for speckeys [09:28:05] *** cydork has joined #opensolaris [09:28:37] * lloy0076 runs [09:28:40] <lloy0076> Reading man pages? [09:28:45] <Tempt> THE HORROR [09:30:56] *** danv12 has quit IRC [09:31:55] <quasi> that's crazy talk! [09:33:25] *** McMurlock|work has joined #opensolaris [09:33:51] <kaiwai> quasi: crazy talk would be taking advice from me [09:34:28] *** g4lt-U60 has quit IRC [09:34:52] *** g4lt-lappy has quit IRC [09:36:52] * quasi should probably get coffee before offering advice also [09:38:56] <e^ipi> hooray for coffee [09:39:01] <e^ipi> it is lifeblood [09:39:04] <coffman> Tempt: nice,thx [09:39:13] <coffman> e^ipi: thumps up [09:40:10] *** danv12 has joined #opensolaris [09:42:02] *** Kprofthreat is now known as Kpr|hates|B [09:42:09] *** Kpr|hates|B is now known as Kpr|hates|Bnet [09:42:26] <quasi> e^ipi: http://www.tmsfeatures.com/tmsfeatures/servlet/com.featureserv.util.Download?file=20070716csbhl-a-p.jpg&code=csbhl [09:43:37] *** Kpr|hates|Bnet is now known as KocksuckingBnet [09:43:55] <lloy0076> Oh, why did they have to make redhat.com so slow :( [09:44:12] <e^ipi> it's running on linux [09:44:15] <e^ipi> what do you expect? [09:44:24] *** KocksuckingBnet is now known as Kpr [09:44:34] <e^ipi> ( that one was too easy ) [09:44:39] <lloy0076> It's not the Linux; it's the fact they've got all the eye candy they don't need :P [09:44:55] <e^ipi> naw, i'm blaming the linux [09:48:59] *** KermitTheFragger has joined #opensolaris [09:49:52] *** sahafeez has joined #opensolaris [09:50:00] *** sahafeez has quit IRC [09:55:15] *** calumb has joined #opensolaris [09:55:47] <kaiwai> meh, blame me, its alot easier [09:57:30] <e^ipi> that would imply you have relevance and ability enough to be at fault [09:59:07] <Kpr> http://www.tectonic.co.za/view.php?id=1633 [09:59:41] <e^ipi> yeah, indiana really did a lot to dilute the opensolaris name [09:59:46] <e^ipi> it was a total cock-up all around [10:00:07] <quasi> yep [10:00:10] *** dmarker has quit IRC [10:00:20] *** dmarker has joined #opensolaris [10:00:25] <e^ipi> i'm not sure if this was Ian's intent, or if he's just incompetent, but that's the result [10:00:33] <dlg> heh [10:00:50] *** lloy0076 has left #opensolaris [10:01:08] <quasi> e^ipi: my money is on malicious intent [10:01:17] <e^ipi> i wouldn't doubt it [10:01:32] <e^ipi> his first loyalty is to Linux [10:05:13] <e^ipi> if RedHat or Canonical hired paul allen, nobody would trust him... not sure why Murdock gets a free pass [10:06:39] <Tempt> Christ. [10:06:44] <Tempt> Sun should muzzle Murdoch. [10:06:50] <Tempt> Keep him around to be seen and not heard. [10:07:40] <Cyrille> someone should point him to this channel, this would make for interesting conversations. [10:08:02] <Tempt> Actually, he's not doing too badly in that article. [10:08:09] <e^ipi> wouldn't quiet me down, i'm quite open with my criticisms [10:08:20] <Tempt> "There is very little compatibility between Linux distributions" = truth [10:08:31] <e^ipi> hell, i've mentioned them on iandiana-discuss, and I know he reads that list [10:08:51] <Cyrille> I wasn't suggesting that as a way to silence criticism. [10:09:11] <e^ipi> i'm still waiting for him to give me any sort of evidence that he's not trying to kill OpenSolaris or the community [10:10:07] <cmang_> if you'll be so kind as to pardon my ignorance, what has he done to kill opensolaris? [10:10:07] <Tempt> vnm,.vc////////////////xd?<<<<<<<<<<,,,,,,,,,,,,,,,,,,,,,,,,,,ffffffffffffff cc cvzdcvcv bbvvvvvvvvvvvvnnbmncnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnvvvvvvvvvnnnnnv , n, m [10:10:12] <Tempt> Oops. [10:10:16] <Tempt> (keyboard repairs) [10:10:42] <Tempt> His point about the best of Linux being distribution model, community and free/open source stands true. [10:10:43] <e^ipi> cmang_: trying to make us in to linux, diluting the name by saying we're trying to be more like linux, etc [10:10:55] <Tempt> He doesn't try to pull the "Linux code is better!" line like Torvalds. [10:12:29] <e^ipi> no, he'd get fired [10:13:10] <kaiwai> hmmm, nothing wrong with murdoch [10:13:35] <Tempt> Although I don't think the Linux community is really worth much anymore. [10:14:00] <e^ipi> so why is he trying to push solaris users out of the community & replace us with linux users ? [10:14:17] <kaiwai> who said he was trying to do that? [10:14:23] <Tempt> Because he wants his old Debian friends back? [10:14:36] <e^ipi> kaiwai: his actions [10:14:41] <Tempt> The crowd that slags him off all over the place because he left Debian? [10:14:42] <kaiwai> because he wants to get people to move over, and allow that transition to be as smooth as possible [10:15:04] <kaiwai> but debian can keep the GPL zealots [10:15:18] <e^ipi> there's a difference between "smooth transition" and "making solaris in to linux" [10:15:39] <Tempt> The problem I see here is a "smooth transition" seems to involve encouraging the same hack/slash attitude that prevails in the Linux community. [10:15:39] <kaiwai> depends on how you define 'makeing solaris into linux' [10:15:40] <Cyrille> I'm not sure what you mean with "in to linux" [10:16:33] <Cyrille> but he's definitely walking a fine line when trying going by the "best of both worlds" line. [10:17:05] <kaiwai> well, if 'like linux' means apt-get with a big ass load of software to choose from, I'd say its a small price to pay [10:17:33] <Cyrille> cool, can we also have the endless "incompatible version of glibc" fest with that too? [10:17:59] <kaiwai> it would be nice if people stopped using "linux userland" (reading osnews.com) [10:18:16] <kaiwai> Cyrille: yeah, and I'd like a boyfriend - its not going to happen anytime soon [10:18:33] <kaiwai> Cyrille: its the same bullshit used for not having a stable driver API in Linux [10:18:58] <Cyrille> are you by any chance implying other APIs are stable? [10:19:19] *** calumb has quit IRC [10:19:43] * coffman hands kaiwai a ubuntu cd [10:20:01] * kaiwai gives coffman a big hug and kiss to scare him off [10:20:28] <kaiwai> Cyrille: atleast if you write a driver for Solaris it isn't a gamble as to whether it works or not with the next release [10:21:01] <Tempt> Oh, let's move to glibc. [10:21:04] <Tempt> That'd be a big step forward. [10:21:19] <e^ipi> hell, let's just move everything over to the GNU userland [10:21:30] <e^ipi> and while we're at it, why not slap down the Linux kernel [10:21:35] <Tempt> What a good idea? [10:21:35] <cmang_> burn, unix, burn [10:21:38] <tsoome> Tempt: its not forward, if you have turned around before stepping [10:21:39] <e^ipi> we can call it "OpenSolaris" [10:21:47] <Tempt> Yes. [10:21:55] <Tempt> Solaris can become another Linux distro to join all the others! [10:22:00] <kaiwai> damn, if there was anything to make someone emo, that would be it [10:22:02] <e^ipi> nobody else gets to use the name, because that'll be the official "OpenSolaris" [10:22:15] <Tempt> And that way, we can ditch all that nasty expensive SPARC hardware. [10:22:41] * kaiwai eyes up an old Amiga 1200 [10:22:47] *** kumamoto has joined #opensolaris [10:22:52] <kaiwai> the good old days [10:23:07] <Tempt> And when vendors list a specific version as a supported platform, it'll actually mean something, because it won't run on anything newer due to glibc problems. [10:24:02] <Cyrille> actually it can even be that it won't work on "updates" from that specific version... [10:24:09] <Tempt> Then then everyone on slashdot can say "woo-hoo, we got rid of Solaris, we're obviously closer to beating Microsoft." [10:25:20] <kaiwai> hmm, mediocrity beating mediocrity [10:26:11] <Tempt> Because while the linux community mantra seems to be about beating Microsoft, they only gains they've had is displacing UNIX. [10:27:54] <cmang_> well, liux is also used in embedded applications that you don't see UNIX in. [10:28:13] <cmang_> linux, that is [10:28:52] *** nemesis_ has joined #opensolaris [10:29:22] <tsoome> and what good has come from that? [10:29:29] <Tempt> It cost QNX marketshare. [10:29:36] <Tempt> Thus beating Microsoft, apparently. [10:29:47] <kaiwai> Tempt: true, but its questionable about how much it has truely affected big vendors like Sun [10:29:53] <tsoome> I just throwed away stupid linksys wifi box [10:30:00] <Tempt> Haha. [10:30:03] <kaiwai> Tempt: Sun's margins are down, but their volumes are up [10:30:20] <Tempt> I struggled with a LinkSys WAG54G for months before tossing it in favour of a router running a real network OS. [10:30:34] <tsoome> Tempt: exactly [10:30:46] <kaiwai> hmm, haven't had any problems with my Belkin 54g router yet [10:30:46] <tsoome> same crap [10:30:49] <kaiwai> works nicely [10:30:52] <kumamoto> tsoome: try using pfsense [10:30:54] <Tempt> Pfft. [10:31:10] <Tempt> Belkin ... I've got a Belkin AP, it needs a reboot weekly. [10:31:30] <tsoome> ok, original fw crashed, tryed those "free" ones, same shit [10:31:55] <kumamoto> am all about recycling old hardware for routers and firewalls including wifi [10:32:00] <tsoome> the box just is unable to stay running [10:32:03] <Tempt> Man, there's an operating system designed to run on routers. [10:32:08] <Tempt> Carefully engineered for that purpose. [10:32:21] <e^ipi> Tempt: IOS ? [10:32:29] <Tempt> e^ipi: Correct! [10:32:52] <_mary_kate_> yes. no-one ever has strange problems with ios. ;-) [10:33:08] <Tempt> Although the stuff that Juniper has done impresses me as well, and there's a heart of BSD in there. [10:33:19] <Tempt> _mary_kate_: Yes, but they tend to be more fixable. [10:33:30] * kaiwai listening to miles davis [10:34:05] * dlg hug his openbsd router [10:35:31] * Tempt can't hug his router, it's in a rack. [10:35:41] <cmang_> About the earlier topic, it kind of sounds to me like having a separate modular gnu'd up distro with Solaris underpinnings makes sense for some people, as long as it doesn't displace the quality or job of the "official" solaris.. [10:35:54] <kaiwai> anyone given the VIA 64bit processor a try with Solaris? [10:36:06] <dlg> via has a 64bit chip? [10:36:09] <kaiwai> cmang_: defined 'gnu'ed' [10:36:16] <dlg> kaiwai: url pls [10:36:19] <cmang_> kaiwai: apt-get, gnu userland [10:36:23] <e^ipi> cmang_: "some people" meaning people too stubborn to learn a better way of doing something? [10:36:29] <Cyrille> isn't that what nexenta is? [10:36:46] *** McMurlock|work has quit IRC [10:36:48] <cmang_> Cyrille: that was my impression. [10:37:22] *** McMurlock|work has joined #opensolaris [10:38:43] <Cyrille> I think that's what's confusing people, if the goal is just to do Nexenta, well good news, someone's already done that. If it's not, then what is it? [10:38:46] <kaiwai> dlg: thought there was, but can't find it, must have never gone ahead [10:39:20] <dlg> kay [10:39:25] <kaiwai> Cyrille: it'll be able making Solaris easier to use most likely [10:39:42] <kaiwai> but unfortunately there is always some dick who equates that to GNU and Linux [10:40:29] <e^ipi> no, not easier to use [10:40:34] <e^ipi> just more like gnu/linux [10:40:43] <e^ipi> don't compound the two [10:41:19] <e^ipi> i find the GNU userland to be totally unintuitive. I'm used to the sysv userland [10:41:34] <Cyrille> to be fair, he seems to restrict the "more like linux" aspect to the distribution model, not that I really know what he means there, and is not really talking about the actual contents. [10:41:34] <WickedWicky> morning people [10:42:15] <kaiwai> Cyrille: maybe the issue at play is 'poor communication' on Ian's part [10:42:22] <kaiwai> 'allo WickedWicky [10:42:26] <Tempt> Getting off the topic of hating political posturing and onto tech issues, has anyone here deployed Confluence? I've got a project wanting to ditch their mediawiki and replace it. [10:42:31] *** calumb has joined #opensolaris [10:42:48] <e^ipi> or perhaps the issue is " Ian is a Linux zealot, and why the fuck are we dumb enough to trust him with anything?! " [10:43:39] <kaiwai> I'm sensing a negative vibe from you e^ipi; don't hold back your feelings [10:44:18] *** yongsun has left #opensolaris [10:44:27] *** nemesis has quit IRC [10:44:53] <WickedWicky> Tempt: not me [10:45:16] *** timsf has joined #opensolaris [10:45:30] <Tempt> Alright. [10:45:37] <timsf> hi all [10:45:40] <Tempt> Quoting Ian here .. "OpenSolaris is a kernel". [10:46:08] <e^ipi> ian is an idiot. he made a blog post to that effect ( "is it a kernel? is it a full OS? oh, so confusing!" ) [10:46:35] <e^ipi> christ... FreeBSD is more than a kernel and less than a full OS too... why is it that hard to grok? [10:47:42] <kaiwai> Tempt: I saw that, "its a kernel....oh its more than that!" [10:48:23] <kaiwai> with that being said, there are bigger idiots at sun besides Ian [10:48:43] <Tempt> Holy hell. Google ads on ianmurdock.com - "Minix 3: a new reliable operating system. Small than Linux. Try it. It's free!" [10:48:48] <Tempt> Minux? Buying ads? [10:48:53] <Tempt> s/nux/nix/ [10:49:15] <kaiwai> from what I understand it is buggy and unstable as buggery [10:49:34] <Tempt> How much do you understand about Minix? [10:50:32] <cmihai> Eh, needs memory managemnet. [10:50:33] <dlg> schwartz is my new hero [10:50:36] <dlg> and mark taylor [10:50:37] <cmihai> And virtual memory. [10:50:39] <e^ipi> dlg: ? [10:50:48] <dlg> everyone else that high up seems confused [10:50:51] <kaiwai> Tempt: unfortunately what I know is from negative feedback :) [10:50:57] <cmihai> dlg: or an asshole :P [10:51:10] <kaiwai> cmihai: or useless [10:51:10] <Tempt> kaiwai: So you know nothing about Minix. [10:51:11] <dlg> or an asshole, yeah [10:51:27] <cmihai> But yeah, Schwartz is quite the diplomat. Loved his answer to Linus [10:51:28] <kaiwai> the marketing manager being number one on the list [10:51:54] <kaiwai> the idea a mouth breather of that quality sharing the same air as I gives me shudders up the spine [10:52:15] <e^ipi> schwartz seems like a cool guy [10:52:33] <cmihai> One of the few CEOs with a pony tail anyway :P [10:52:46] <dlg> mark taylor arranged neptune doco for me without requiring an nda [10:52:55] <kaiwai> schwartz does need to learn spin though [10:53:12] <kaiwai> neptune? [10:53:29] <dlg> suns 10Gb chip [10:53:38] <Tempt> Doing some coding? [10:53:45] * dlg always coding [10:53:47] <Tempt> Is Neptune supported through NEMO? [10:54:04] <dlg> dunno :) [10:54:13] <kaiwai> should do, aren't all of them being written for the new API? [10:54:22] <kaiwai> or is Sun doing a Microsoft? [10:54:27] <Tempt> dlg: Watcha ya cooking up with Neptune? [10:54:32] <kaiwai> "here is a new API, but we won't use it" [10:55:38] <e^ipi> kaiwai: how would you know, you can't even code? [10:55:43] <dlg> Tempt: openbsd drivers [10:55:52] <Tempt> dlg: Good work! [10:56:02] <kaiwai> e^ipi: who said I can't? [10:56:09] <e^ipi> kaiwai: you did [10:56:10] <dlg> Tempt: you can say that after it works [10:56:15] <Tempt> kaiwai: This your opportunity to hand over some code. [10:56:25] <Tempt> dlg: Well, good effort then! ;) [10:56:31] <kaiwai> e^ipi: no, I said, I can't programme in C or C++, but it doesn't mean I can't code or can't understand concepts [10:56:32] <dlg> heh [10:56:47] <ofu> lol http://www.newlaunches.com/archives/what_happens_when_you_turn_the_ac_off_in_the_server_room.php [10:57:02] * dlg would like it if you mailed mark taylor and said nice things too [10:58:32] <Tempt> dlg: /msg me the details and I'll send the mail. [11:00:53] *** vmlemon has joined #opensolaris [11:03:05] <Tempt> The joy of BSD licensing means that'll turn up in FreeBSD soon. [11:03:23] *** kaiwai has quit IRC [11:03:30] <dlg> ? [11:03:51] <Tempt> Well, as soon as you've got your driver done, the porting process can get it over to FreeBSD. [11:04:04] <dlg> ah [11:04:06] <Tempt> I wouldn't mind seeing a bit more BSD around, to be honest. [11:04:14] <Tempt> If I can't have Solaris, I want BSD. [11:04:19] *** LuckyLuke has joined #opensolaris [11:04:22] *** yongsun has joined #opensolaris [11:04:43] <dlg> maybe we should make a project to make solaris feel more like bsd [11:04:48] <dlg> project california [11:04:57] <e^ipi> i'm liking my admittedly limited experience with AIX so far [11:05:10] <Tempt> export PATH=/usr/ucb:/bin ... [11:05:11] <e^ipi> let's go with that [11:05:14] <ofu> california sounds more interesting than indiana... [11:05:17] * dlg not sure the other bsds like his stuff [11:05:18] <e^ipi> project "keep doing what you're doing" [11:05:18] <cmihai> heh [11:05:21] <Tempt> The BSD style userland is still there. [11:05:22] *** tsoome has quit IRC [11:05:26] <cmihai> That's project "let's go back to SunOS" :-) [11:06:05] <e^ipi> not a bad project [11:06:10] <cmihai> eh\ [11:06:25] <cmihai> No real point. Oh well, I guess they could update the ucb tools [11:06:39] <cmihai> But apart from that, nut much to be done. I don't want BSD init, I like SMF. [11:06:52] <dlg> bsd init means different things in different places [11:06:53] <_mary_kate_> solaris should use initng!! [11:07:07] <e^ipi> _mary_kate_: suggest that to indiana-discuss [11:07:11] <e^ipi> they'd probably go for it [11:07:14] <cmihai> And pkgsrc can get you a nice toolchain and apps. [11:07:15] <Cyrille> or a completely crazy idea, we could have project "OpenSolaris" [11:07:56] <Tempt> Couldn't hurt to give the ucb tools a coat of paint. [11:08:06] <cmihai> Yeah [11:08:08] <cmihai> Or port PF :-) [11:08:17] <cmihai> And CARP and all that. [11:08:17] <Tempt> ipfilter is good enough ;) [11:08:18] <e^ipi> i thought someone was working on that [11:08:26] *** boro has joined #opensolaris [11:08:26] <e^ipi> one of the SoC students [11:08:27] <Tempt> Unless you want all that fancy layer 2 filtering stuff. [11:08:41] <cmihai> What fancy layer 2 filtering? That's all bridge stuff :-) [11:08:53] <cmihai> FreeBSD's NetGraph stack is nice too. [11:08:56] <_mary_kate_> i hear U4 will finally fix ipf return-rst on x86 [11:09:16] <dlg> cmihai: LIAR [11:09:32] <cmihai> It's not nice? [11:09:37] <dlg> no, not really [11:09:50] <cmihai> :-( [11:09:53] <cmihai> But it's modular :-\ [11:10:44] <dlg> heh [11:10:45] <dlg> so are streams [11:11:47] *** Snake007uk has joined #opensolaris [11:13:31] *** baijiutong has quit IRC [11:15:20] *** calumb has quit IRC [11:18:30] *** Fish has joined #opensolaris [11:19:34] <Fish> hello [11:19:55] *** nemesis has joined #opensolaris [11:20:30] *** LuckyLuk1 has quit IRC [11:24:23] *** trip_ has joined #opensolaris [11:24:32] *** phalenor has quit IRC [11:33:34] *** nemesis_ has quit IRC [11:35:35] *** Murmuria has joined #opensolaris [11:41:56] *** cmang_ has quit IRC [11:42:03] *** g4lt-U60 has joined #opensolaris [11:42:10] *** phalenor has joined #opensolaris [11:42:55] *** g4lt-lappy has joined #opensolaris [11:45:34] *** cmang has joined #opensolaris [11:47:35] *** Drone has quit IRC [11:47:52] *** Drone has joined #opensolaris [11:48:27] <Tempt> Jesus. Read some of the comments at http://ianmurdock.com/2007/06/08/where-do-i-download-opensolaris/#comments [11:49:03] <Tempt> "Solaris is too hard to install for the neophyte, the first thing that gets them is the slices, where are the partitions and what are slices. Big hang up for Windows trained Linux types." [11:49:16] <Tempt> Man, when did partitions/slices get so hard? [11:50:20] *** nemesis has quit IRC [11:50:20] *** Teknomancer has joined #opensolaris [11:50:33] <Teknomancer> hi all, i'm just going to install opensolaris (first time) [11:51:05] <Teknomancer> i'm porting an app for it that requires qt3, libsdl, IASL, libxerces, and many more libs... [11:51:26] <Tempt> What's the app? [11:51:32] <Teknomancer> does open solaris has something like a package manager where i can pull those libraries? [11:52:48] <Teknomancer> Tempt: i'm don't think i'm allowed to reveal that info :( (you know companies, really sorry as i dont want to be sued even before i start :( ) [11:53:28] <Tempt> Fair enough. [11:54:10] <Teknomancer> hm, i have only 10.95 gigs to spare ... site says 10 GB is a reasonable minimum :) so i hope that's really true :P [11:58:33] *** e57181 is now known as estibi [11:59:55] *** bnitz has joined #opensolaris [12:02:42] <tomww> Teknomancer: you could look at the build environment used to build the GNOME stuff on Solaris. [12:03:33] *** danv12 has quit IRC [12:03:39] <tomww> there is a package system, but it is not with a comfortable User-interface [12:03:44] *** cmang_ has joined #opensolaris [12:03:45] <Teknomancer> oh [12:04:03] <Teknomancer> worst come worst, i guess i'll then google for each library [12:04:10] <tomww> it is being worked on, so this should not a big problem with your tasks [12:04:16] <Teknomancer> hopefully most of them should be available [12:04:38] <tomww> I think you want a place, where you can get the missing libs to compile (if necessary) and install [12:04:46] <Teknomancer> tomww: yeah [12:05:09] <tomww> I myself use the common build environment (the GNOME people use). [12:05:36] <tomww> THere is a sourceforge repository with build-recipes. with the compile around you type command like this: [12:06:28] <tomww> pkgtool --download --interactive build SFElivemedia.spec [12:06:35] <Teknomancer> we have our home-grown build environment, the guy who maintains it said, it already works in opensolaris, so hopefully, most of the more icky stuff like specialized compilers should already be available to me [12:06:56] <Teknomancer> tomww: i see.. will check [12:06:59] <tomww> this fetches the sources, decompresses, patches, configures, compiles, installs with a RPM-style recipe [12:07:16] <tomww> see this link for an example (including compiler install) [12:07:23] <tomww> http://sarine.nl/gmpc-solaris-build [12:08:01] <Teknomancer> k checking [12:09:20] *** CSFrost has joined #opensolaris [12:09:27] *** jwest- has quit IRC [12:09:33] <tomww> some days before ther was a problem with the sourcerge ip address for svn, you can make your own entry in /etc/hosts as a workaround [12:09:44] <tomww> *sourceforge [12:10:25] *** halton has left #opensolaris [12:14:56] *** sbahra has quit IRC [12:18:46] *** cmang has quit IRC [12:22:32] *** tsoome has joined #opensolaris [12:23:36] *** phalenor_ has joined #opensolaris [12:23:37] *** MattMan has joined #opensolaris [12:23:59] *** NikolaVeber has joined #opensolaris [12:31:54] *** sbahra has joined #opensolaris [12:32:07] *** Kpr is now known as Kprofthreat [12:32:35] *** tombhadAC has joined #opensolaris [12:32:40] *** Teknomancer is now known as _Teknomancer [12:35:10] *** phalenor has quit IRC [12:38:24] *** calumb has joined #opensolaris [12:44:07] *** boro has quit IRC [12:48:23] *** MattMan has quit IRC [12:49:15] *** coffman has quit IRC [12:56:40] *** Samy_ has quit IRC [12:57:02] *** Samy_ has joined #opensolaris [12:57:13] *** calumb has quit IRC [13:01:05] *** deather_ has joined #opensolaris [13:04:43] *** nemesis has joined #opensolaris [13:04:57] *** trip_ is now known as triplah [13:08:33] *** noyb has joined #opensolaris [13:13:37] *** yongsun has quit IRC [13:17:57] *** deather has quit IRC [13:18:50] *** cmang_ has quit IRC [13:23:41] *** dme has quit IRC [13:33:29] *** derchris has quit IRC [13:33:42] *** derchris has joined #opensolaris [13:33:58] <_mary_kate_> hmm, the iscsi initiator seems to use a lot of system cpu [13:37:15] *** ShanghaiScott has joined #opensolaris [13:37:57] *** deather_ is now known as deather [13:38:18] *** Fullmoon has joined #opensolaris [13:38:33] *** iMax has quit IRC [13:40:06] *** dme has joined #opensolaris [13:40:46] *** cmang has joined #opensolaris [13:42:58] *** CIT has joined #opensolaris [13:44:30] *** ShanghaiScott has quit IRC [13:44:51] *** iMax has joined #opensolaris [13:44:54] <CIT> Hey [13:45:01] *** CIT is now known as edwardocallaghna [13:45:05] *** edwardocallaghna is now known as edwardocallaghan [13:45:11] <quasi> hey is for herses [13:45:43] <edwardocallaghan> Hey quasi [13:46:23] <quasi> howdy edwardocallaghan [13:46:33] *** halton has joined #opensolaris [13:48:46] *** Symmetria has quit IRC [13:51:10] *** jambock has joined #opensolaris [13:51:38] <edwardocallaghan> dclarke:Awake yet ? [13:52:46] *** cmang has quit IRC [13:53:12] *** cmang has joined #opensolaris [13:55:28] *** kaiwai has joined #opensolaris [13:55:56] <pjd-> Anyone is familiar with ZFS testsuite? [13:56:20] <pjd-> What is STF? How hard will it be to move this testsuite to FreeBSD? [13:58:50] <quasi> pjd-: http://opensolaris.org/os/community/testing/testsuites/stf/ - as for porting it I don't know [14:00:10] <quasi> and http://src.opensolaris.org/source/xref/test/ontest-stc2/src/tools/stf/ for the source [14:00:18] *** Pietro_S has joined #opensolaris [14:01:08] <pjd-> Ok, thanks. [14:03:13] *** sparc-kly has quit IRC [14:03:41] *** sparc-kly has joined #opensolaris [14:03:42] *** ChanServ sets mode: +o sparc-kly [14:07:16] *** master_o1_master has joined #opensolaris [14:09:31] *** jlc has quit IRC [14:09:54] *** coffman has joined #opensolaris [14:12:32] *** sparc-kly_ has joined #opensolaris [14:13:33] *** pjlv has quit IRC [14:13:41] *** sparc-kly has quit IRC [14:17:32] *** master_of_master has quit IRC [14:19:08] *** hohum has joined #OpenSolaris [14:19:31] <coffman> hmpf [14:20:42] *** RobNyc-work has joined #opensolaris [14:21:13] *** jpdrawneek has joined #opensolaris [14:21:31] *** cmihai has quit IRC [14:22:14] *** obsethryl has joined #opensolaris [14:23:39] *** estibi has left #opensolaris [14:23:40] *** estibi has joined #opensolaris [14:25:10] *** triplah has quit IRC [14:25:38] <timsf> pjd, I'm familiar with the ZFS test suite [14:25:42] <timsf> (I wrote bits of it) [14:25:53] <timsf> - probably quite hard to port to BSD [14:26:03] <timsf> though there's a commands.cfg file that contains all the commands we run [14:26:26] <timsf> you might be able to write a set of Solaris compat wrappers in there to suit BSD ? [14:28:12] <pjd-> timsf: I don't really feel yet how it works, I just wanted to try it on my Solaris box, but only have one spare disk in there... [14:28:31] <timsf> bzzt. Sorry, needs at least two disks [14:28:54] <timsf> should work with iscsi targets though if you can get your Solaris box to see those ? [14:28:59] <pjd-> In FreeBSD we use perl's prove for regression tests, which is very flexible, but not really powerful. [14:29:23] <timsf> Oh yeah, forgot about having to port SUNWstc-stf as well :-/ [14:29:43] <timsf> (don't know much about it's internals) [14:29:58] <pjd-> Bascially this testsuite is mostly based on shell scripts or C programs? [14:30:04] <timsf> Mostly shell scripts [14:30:30] <timsf> There's a small bit of C, which was around before I started in 2005ish, [14:30:46] <timsf> but it's mostly shell, which actually does a fairly good job on zfs test coverage.. [14:31:17] <pjd-> Maybe I should simply change ZPOOL to 'ssh freebsd-box zpool', etc. ?:) [14:31:26] <timsf> Heh [14:32:28] <timsf> porting usr/src/cmd/* and usr/src/lib/* might do the trick too! [14:33:00] *** curlyman has joined #opensolaris [14:33:01] <pjd-> I already went this route:) [14:33:35] <pjd-> Can you advice where should I start? From porting stc-stf? [14:33:49] * timsf thinks [14:34:17] <timsf> Okay - you could probably run the test suite as individual shell scripts, without the STF harness [14:34:51] <timsf> I've got a tiny bit of ksh that implements many of the functions contained in stf that our test suite needs, hang on [14:35:39] <timsf> (hrm, it's gone, I'll dig about) [14:35:59] <timsf> basically it wraps the log_fail, log_pass, etc. functions for you [14:36:20] <timsf> then I think going through the commands.txt file to see which commands we're using would be useful [14:36:46] <timsf> Brendan Greg's execsnoop.d should show you exactly what commadns are being run during a test execution run [14:37:03] *** triplah_ has quit IRC [14:37:05] <timsf> then it becomes a case of working out where BSD and Solaris differ.. [14:43:02] <pjd-> Ok... [14:43:17] <timsf> Sorry, not much help I know [14:43:20] <pjd-> I'll try from running some easy test, like functional/atime. [14:44:26] <timsf> remember to source .cfg files at the top level of the suite [14:44:36] <timsf> and set STF_SUITE to be that top level directory [14:44:47] <pjd-> Ok. [14:45:22] <pjd-> 'typeset' is ksh's built-in command? [14:45:29] <timsf> Yep [14:45:48] *** mikefut has quit IRC [14:46:08] *** Murmuria has quit IRC [14:46:11] *** monzie has quit IRC [14:46:27] *** mikefut has joined #opensolaris [14:46:46] *** curlyman has left #opensolaris [14:47:00] *** xushi_ has quit IRC [14:47:54] <pjd-> timsf: Ok, I'll see what I can hack-up, thanks for your help. [14:48:15] <timsf> No problem - if I find that shell wrapper for the various log_xxx functions, I'll send it on [14:48:39] <timsf> (all they do is just run a command with their arguments, logging pass/fail messages where appropriate) [14:49:45] <pjd-> Cool, I can easly convert them to generate output that is recognized by prove. [14:52:55] *** Gman has joined #opensolaris [14:54:15] <kaiwai> hi gman [14:54:25] <Gman> kaiwai, hey dude [14:54:55] <kaiwai> www.cuteoverload.com <-- for those of us who have cutegasms [14:55:34] *** trs81 has quit IRC [14:55:36] *** trs81 has joined #opensolaris [14:56:48] <kaiwai> so, hows things gman? [14:56:54] <Gman> good [14:57:01] <Gman> attending guadec, the gnome conf this week [14:57:06] <Gman> hanging out with my gnome bros [14:57:22] <kaiwai> cool, same here, got my camera working with ptp; so, whats the vision for GNOME in the future? [14:57:25] <timsf> (as opposed to your actual bros, which was last weekend :-) [14:59:31] <kaiwai> I wouldn't mind seeing mono being merged into JDS/Solaris [15:04:00] <Gman> kaiwai, online desktop 2.0! [15:04:01] <Gman> :) [15:04:29] <Gman> http://online-desktop.org/wiki/Online_Desktop [15:05:16] *** terdmonk has quit IRC [15:05:30] *** inaddy has joined #opensolaris [15:05:57] <kaiwai> its a nice vision, but I doubt it will occur - sorry to sound pesimistic [15:07:46] <Gman> i bet it will [15:08:28] <kaiwai> well, lets remember, this is opensource, its like a group of women - common ground, but generally hate each other over petty little things [15:08:37] *** obsethryl has quit IRC [15:08:47] *** edwardocallaghan has quit IRC [15:09:09] <timsf> Gman, is this something like SunRay over WAN or a VNC desktop + LiveCD sort of thing? [15:09:21] <Gman> read the link [15:09:32] <timsf> I did, can't work out where the processing happens.. [15:09:55] <kaiwai> Gman: sounds similar to KDE's Web 2.0 like technologies [15:10:02] <timsf> Ooh, I get it. [15:10:09] <timsf> Dickless^H^H^H^H^HDiskless clients. [15:10:28] <kaiwai> they were the good old days [15:10:30] <kaiwai> users new their place [15:11:04] <kaiwai> *knew [15:11:26] <kaiwai> and MBA wizzkids were running for coffee instead of running companies into the ground [15:12:14] <kaiwai> gman: the main problem I have is that it'll be too americanised in terms of integration [15:12:31] <Gman> too americanised? [15:13:19] <kaiwai> in terms of lack of integration into services offered outside that of the US [15:14:39] <kaiwai> an example of that problem was with Apple's own widgets and only interacting with for example, US financial data [15:14:51] <Tempt> I do like the idea of an Online Desktop. [15:15:46] <kaiwai> depends how one defines 'online desktop' - cruddy stuff like 'widgets floating in a browser'? [15:15:56] <Gman> timsf, http://developer.mugshot.org/wiki/Big_Board_Design [15:16:17] <Gman> kaiwai, no, that's arse [15:16:23] <Tempt> Well, depends how cruddy the widgets are. [15:16:26] <Gman> it's about using existing web services within your desktop environment [15:16:34] <Gman> connecting, sharing, collaborating [15:16:35] <Tempt> But today's cruddy widget is tomorrow's nice application. [15:16:36] <kaiwai> firefox 2.0.0.5 is out for solaris [15:16:37] <Gman> etc. [15:17:03] <timsf> HAH! [15:17:07] <kaiwai> Tempt: true, but which is going to be used, Silverlight, Flash or JavaFX [15:17:10] <timsf> It's Views all over again :-) [15:17:28] <timsf> (which I loved) [15:18:05] <Tempt> Hey, I'm not getting in that crap. Straight Java seems to be able to deliver the goods, as an SSGD user can tell you. [15:18:52] <timsf> http://acm.org/sigchi/chi97/proceedings/briefing/dg.htm [15:19:19] <timsf> (scroll down for the screenshot, then compare with http://developer.mugshot.org/wiki/Image:Big_Board_Desktop_with_Task_Switcher.png ) [15:19:37] <kaiwai> Tempt: I'm not fussed as to how it is delivered, as long as I can access it with my OS of choice, I'm happy [15:20:06] <timsf> Sounds like an interesting project though, will watch that space. [15:20:32] *** SirFunk has quit IRC [15:21:38] <kaiwai> eh [15:21:39] <kaiwai> its ok [15:21:46] <kaiwai> I prefer the amiga style desktop [15:23:01] <Gman> timsf, i know which i prefer [15:23:06] <kaiwai> as long as it doesn't turn into a dumping ground for advertisers [15:23:24] <timsf> Well add 10 years to any project, and it'll look better, [15:23:31] <timsf> (mostly) [15:23:50] <timsf> but it's the same idea I think... I was a big fan of Views and the ideas they were coming out with [15:24:23] <Gman> timsf, hrm, i don't really grok the similarities [15:24:31] <Gman> one focuses a lot more on presence [15:24:58] <kaiwai> gman: the desktop in the first link looks very similar to Triteal's java desktop back when network computers were the fad [15:25:00] <timsf> Right, but the "no local desktop state", "simple application bar down one side", [15:25:30] <Gman> online desktop is so much more than that though [15:25:39] <timsf> sure, the idea of presence, [15:25:57] <timsf> but that's only appeared since IM really took off ( and everyone was on the LAN, so to speak ) [15:26:01] <Gman> storing local desktop state is just a side effect of having the network [15:26:06] <timsf> (with fast broadband) [15:26:09] *** rbrown has quit IRC [15:27:22] <kaiwai> gman: some of us prefer to keep a certain amount locally stored [15:27:27] *** dunc_ has joined #opensolaris [15:27:48] <timsf> Anyway, read the rest of the ACM paper - I think there's a parallel in there, but I admit to only hearing about the gnome effort 10 minutes ago! [15:28:20] *** Dink has joined #opensolaris [15:28:43] <Tempt> I'd rather have all my state on my server, and have my clients ignorant. [15:29:24] *** halton has quit IRC [15:30:33] *** calumb has joined #opensolaris [15:30:40] <kaiwai> Tempt: oh, for companies, sure, but for me as an individual sitting under a rock [15:30:52] <Tempt> Even for my personal stuff. [15:32:08] <Tempt> I'd rather all my valuable stuff was kept on my server, which I can access from where I need it. [15:34:59] <kaiwai> gonna give the new build a go [15:35:03] * dlg agree [15:36:04] <Tempt> My ideal desktop would be similar to SSGD, but with the option of a dedicated client instead of using a browser. [15:36:21] <Tempt> So you use a browser on the go, but have a dedicated client on hosts you use a lot. [15:36:40] <Cyrille> I think there's a native client or whatever the proper name for it is in SGD. [15:37:15] <Tempt> No, the native client just launches a browser. [15:37:25] <Cyrille> my mistake then. [15:37:35] <Tempt> It just pre-installs all the java applets in case you have a locked-down environment. [15:37:36] *** cmang has quit IRC [15:37:40] *** cmang has joined #opensolaris [15:38:01] <kszwed> there was a c-based native client for tarantella that may still work. [15:38:16] <Tempt> The real barrier to centralised desktop/NC stuff is bandwidth. [15:38:28] <Tempt> kszwed: Not so good. Things have advanced a lot since then. [15:39:15] <kaiwai> Tempt: and reliable cheap connection [15:39:34] <Tempt> Well, it all amounts to the same thing. [15:39:56] <kszwed> Tempt: not really, rdp over aip still blows [15:39:57] *** _Teknomancer is now known as Teknomancer [15:39:58] <Tempt> If I was in Japan, I could pull myself at least 100mbit of bandwidth into home and be able to reasonably access everything at a good speed from whereever [15:40:00] <JWheeler> what command do I use to check which cpu a binary has been optimised for? [15:40:15] <Tempt> file [15:40:18] <Tempt> file $filename [15:40:26] <Tempt> Will return enough header to get the version [15:40:31] <JWheeler> I need to be a bit more specific, ie i386/i586 etc [15:40:36] <Tempt> aah [15:40:43] <Tempt> You'll get a bit more breakdown easily on SPARC. [15:40:54] <Tempt> /opt/PCOWtools/bin/screen: ELF 32-bit MSB executable SPARC32PLUS Version 1, V8+ Required, dynamically linked, not stripped [15:41:04] <JWheeler> I'm building software for multiple arches, and messing about with scripting an $ISAEXEC friendly pkg [15:41:12] <JWheeler> ah yes [15:42:15] <_mary_kate_> file should tell you if it has specific Intel instructions in brackets [15:42:17] <coffman> yeah, if the bandwidth gets better companys could get there selfs sunrays only and buy server time from isp over the internet [15:42:22] <_mary_kate_> i can't find an example now though [15:42:55] <Tempt> Ummn, how much optimization are you doing on your compiles? [15:43:15] <JWheeler> well, I was planning on doing i3/5/686 for x86 [15:43:16] <Tempt> Nobody is going to be running the binaries on anything short of a P2 these days, so ... i686 would be a suitable minimum binary support [15:43:28] <Tempt> I mean, who's running Solaris on i386 anymore? [15:43:33] <_mary_kate_> i doubt you can even install solaris on i486 [15:43:36] <JWheeler> I don't really know much about sparc... I guess I'll learn that when I hit it [15:43:37] <_mary_kate_> i586, possibly [15:43:38] <JWheeler> ...well I am [15:44:00] <JWheeler> _mary_kate_, I'm building mplayer for blastwave, we must support solaris 8, which technically supports i486 :/ [15:44:23] <_mary_kate_> JWheeler: talk to delewis, he maintains solaris port of mplayer [15:44:26] <JWheeler> can't say I really agree with that one myself, but it doesn't take much more effort to add that cpu type to my for loop [15:44:36] <JWheeler> oh cool [15:46:09] <Tempt> How much media are you going to play on a 486? [15:46:22] <JWheeler> :) a valid point! [15:46:49] <Tempt> So, I'd just build it with a modern optimization. [15:47:02] <_mary_kate_> why don't you just use the cpu auto detect stuff [15:47:13] <JWheeler> well that's ok too, I still need to work out what my binaries are being spat out as though [15:47:42] <JWheeler> I'm not sure quite how clever it is. [15:47:43] <_mary_kate_> if it says 80386 and says nothing in brackets, it'll work on a 486 [15:48:18] <JWheeler> To be completely honest, I'm taking over from the previous maintainer, that isn't contactable. He built it with isaexec, so I'm just doing the same [15:48:37] *** sartek has joined #opensolaris [15:49:07] <Tempt> Oh, fair cop. [15:49:22] <Tempt> It'll be good to see some updates to mplayer. [15:49:24] <JWheeler> but hey, I'm always open to suggestions :) [15:49:31] <Tempt> Can you compile a SPARC version with v9+vis? [15:49:40] <JWheeler> absolutely. I use it every day, lets get the most modern builds of it out there! [15:50:04] <_mary_kate_> i'll start using mplayer again when it learns about dvd menus. ;-) [15:50:06] <Tempt> I'm logged in an x86 box at the moment trying to work out which arch the binary is for. [15:50:20] <kszwed> you'll need to start hacking code to leverage vis [15:50:38] <_mary_kate_> i thought it already had vis optimisations [15:50:45] <kszwed> not sure what the future of vis is in the new niagara line [15:51:27] <Tempt> the instructions will stay otherwise it'll break binary compatibility. I can't see Sun wanting to do that. [15:53:10] <kszwed> yes, but the whole point of vis is to do things faster. [15:53:27] <kszwed> just because they exist doesn't mean they're useful anymore. [15:53:55] <kszwed> i mean, look at floating point on the T1 :) [15:54:27] <Tempt> Haha! [15:56:08] <jpdrawneek> Thats is what Rock is for [15:56:36] <Tempt> I want my 16 socket machine with 64 threads on each processor. And I want a golden levitating pony. [15:56:37] <pjd-> # ./atime_001_pos [15:56:38] <pjd-> 15:56:02 ASSERTION: Setting atime=on, the access time for files is updated when read. [15:56:41] <pjd-> 15:56:02 SUCCESS: /usr/bin/touch /mnt/test/testfile [15:56:43] <pjd-> 15:56:02 SUCCESS: /sbin/zfs snapshot testpool/testfs@testsnap [15:56:46] <pjd-> 15:56:02 SUCCESS: /sbin/zfs clone testpool/testfs@testsnap testpool/testclone [15:56:49] <pjd-> 15:56:02 SUCCESS: /sbin/zfs set atime=on testpool/testfs [15:56:52] <pjd-> 15:56:03 SUCCESS: /bin/cat /mnt/test/testfile [15:56:53] <pjd-> 15:56:03 SUCCESS: check_atime_updated /mnt/test/testfile [15:56:56] <pjd-> 15:56:03 SUCCESS: /sbin/zfs set atime=on testpool/testclone [15:56:58] <pjd-> 15:56:04 SUCCESS: /bin/cat /testpool/testclone/testfile [15:57:01] <pjd-> 15:56:04 SUCCESS: check_atime_updated /testpool/testclone/testfile [15:57:04] <pjd-> 15:56:05 SUCCESS: /bin/cat /mnt/test/.zfs/snapshot/testsnap/testfile [15:57:07] <pjd-> 15:56:05 SUCCESS: check_atime_updated /mnt/test/.zfs/snapshot/testsnap/testfile exited 1 [15:57:10] <pjd-> 15:56:05 NOTE: Performing local cleanup via log_onexit (cleanup) [15:57:11] <pjd-> 15:56:05 SUCCESS: /sbin/zfs destroy testpool/testclone [15:57:13] <Tempt> pjd-: STOP FLOODING! [15:57:14] <pjd-> 15:56:05 SUCCESS: /sbin/zfs destroy testpool/testfs@testsnap [15:57:17] <pjd-> 15:56:05 Verify the property atime=on passed. [15:57:20] <pjd-> timsf: :) [15:57:25] <timsf> (phew) [15:57:33] *** calumb has quit IRC [15:57:42] <timsf> Glad that works for you [15:57:44] <coffman> pjd-: use a pastebin for such! is very rude! [15:59:03] <kszwed> pjd-, FAILED: irc etiquette test [15:59:09] *** SirFunk has joined #opensolaris [15:59:27] *** kaiwai has quit IRC [16:00:13] *** estibi is now known as e57181 [16:00:25] <timsf> Next step, get the rest of the test suite working ;-) [16:01:26] *** hohum_ has joined #OpenSolaris [16:02:32] <pjd-> Yeah. [16:02:56] *** hohum has quit IRC [16:03:09] <pjd-> Some of them surely won't work, but I really need such testsuite to verify everything is fine after each integration of changes from OpenSolaris. [16:03:11] <Gman> more gnome web stuff being announced - http://pyrodesktop.org/Main_Page [16:03:17] *** hohum_ is now known as hohum [16:03:44] <timsf> Yep, [16:04:02] <timsf> the OpenSolaris test farm could be of use here, [16:04:09] <timsf> in that if your changes get back into the main zfs tree, [16:04:26] <timsf> you'd be able to run your Solaris BFU archives against the test suite [16:04:41] <timsf> which would prove at least that your changes haven't broken Solaris - [16:04:45] <timsf> (well, it's a start) [16:05:24] <pjd-> timsf: I don't have any major changes to go back to Solaris, but I'm pondering some. [16:05:30] *** crib has quit IRC [16:05:33] *** calumb has joined #opensolaris [16:06:35] *** crib has joined #opensolaris [16:06:48] *** dunc_ has quit IRC [16:07:43] <pjd-> timsf: BTW. I've file system regression testsuite that I created while porting ZFS. It tests correctness of file system operations like rename/mkdir/rmdir/unlink/etc. [16:08:00] <pjd-> timsf: It verifies if returned errors match POSIX, etc. [16:08:20] <pjd-> And now it supports {FreeBSD,Solaris}/{ZFS,UFS}. [16:08:27] <Pietro_S> isn't there any OpenSolaris Media Center pack? [16:08:30] <timsf> Neato. I've a feeling that sort of stuff is already covered in Solaris test (mstc) [16:08:39] <timsf> but I'm not sure that's open sourced :-/ [16:08:54] *** CSFrost has quit IRC [16:09:15] <pjd-> timsf: I'm sure Sun have something like this, because I found no problems on Solaris/ZFS. [16:09:51] <timsf> ZFS is pretty clean from that standpoint at this stage (after all, it's been in production for a while) [16:10:12] <timsf> but I know there's some conformance/compliance tests floating about - never had to run them [16:10:15] <timsf> (thank God) [16:10:57] <pjd-> timsf: Actually some corrner cases are hard to find without targeted testing, eg. returning correct error in some unusual situations. [16:11:16] <timsf> I can well believe it. [16:11:39] <pjd-> But yeah, my tests confirms ZFS is clean. [16:12:24] <timsf> Do you have a pointer to those tests pjd- ? [16:12:32] *** EchoBinary has quit IRC [16:13:04] <pjd-> timsf: http://people.freebsd.org/~pjd/fstest/ [16:13:07] <Pietro_S> any known port of MythTV to solaris? [16:13:09] <timsf> Thanks! [16:13:25] <pjd-> timsf: There should be a README inside. [16:13:36] <timsf> Cool, I'll check it out. [16:16:22] *** reflect_ has joined #opensolaris [16:17:39] <reflect_> does an up-to-date list of hardware chipsets supported by OpenSolaris exist? (the hcl is outdated and extremely hard to find relevant information with) [16:17:51] *** kaiwai has joined #opensolaris [16:17:54] <kaiwai> :) [16:17:59] <kaiwai> purrfect [16:18:01] *** Drone has quit IRC [16:19:56] *** kumamoto has left #opensolaris [16:20:00] *** LuckyLuk1 has joined #opensolaris [16:20:12] <kaiwai> hi LuckyLuk1 [16:20:43] *** Drone has joined #opensolaris [16:22:12] <Teknomancer> time to install opensolaris... [16:22:59] *** Teknomancer has left #opensolaris [16:23:10] *** calumb has quit IRC [16:23:52] *** calumb has joined #opensolaris [16:28:48] <timsf> pjd- cool they work "openssl: not found" on Solaris though ;-) [16:28:53] <timsf> Will sort that out [16:29:22] <timsf> (yeah, /usr/sfw/bin) [16:30:15] <kaiwai> hmm [16:30:39] <kaiwai> LDFLAGS="-L/usr/sfw/lib -R/usr/sfw/lib" [16:31:32] <timsf> (no need, it's just calling the openssl binary) [16:32:07] <timsf> Failed 7/184 test scripts, 96.20% okay. 33/1955 subtests failed, 98.31% okay. [16:32:10] *** szt has joined #opensolaris [16:35:11] *** EchoBinary has joined #opensolaris [16:35:20] *** LuckyLuke has quit IRC [16:35:46] <kaiwai> timsf: whats this for? [16:36:02] <timsf> that test suite that pjd- just pointed me to [16:36:24] <kaiwai> ah [16:36:28] <timsf> dunno what the baseline is though, so those numbers aren't telling the whole story. [16:36:40] <timsf> (could be broken tests, or known bugs in zfs) [16:43:53] <pjd-> timsf: Probably the former, I haven't run them for a while, but from what I remember zfs on solaris passed all tests at some point. [16:44:33] <pjd-> timsf: You edited config, right? [16:44:38] <timsf> Yep [16:45:49] <pjd-> Let me try. [16:50:07] <timsf> It's no big deal really - just curious [16:50:50] <pjd-> I'm curious too:) [16:50:57] <pjd-> But I see the same result. [16:54:57] <axisys> how do I install Xsun.. my peers installed Xorg by mistake on his ultra 20 and now he has no display [16:55:29] <Tempt> Doesn't it install by default? Just run kdmconfig? [16:56:15] <axisys> Tempt: i remmeber the option between the two from my install a while ago.. but he sure is getting tons of xorg related error [16:56:25] <axisys> so kdmconfig will fo the magic? [16:57:05] <axisys> s/fo/do/ [16:57:32] <Tempt> should do [16:57:44] <Tempt> Unless some recent release of SX has deprecated it. [16:58:00] *** pfa3rh has joined #opensolaris [16:58:49] <axisys> Tempt: forgot to mention it is sol10u3 x86 [16:59:25] <Tempt> well, I fixed up a problem with that very release today. [16:59:33] <Tempt> Using kdmconfig to fix a screwed video config [16:59:41] *** calumb has quit IRC [16:59:48] <Tempt> So it should be ideal. [17:00:53] *** Gman has quit IRC [17:01:25] *** nachox has joined #opensolaris [17:04:52] *** calumb has joined #opensolaris [17:07:09] *** Gman has joined #opensolaris [17:07:11] <axisys> Tempt: that did it.. thanks a lot [17:07:16] <Tempt> No worries. [17:09:06] <pjd-> Hmm, interesting. [17:09:24] <nachox> morning [17:09:29] <pjd-> When file permissions are set to 06, noone can read/write to the file. [17:09:55] <pjd-> On Solaris. [17:10:33] <timsf> Even the owner ? [17:10:36] <nachox> 06? [17:11:06] *** LuckyLuke has joined #opensolaris [17:11:28] <timsf> Looks like it.. [17:11:58] <timsf> (same on tmpfs) [17:12:17] <reflect_> does an up-to-date list of hardware chipsets supported by OpenSolaris exist? or is anyone working on creating one? (the hcl is outdated and extremely hard to find relevant information with) [17:12:18] <pjd-> This works on FreeBSD - other users can open the file. [17:12:40] <pjd-> I don't think POSIX defines this behaviour. [17:13:00] <Tempt> http://www.theregister.co.uk/2007/07/18/macbook_apple_support_hell/ [17:13:20] *** jgay has joined #opensolaris [17:14:50] *** nostoi has joined #opensolaris [17:15:21] <timsf> I see that too on Solaris - other users can read the file, just not the file owner. [17:16:59] <pjd-> They can? [17:17:15] <timsf> Yeah, at least on my machine (erm, zfs root) [17:17:16] <pjd-> Not in my test... [17:17:24] <pjd-> On ZFS also. [17:17:46] <timsf> (was just using a shell to cat, or echo >> the file) [17:17:51] *** calumb has quit IRC [17:18:09] <pjd-> timsf: You may try using fstest: [17:18:12] <pjd-> mkdir foo [17:18:21] <pjd-> chown 65534:65534 foo [17:18:22] <pjd-> cd foo [17:18:39] <pjd-> fstest -u 65534 -g 65534 create bar 0006 [17:18:53] <pjd-> fstest -u 65533 -g 65533 open bar O_RDONLY [17:19:29] <pjd-> I've even more scary case. [17:20:07] <pjd-> File owned by 65534:65534: [17:20:07] *** estibi has joined #opensolaris [17:20:13] <quasi> Tempt: sounds like the perfect time to buy a macbook ;) [17:20:18] <pjd-> fstest -u 65534 -g 65534 chmod bar 0772 [17:20:38] <pjd-> fstest -u 65533 -g 65533 open ${n1} O_RDONLY [17:20:44] <pjd-> s/${n1}/bar/ [17:20:49] <pjd-> It succeeds. [17:21:02] <timsf> Odd. [17:21:41] <pjd-> Indeed. [17:23:12] <pjd-> Ops, sorry. [17:23:31] <pjd-> fstest -u 65534 -g 65534 chmod ${n1} 0774 [17:23:39] <pjd-> fstest -u 65533 -g 65533 open ${n1} O_WRONLY [17:23:42] <pjd-> This one succeeds. [17:24:05] <timsf> Could be worth taking it up on zfs-discuss if there's something amiss (or file a bug :-) [17:24:20] <timsf> There could be weird reasons why ZFS is doing things the way it is... [17:24:29] <timsf> (cf. recent bug about atq [17:24:33] * timsf digs for url [17:25:16] <timsf> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6563206 [17:27:45] *** LuckyLuk1 has quit IRC [17:28:51] <pjd-> Hmm. [17:29:05] <pjd-> directory is owned by 65534:65534 [17:29:18] *** tsoome has quit IRC [17:29:21] <pjd-> I create file in there as 65534:65534, but it creates file with 65534:0 [17:29:44] <pjd-> I'm using setgroups(2), so it should remove 0 from groups. [17:31:15] <timsf> I'm not familiar enough with syscalls to really be able to offer advice, sorry pjd- [17:31:52] <pjd-> n/p just thinking loudly. [17:32:26] <pjd-> Maybe I need to change gid separately. [17:32:57] *** JWheeler has quit IRC [17:34:15] *** mega has joined #opensolaris [17:38:30] *** Plouj has left #opensolaris [17:40:14] *** phalenor_ is now known as phalenor [17:41:39] *** stevel has joined #opensolaris [17:41:39] *** ChanServ sets mode: +o stevel [17:41:46] *** mihaic has joined #OpenSolaris [17:44:29] *** Gman has quit IRC [17:46:03] *** mega has quit IRC [17:46:33] *** tsoome has joined #opensolaris [17:47:33] <pjd-> timsf: Ok, all those problems were because Solaris' setgroup(2) doesn't change effective gid. [17:47:53] <timsf> I've learnt something new - thanks pjd- [17:52:05] *** kaiwai has left #opensolaris [17:55:01] <trygvis> quasi: ping [17:55:33] <trygvis> I'm seing the same issue with the supermicro card (with the marvell chip). got any workarounds or other info? [17:55:52] *** Gman has joined #opensolaris [18:02:23] *** calumb has joined #opensolaris [18:03:59] *** sparc-kly has joined #opensolaris [18:04:00] *** ChanServ sets mode: +o sparc-kly [18:04:03] *** sparc-kly_ has quit IRC [18:07:55] *** sparc-kly_ has joined #opensolaris [18:08:06] *** sparc-kly has quit IRC [18:08:52] *** mihaic is now known as cmihai [18:10:20] *** nostoi has quit IRC [18:15:22] *** sparc-kly_ has quit IRC [18:15:28] *** sparc-kly has joined #opensolaris [18:15:29] *** ChanServ sets mode: +o sparc-kly [18:16:26] *** alanc-away is now known as alanc [18:17:28] <quasi> trygvis: you've got one or two in the machine? [18:18:06] <trygvis> one [18:18:29] *** bondolo has joined #opensolaris [18:19:05] *** sparc-kly_ has joined #opensolaris [18:19:37] <quasi> it was postwait who was complaining about resets from time to time - I haven't heard if they got any further [18:19:54] *** sparc-kly has quit IRC [18:20:35] *** sparc-kly_ has quit IRC [18:20:45] *** sparc-kly_ has joined #opensolaris [18:24:01] *** Cyrille has quit IRC [18:27:29] <Abe_Froman> http://blogs.sun.com/perrin/entry/slog_blog_or_blogging_on [18:27:34] <Abe_Froman> that right there is the shit [18:30:14] *** mikefut has quit IRC [18:31:27] *** slowhog has joined #opensolaris [18:31:41] *** obsethryl has joined #opensolaris [18:46:05] *** sahafeez has joined #opensolaris [18:48:46] *** sommerfeld has joined #opensolaris [18:48:47] *** ChanServ sets mode: +o sommerfeld [18:49:11] *** BadKarma has joined #OpenSolaris [18:54:57] *** calumb has quit IRC [18:57:01] *** jgilje has quit IRC [18:58:55] *** reflect has joined #opensolaris [19:01:07] *** reflect_ has quit IRC [19:07:48] * alanc dreams of someday not having to type "See description" in bug evaluations over and over again [19:08:27] *** Snake007uk has quit IRC [19:08:33] <richlowe> alanc: :) [19:09:46] * timeless doesn't understand [19:10:08] <bda> Heh. [19:10:36] * stevel dreams of someday killing his "See comments" nag script [19:10:43] <bda> Co-worker complained that one of his scripts is kinda slower than he expected. Databasen stuff, so he's calling fsync "often". [19:10:45] <timeless> heh [19:10:49] <bda> Like 40/s. [19:11:06] <timeless> why in the world would he do that? [19:11:13] <timeless> is the system that unstable/untrustworthy? [19:11:16] <bda> Making sqlite transactional. [19:11:36] <bda> No, just paranoia. I suggested he backoff and write something to recover the indexes from the master in the event of a failure. [19:13:21] <richlowe> timeless: you're a bugzilla person, you wouldn't. [19:13:31] <richlowe> timeless: hack up a bz4 that hides any useful information, and suddenly, it'll become clear. [19:13:36] <richlowe> or rather, won't. [19:13:39] <timeless> heh [19:14:11] <richlowe> stevel: kill the nag script, then kill the violators. [19:14:14] <richlowe> stevel: problem solved. [19:14:20] <Pietro_S> bda: sqlite3 is transactional ... [19:14:41] <timeless> btw, are there any features you actually miss from the sun bug tracker when you look at bugzilla? [19:14:56] <stevel> richlowe: i'm pretty sure that violates some sort of company HR policy [19:15:30] <wesolows> timeless: I really wish I could do a better job of explaining why I hate bugzilla so much. [19:15:32] <bda> Pietro_S: Yeah, I think he is doing something "crazy". [19:15:35] *** bnitz has left #opensolaris [19:15:36] <wesolows> timeless: But I hate it nonetheless. [19:15:49] <wesolows> Bugster's a pile of crap, but I hate bugzilla much more. [19:16:09] <bda> bugzilla-- [19:16:14] <st3fan> jira++ [19:16:29] <richlowe> from what I can tell, they all suck. [19:16:33] <richlowe> it's just a matter of how badly, and where. [19:16:42] <WickedWicky> evening all [19:16:52] <st3fan> jira is nice :) [19:17:20] <timeless> wesolows: how about a google docs page where you just jot notes? [19:18:18] <alanc> timeless - at least in the bugzillas I've used, it's much harder than bugster to track bugs that affect multiple releases [19:18:31] <timeless> alanc: *nod* [19:18:41] <alanc> perhaps due to the common open source method of "fix it in the newest only, make everyone upgrade to that" [19:18:42] <timeless> how does bugster handle that? (url or 2 please?) [19:18:56] *** obsethryl has quit IRC [19:19:03] <richlowe> timeless: you're unlikely to find urls you can hit that explain bugster/bugtraq in useful amounts of detail [19:19:05] <alanc> instead of the commercial method of "backport to all releases with customers paying for support" [19:19:05] <timeless> i think it's more because it's hard to make a decent ui for that [19:19:17] <richlowe> as far as patches go, it seems to use SubCR's [19:19:20] <timeless> richlowe: ii don't need explanations, just a page that shows it [19:19:21] <alanc> timeless: bugster is completely Sun internal, so the URL's explaining it are too [19:19:29] <bda> I like trac, but it is missing some really necessary features. [19:19:48] <timeless> is it possible to take a page that shows how it's done and replace all the text with random words? [19:19:49] <richlowe> so Introduced In is left accurate, it's fixed in the top-most gate, then SubCR's are filed to drag it back down. [19:19:56] <richlowe> (though that could be entirely wrong.) [19:20:22] <timeless> what happens if a bug is found to exist in an older version? [19:20:29] <timeless> is introduced in changed? [19:20:43] <timeless> and what happens if a bug isn't present in some intermediate version? [19:21:09] <timeless> in bmo we basically use flags to do approximately that [19:21:19] <timeless> version is where it was found, and can be moved to earlier [19:21:22] <timeless> we fix on trunk [19:21:32] <alanc> but as richlowe said, the status for which releases a fix goes in is a table - if it's in S8, S9, S10, then there's three table entries, each showing which build/update it's fixed in for that release (or just "patch" for the patch gates, since adding each patch as a separate build would suck too hard in bugster) [19:21:33] <timeless> and then set flags asking for approval to commit to older versions [19:21:42] <timeless> generally starting w/ the most recent working backwards [19:21:52] <timeless> w/ approval getting harder the older the branch is [19:22:55] <alanc> in fact, all the bugzilla's I've used just track the bug state as FIXED and don't even have a "Fixed in release/build" field at all like bugster does [19:23:15] <timeless> alanc: target milestone generally can be used for that [19:23:21] <timeless> bugzilla product itself does that [19:23:33] <timeless> although most open source projects don't spend time on it [19:23:48] <timeless> in bugzilla product we move target milestone back as things get back ported [19:23:50] *** rbrown has joined #opensolaris [19:24:15] <timeless> and yes, i'm aware it's not great [19:24:32] <timeless> when you have a bug that you're looking to backport a fix on [19:24:33] <alanc> we want to be able to easily search the db for all the bugs fixed in a given build or which build a given bug is fixed in, hence the multiple-release system instead of a single field [19:24:53] <timeless> how often is the patch set significantly different from version to version? [19:25:00] <timeless> and does that affect things? [19:25:09] <timeless> also, do you generally get patches right the first time? [19:25:44] <alanc> timeless: given the amount of change between Solaris 8 and Nevada, there's often a lot of difference, and Sun tries very very hard to get it right the first time [19:25:58] <timeless> err [19:26:10] <timeless> i know that sun tries to get the patch that's shipped to customers right the first time [19:26:13] <alanc> the first time it's integrated that is [19:26:18] <richlowe> when patches aren't right, their wrongness is another bug. [19:26:21] <timeless> but do engineers post interim patches while iterating [19:26:30] <richlowe> introduced in <that patch>, as best as I can tell. [19:26:38] <richlowe> and yes, T-patches and IDRs. [19:26:39] <sommerfeld> timeless: really depends on how much other change happened which overlaps with the patch [19:26:44] <alanc> we don't put the code changes in bugster [19:26:49] <timeless> ah [19:26:55] *** Fullmoon has quit IRC [19:26:57] <alanc> at least not as we're working on them - some groups do at review time [19:27:14] <timeless> http://rt.cpan.org/Public/Dist/Display.html?Name=poe [19:27:21] <timeless> consider that poe example [19:27:30] <timeless> i found its output terrible (see the broken in bit) [19:27:45] <sommerfeld> timeless: IDR's are "interim/diagnostic relief". a specially marked patch explicitly intended to put out a fire. [19:28:00] <sommerfeld> T-patches are early releases of the "real" patch [19:28:14] * timeless nods [19:28:31] <timeless> so basically in modern sun the evolving work would be in Hg [19:28:49] <timeless> where does conversation about work go? office to office? email? [19:28:55] <alanc> also, where bugzilla just has comment fields, bugster breaks them down the into 5 types of "notes" you can add to a bug report - "Description", "Comments", "Evaluation", "Suggested Fix", and "Work Around" [19:29:02] * timeless nods [19:29:09] <timeless> hold on one sec [19:29:24] *** tombhad-AC has joined #opensolaris [19:29:53] <timeless> http://viper.haque.net/~timeless/blog/bugzilla.example.com/show_bug.html [19:30:27] <alanc> the whole bit about "See Comments" & "See Description" is that "Description" fields are published on bugs.o.o & sunsolve - Comments are explicitly Sun Confidential and never published outside the company, so people were trained to put "See Comments" in bug descriptions and put juicy bits in the Comments field - which totally doesn't work for external OpenSolaris developers [19:30:46] <timeless> ah [19:31:18] <richlowe> or anyone. [19:31:27] <timeless> how often do workaround/suggested fix/evaluation change [19:31:40] <timeless> and do they grow, iterate, morph, polymorph... [19:31:46] <timeless> do people care what values they used to have? [19:31:54] *** mikefut has joined #opensolaris [19:32:18] <alanc> unfortunately, "See Comments" was originally supposed to just be for senstive data, where nothing at all could be said publically, but some people misunderstood and trained others to use it for everything, because just telling anyone we have a bug could be seen as bad PR [19:32:19] <timeless> and are they likely to be essays (pages long) or merely 8 lines? [19:32:25] <EchoBinary> i may have asked this before - but i forget - is there a package management system for solaris express, to add new applications and things, similar to freebsd's ports, or debians apt-get? [19:32:42] <timeless> echo: a number of distributions have [19:32:45] <alanc> they can be up to 32k [19:32:53] <timeless> nexenta has apt-get, one has pkg-get [19:33:06] <richlowe> pkg-get is a general thing, I don't think anyone delivers it by default. [19:33:10] * timeless wonders if port install is available somewhere [19:33:13] <richlowe> it'll pull packages from blastwave (or wherever it's told to pull them from) [19:33:34] *** bubbva has joined #opensolaris [19:33:36] <timeless> out of the box, sxce doesn't have one [19:34:04] <timeless> hrm [19:34:09] <richlowe> timeless: depending on who you ask, Suggested Fix should contain the source patch. [19:34:30] <timeless> the patch as diff, or as a link to a changeset? [19:34:35] <richlowe> diff. [19:34:38] <richlowe> given TeamWare is changeset-less [19:34:50] <timeless> sorry, i don't know much about teamware :) [19:34:56] <richlowe> lucky you ;) [19:35:13] <timeless> in a few years new employees might not either, right? :) [19:35:31] <timeless> anyway, aren't diffs in web page text fields likely to corrupt? [19:35:39] <timeless> is that not a problem that people encounter? [19:35:51] <timeless> or is it more of a historical "you should have done X" than a reality? [19:36:31] <timeless> i really wish someone could sanitize a single bug that dates from s7 to s10+1 and is partly fixed [19:36:45] <timeless> thttp://viper.haque.net/~timeless/blog/qa-bugzilla.example.com/show_bug.html [19:37:02] <timeless> was a bug i sanitized from a previous company in order to show what they had done [19:37:50] * timeless ponders [19:37:53] <axisys> i wish i could do zfs send | zfs recv recursively [19:38:07] <timeless> zfs list|grep [19:38:23] <timeless> a bit of work after that involving sorting and perl or something [19:38:24] <axisys> timeless: heh [19:38:31] <timeless> axi: i've had problems like that [19:38:41] <timeless> although my version was actually zfs volume moves [19:38:53] <timeless> it was kinda fun [19:38:57] <timeless> i think it's in hanoi.pl [19:38:58] <axisys> timeless: wouldnt it be nice if we could send recursively? [19:39:34] <timeless> i don't have a place to send it to :) [19:39:53] <axisys> timeless: very funny [19:40:02] <axisys> timeless: so what is hanoi.pl ? your site? [19:40:05] <timeless> besides, i only have 160 or so volumes [19:40:18] <axisys> timeless: wow! sarcastic too huh? [19:40:27] <timeless> axisys: i ran out of space and wanted to rearrange my file system [19:40:30] <timeless> timeless@swift:~% /sbin/zfs list|grep -v '@'|wc -l [19:40:30] <timeless> 158 [19:40:40] <timeless> i couldn't do it because i didn'thave enough space [19:40:50] <timeless> so i would create small volumes, move things around [19:40:53] <timeless> slowly [19:41:11] <timeless> finally, once i had things in bite sized pieces, i wanted to move the tree to where it belonged [19:41:22] <timeless> to do that, i was considering a hanoi style move [19:41:25] *** theRealballchalk has joined #opensolaris [19:41:49] <timeless> in the end, i cheated and just moved everything to a flat numbered n1...nX thing and reassembled the towers [19:42:11] *** theRealballchalk has quit IRC [19:42:38] *** theRealballchalk has joined #opensolaris [19:43:57] <bda> Pietro_S: Doesn't appear crazy. Huge front-end select from another system, then several thousand inserts to sqlite every run. shrug. [19:44:10] <alanc> timeless: bugster isn't web-based [19:44:23] <timeless> alanc: oh, right! [19:44:29] <timeless> um, hrm [19:44:29] <alanc> there's a web interface to view edit, but editing is done via CLI's or the Java GUI [19:44:43] <alanc> bah, "there's a web interface to view bugs, but..." [19:44:45] <timeless> is the java gui live? [19:44:56] <alanc> live? [19:45:07] <alanc> we've been using it for 3-4 years now [19:45:09] <timeless> or can you scrible over most of the fields w/o it committing your changes long enough to snap a screen shot [19:45:16] <bda> (er, thousands of inserts across 20k files or so.) [19:45:30] <alanc> oh, you have to click the save button to commit changes to the database [19:45:35] <richlowe> it's X, who cares about their bugs anyway? :) [19:45:38] *** sartek has quit IRC [19:45:41] * timeless rotfl [19:46:07] <timeless> so... do any of these products have version tables w/ hundreds ofversions for a single bug? [19:46:13] <alanc> no [19:46:18] <timeless> or are things likely to be limited to say not more than 10 verions? [19:46:30] <richlowe> alanc: you could bring up one of the bugs filed from out here, and it (should) be clean. [19:46:46] <alanc> Sun doesn't usually support more than 4-5 versions at the same time of any given product - often more like 2-3 [19:47:16] <alanc> we're only up to about the 12th version of Solaris now [19:47:25] <timeless> but if a bug is filed in say sol5, it'd still have a version entry for 5 even if it's never going to be fixed in 5, right? [19:47:33] <alanc> (ignoring the SunOS 4 rebranding as Solaris) [19:47:36] <timeless> or do you delete it once it gets obsoleted? [19:47:52] <alanc> no - we generally only add the version entries for the versions we plan to fix [19:48:02] <timeless> hrm [19:48:07] * timeless tries to draw this [19:48:14] <timeless> so i file a bug in s9 [19:48:28] <timeless> it's found later to occur in s8, the oldest version field is moved to s8 [19:48:33] <timeless> someone works on fixing it for opensolaris [19:48:39] <timeless> marks a suggested fix [19:48:39] <alanc> then there's a "service record" recording that customer "Timeless" hit this bug in S9 [19:49:01] *** WickySolaris has joined #opensolaris [19:49:04] <WickySolaris> ellows [19:49:11] <WickySolaris> quick question before destroying Linux [19:49:12] <timeless> it's fixed in s10+1 or whatever, and a version field is added for s10+1 w/ the fix version info [19:49:18] <WickySolaris> can I read XFS filesystems under Solaris? [19:49:30] <oxygene> WickySolaris: not easily [19:49:31] <timeless> and then someone suggests/nominates fixing it in s10 by adding another version field? [19:49:32] *** tombhadAC has quit IRC [19:49:35] <timeless> s/field/row/ [19:49:57] <timeless> is this done manually as you choose to add backports? [19:50:08] <timeless> and can i choose to nominate for s8 before i nominate for s9? [19:50:24] <WickySolaris> oxygene: I need to know since all my mp3s and pr0n is on XFS, if its not possible or cant do it without killing my neighbourhood I'll have to do some kind of data swapping first [19:50:32] <timeless> and again (i suspect i asked earlier), how would one indicate that it e.g. doesn't happen in s10 even though it's in OS and s9 [19:50:57] <timeless> it certainly sounds like a versions table might not be that bad, and possibly not even that hard to maintain [19:51:05] <Pietro_S> bda: sqlite3 can force flushing to disk, if he worry about losing data, but even with it I doubt it, but what is clear is that he probably use bad database ;-) [19:51:13] *** WickedWicky has quit IRC [19:51:16] <timeless> assuming i'm not too far off from how you're doing things [19:51:25] <oxygene> WickySolaris: linux is quite a rare operating system with its dozens of file systems supported - solaris itself doesn't know xfs, and I don't know if there's any third party support [19:51:27] <timeless> it's a lot scarier if i have to worry about hundreds of versions at a time [19:51:28] <alanc> timeless: remember that bugster was designed for a business, so people don't go around nominating bugs for backports - customers escalate and we do it [19:51:43] <timeless> alanc: hrm [19:51:59] <alanc> no escalation, then unless it's a security bug or otherwise serious, no backport [19:52:00] <timeless> so it's trivial for a customer to escalate for s7 before one escalates for s8 [19:52:01] *** Gman is now known as gmanAFK [19:52:14] <timeless> or in fact it's trivial for the s8 to never happen [19:52:27] <timeless> and yes, i understand some of these aren't supported, i need numbers to juggle :) [19:52:38] *** hile_ has quit IRC [19:53:04] <alanc> yes and no - customers escalate for the version they are running, then the escalations people make sure it is forward ported to the releases after that so the customer can upgrade in the future, because we like customers to upgrade to new releases [19:53:12] <timeless> is escalation always done a version at a time, ignoring security things, or are they often done as a batch? [19:53:29] <timeless> alanc: oh, heh, right :) [19:54:07] *** _plats has joined #opensolaris [19:55:04] <_plats> hi i have a second drive in my machine which i have partitioned and mounted ( has data on ) but the machine has since rebooted and ive gorgotton the device name i mounted [19:55:22] <bda> Pietro_S: He is rolling over 20k sqlite files. The inserts on each happen very quickly for each file, so it's basically the expected behavior afaict. [19:55:39] <bda> Pietro_S: But yeah, "fixing" it would be good. ;) [19:56:07] <oxygene> bda: 20k separate databases?!? that's nuts [19:56:24] <bda> oxygene: Per-user spam indexes. [19:56:24] <bda> shrug [19:56:31] <bda> Our workload is weird. [19:56:55] <WickySolaris> oxygene: thank you for your answers, I'll try something else [19:57:04] <bda> It's all stored in a pgsql database which is dog-slow. So he's building per-user indexes off it, which are super fast. [19:57:42] <_plats> i have like 42 choices and i am not sure how to find the one i need to mount [19:59:06] *** mikefut has quit IRC [20:00:38] <CIA-26> rie: PSARC/2007/413 Add -zglobalaudit option to ld, 6568745 segfault when using LD_DEBUG with bit_audit library when instrumenting mozilla [20:03:17] *** phonohead has left #opensolaris [20:03:51] *** Fullmoon has joined #opensolaris [20:05:08] *** vmlemon_ has joined #opensolaris [20:05:18] *** baijiutong has joined #opensolaris [20:07:32] *** vmlemon has quit IRC [20:07:42] *** vmlemon_ is now known as vmlemon [20:09:17] <_plats> so how can i list mountable partitions on say c1d0s* [20:09:38] *** cydork has quit IRC [20:09:45] *** switch_ has joined #opensolaris [20:09:59] <_mary_kate_> is there a reason solaris couldn't have a larger default maximum window size for tcp? it really makes a difference for bulk data transfer over long lines [20:11:19] *** switch has quit IRC [20:11:22] *** g4lt-lappy has quit IRC [20:11:32] *** tombhad-AC has quit IRC [20:14:06] <_plats> I have a list from c1d0p0 to c1t0d0s9 and the format . partition . print does not give me a device name not that i can see [20:15:01] <_mary_kate_> _plats: when part shows slice 1 on disk c1d0, the device is c1d0s1 [20:16:04] *** inaddy has quit IRC [20:16:21] <coffman> _plats: c1 is for the controller, d0 for the disk, s1 is for the slice and p1 would be a partition [20:24:26] *** razrX has quit IRC [20:25:40] *** karrotx has joined #opensolaris [20:25:55] <karrotx> does anyone know if the iscsi-target support was pushed into solaris 11/06? [20:27:53] *** razrX has joined #opensolaris [20:28:02] <bda> karrotx: No. It can only act as an initiator. [20:29:07] *** switch_ has quit IRC [20:31:40] <timeless> alanc: so.. [20:31:48] <timeless> can SubCRs have different responsible people? [20:32:08] <timeless> and can they have different Category:Subcategory or Synopsis? [20:32:25] * timeless is trying to figure out how much is actually maintained distinctly [20:33:29] <delewis> did someone ping me earlier? my scrollback buffer doesn't go back very far. [20:33:38] <timeless> delewis: sorry [20:33:48] * timeless finds distinct responsible engineers [20:33:58] <_mary_kate_> delewis: i told jwheeler to talk to you about mplayer [20:40:23] <reflect> are there any plans to make a HCL for opensolaris (as it is further along than solaris and may very well have more drivers) [20:40:37] <cmihai> reflect: the one for SXDE [20:41:07] <reflect> no offence, but that one really isn't very good [20:41:39] <cmihai> Well, then you'll have to check the code [20:41:47] <cmihai> Just use opengrok to grep. [20:41:54] <timeless> alanc, richlowe: http://docs.google.com/Doc?id=dhmd4jxt_56g8gkxt [20:42:24] <reflect> the right thing to do would be to create a proper HCL [20:43:18] <axisys> how do I force hme0 to full duplex on sol 10 w/o using the ndd commands? [20:44:05] <EchoBinary> coffman: used the switch to downgrade that drive to sata1, still no go, just used a pata drive, all is well [20:44:37] <dwc_> yea, the HCL is pretty ugly to use [20:45:34] <reflect> so who do I talk to to get involved in making it better? I believe Sun controls the HCL there.. why doesn't OpenSolaris have its own? [20:45:50] <richlowe> because nobody has set one up. [20:45:59] *** nostoi has joined #opensolaris [20:46:30] <reflect> anyone interested in helping me create one, then? [20:48:36] *** noyb has quit IRC [20:48:41] <e^ipi> why do you need help? just set one up & add all the hardware that you can find in the source [20:48:48] <cmihai> reflect: you start the wiki and allow people to contribute with comments [20:48:51] <e^ipi> usr/src/uts [20:50:34] <reflect> e^ipi: cause I would appreciate comments from people who also thinks the HCL needs an overhaul. two are better than one. besides, I'm not a coder and "browsing the source" isn't one of the things I do best. [20:51:42] <cmihai> http://www.opensolaris.org/os/project/opengrok/ [20:51:45] <cmihai> Use that6 [20:53:45] <e^ipi> yeah, i find it's far more effective to just find the chipset in the source rather than dredge through some halfass HCL that lists 20 variants of the exact same card [20:54:09] <reflect> that's why I propose to make a -real- HCL [20:54:18] <e^ipi> i've yet to ever see one that isn't shite [20:54:30] <e^ipi> so... surprise me [20:55:14] <quasi> reflect: the hcl has a link to a list of supported pci devices - usually that's a whole lot more handy [20:55:22] *** sparc-kly has joined #opensolaris [20:55:22] *** ChanServ sets mode: +o sparc-kly [20:55:54] <e^ipi> and if you don't keep up with integrations, it becomes useless /very/ fast [20:55:59] <coffman> EchoBinary: did you installed the drivers? [20:56:12] <reflect> quasi: where is this link? [20:56:27] *** sparc-kly_ has quit IRC [20:56:58] <reflect> e^ipi: yes, that's why I think it has to be done proper from the start, ie, it shouldn't be done by some outsider, it should be done by someone overseeing code contributions (or have it done automatically somehow) [20:58:19] <e^ipi> set one up, dredge through the source to find supported chipsets, and read the changelogs every two weeks [20:58:40] <e^ipi> ( or whatever schedule the builds come out on ) [20:59:34] <quasi> reflect: http://www.sun.com/bigadmin/hcl/driverlist.html (linked from http://www.sun.com/bigadmin/hcl/ ) [20:59:45] *** cub2000 has joined #opensolaris [21:00:20] <cub2000> does anybody know the network driver used on port #2 of a Sun 2200 ? [21:00:33] <reflect> oh man.. that's what I was looking for.. [21:00:38] <cub2000> nge0 was what I used, but it's not detecting [21:00:50] <reflect> thank you, quasi.. that's ten times better than the HCL ;) [21:00:56] <quasi> cub2000: bge [21:01:10] <cub2000> bge, coooool....thanks a lot quasi [21:01:28] <quasi> cub2000: ehrm, 0 and 1 are bge, 2 and 3 are nge [21:01:45] <cub2000> ohh...I'm on #2 [21:01:51] <quasi> cub2000: if you go by the solaris numbering [21:01:56] *** switch has joined #opensolaris [21:02:07] <cub2000> nge0 is not being detected when I pull the cable out of it [21:02:12] <cub2000> not showing in messages [21:02:49] <quasi> cub2000: you might have to tweak the bios - it is possible to turn them off [21:03:14] <quasi> cub2000: or you could grep for nge and bge in /etc/path_to_inst [21:03:27] <quasi> cub2000: and you did plumb the interface? [21:03:58] <cub2000> of course i plumbed it [21:04:08] <cub2000> let me greb [21:04:11] <cub2000> grep [21:04:59] <EchoBinary> coffman: no not yet, im doing that tonight [21:05:08] <EchoBinary> its my first opensolaris install ever [21:09:59] <axisys> sol 10 does not necessarily read /etc/system file.. so will it read this `set hme:hme_adv_100fdx_cap=1' ? [21:10:52] *** bigjohnto has quit IRC [21:16:04] <cub2000> hmm....interesting [21:16:17] <cub2000> quasi, it's numbered backward on the last 2 ports [21:16:26] <_mary_kate_> axisys: what do you mean, it doesn't necessarily read it? [21:16:28] <cub2000> so nge0 is #3 and nge1 is #2 [21:16:38] <cub2000> thanks [21:16:39] *** cub2000 has quit IRC [21:16:42] <sommerfeld> solaris 10 always reads it. settings may not necessarily behave the same as older releases [21:17:42] *** KermitTheFragger has quit IRC [21:17:57] <alanc> timeless: sorry - was at lunch - SubCRs have same cat/subcat/synopsis, but can have different responsible engineers (bugs.os.o is broken in that it doesn't show the responsible engineers for subcrs, just the main one) [21:18:26] <timeless> not true [21:18:35] <timeless> the bug richlowe showed me had a different RE [21:19:26] <timeless> http://bugs.opensolaris.org/view_bug.do?bug_id=2144965 [21:19:29] <timeless> http://bugs.opensolaris.org/view_bug.do?bug_id=6506010 [21:19:30] <timeless> http://bugs.opensolaris.org/search.do?process=1&sortBy=relevance&perPage=10&textSearch=%22libfont.so.1+in+recent+patches+has+undefined+symbol+%22Xcalloc%22%22 [21:19:43] <richlowe> alanc: see, told you it was a good example. :) [21:20:14] <timeless> ok, so really the only fields that are fixed are cat/subcat/synopsis? [21:20:32] <timeless> i presume when things are SubCRd they get as initial values the original's current? [21:20:55] <timeless> or do they get blanks? [21:22:01] <alanc> oh ah, forgot to click through to the subcr [21:22:04] <alanc> guess that works [21:22:16] <timeless> the buglist view is definitely visibly broken [21:22:23] <sommerfeld> timeless: depends on the CR field. [21:22:25] <timeless> but i'm not 100% certain i can define how ti's broken [21:22:50] <timeless> sommerfeld: http://docs.google.com/Doc?id=dhmd4jxt_56g8gkxt [21:23:10] <timeless> if i gave you edit access could you be more specific? or could you indicate here? [21:23:22] <alanc> if I remember correctly, the shared fields are synopsis, product/cat/subcat, introduced in release/build [21:23:56] *** switch_ has joined #opensolaris [21:24:04] <timeless> http://bugs.opensolaris.org/view_bug.do?bug_id=2144965 [21:24:07] <timeless> shows no introduced in [21:24:10] <alanc> oh, and I think service records are shared [21:24:18] <timeless> Reported Against solaris_8 [21:24:30] <timeless> oh, nm [21:24:36] <timeless> this bug doesn't have introduced in anywhere [21:24:48] <alanc> the introduced in field is blank in that bug - in the main and all subcrs [21:24:49] <timeless> service records are basically "private comments"? [21:25:15] <alanc> service records are customer name/customer contact info and what OS & hardware that customer hit the bug on [21:25:31] * timeless ponders [21:25:44] <alanc> bugs.os.o pulls the OS info out of the SR and shows it as "Reported Against" [21:26:02] <timeless> hrm [21:26:09] <timeless> can reported against have dozens of elements? [21:26:45] <sommerfeld> yes [21:27:10] <sommerfeld> or, well, bugster can. dunno what b.o.o. does to it offhand [21:27:19] <richlowe> the wrong thing, I think. [21:27:25] <richlowe> it'd fit the precedent, anyway. [21:29:02] *** Bart__M has joined #opensolaris [21:30:18] <sommerfeld> (sensible thing would be to sort -u across the CR's) [21:30:24] *** Bart_M has quit IRC [21:31:14] <timeless> sommerfeld: eh? [21:31:19] <axisys> _mary_kate_: well some of the parameters that /etc/system used to use for older version on sol .. not read anymore [21:31:53] <axisys> _mary_kate_: i remember reading something in that line.. may be I am mis phrasing it here [21:32:43] *** Pietro_S has quit IRC [21:32:48] <alanc> sommerfeld: seems to pull the oldest, since internally 6506010 has 13 SR's with reports against 8, 9 & 10, but boo just makes that "Reported against: Solaris 8" [21:33:14] <timeless> x-fade: talk to sp3000 [21:33:17] <timeless> ooops, wrong window [21:33:18] <sommerfeld> timeless: each CR in the database has one or more SR ("support requests"); one field in the SR is the operating system version where the problem was seen [21:34:21] <timeless> sommerfeld: can an SR be attached to just a SubCR? [21:34:41] <sommerfeld> a conceptual "sort -u" (or, more likely, it's SQL equivalent) would show the set of OS releases where the bug had been observed. [21:34:46] *** obsethryl has joined #opensolaris [21:34:58] <sommerfeld> timeless: no. SR's attach to the main CR only [21:35:14] *** switch has quit IRC [21:35:37] <sommerfeld> (within bugster, this isn't ambiguous assuming whoever attached the SR got the OS version right) [21:36:11] * timeless ponders [21:37:07] *** RobNyc-work has quit IRC [21:37:47] <timeless> do the SRs have comments from the requestee? [21:37:55] <timeless> and do they have their own running chains of comments? [21:37:56] <sommerfeld> They can but usually don't [21:38:07] <sommerfeld> just a one-shot editable notes field [21:38:41] <timeless> is it version controlled? [21:38:43] <sommerfeld> generally nothing of interest ends up in there [21:38:44] *** reflect_ has joined #opensolaris [21:38:51] <timeless> or are people assumed to never accidentally corrupt it? [21:39:56] *** tomww has quit IRC [21:41:23] *** reflect has quit IRC [21:41:52] *** reflect_ is now known as reflect [21:43:03] *** vmlemon has quit IRC [21:43:11] *** Arnald has joined #opensolaris [21:45:08] <alanc> most requestees can't see bugster, since it's Sun internal only - comments from internal requesters end up in the comments field usually [21:47:03] <timeless> sure [21:47:25] <timeless> but people can't make mistakes? [21:47:58] <e^ipi> i still think the bug tools should be external, and if a report contains sensitive information it should be stripped out by the responsible engineer & posted to the external site [21:50:18] <axisys> i am looking for a page where it shows what parametes sol10u3 does not read from /etc/system anymore [21:50:23] <axisys> anyone know of one? [21:50:45] <e^ipi> what we've got now is totally useless rubbish [21:50:53] <e^ipi> unless you're a sun employee [21:53:35] <sommerfeld> e^ipi: i think everyone agrees it's a problem. we're in the "who will bell the cat" stage now. [21:54:26] <timeless> ? [21:55:04] <timeless> e^ipi: i'm trying to understnad which features are required in order to have my prefered product seen as a viable option [21:55:33] <e^ipi> the only feature I'm looking for is "not hidden behind Sun's firewall" [21:56:53] <timeless> yeah, you're easy to please conceptually [21:57:01] <wesolows> I think the OGB should confer on it with Stephen's latest draft as the basis, and set a deadline for proposals. [21:57:15] <timeless> url? [21:57:26] <wesolows> Ideally the Tools CG could do this by itself, but I suspect there are too many stakeholders. [21:57:42] <wesolows> timeless: check the tools-discuss mailing list archives [21:58:01] <wesolows> URLs that start with mailbox://wesolows/mbox/... aren't too useful. [21:58:20] <timeless> heh [21:58:39] <timeless> how about a subject that'd help google [21:59:07] <wesolows> Draft DTS requirements, 2 [21:59:11] <wesolows> from 19 June [21:59:26] <timeless> thanks [22:00:04] *** dunc has joined #opensolaris [22:00:23] <timeless> i hate to ask, but DTS=defect tracking system? [22:00:28] <CIA-26> dduvall: 6581614 create_lock() in nightly is racy [22:00:29] <wesolows> yes [22:01:04] <timeless> smi? [22:01:10] <wesolows> Sun Microsystems, Inc. [22:01:10] <richlowe> Sun. [22:01:13] <alanc> timeless: the bugster text fields are not version controlled, but they're mostly set up so that you get a "Are you sure you want to do this?" when editing an existing comment instead of adding a new one [22:01:31] <trygvis> quasi: postwait? [22:01:34] <timeless> alanc: ok [22:01:39] <trygvis> !seen postwait [22:01:41] <Drone> postwait (postwait!n=postwait at pool-71-163-48-241 dot washdc.fios.verizon.net) was last seen in #opensolaris on Sun 08 Jul 2007 20:14 GMT, saying 'the sparc stuff doesn't seem to complain'. [22:01:57] *** quasi has quit IRC [22:04:36] <alanc> pretty much anything external would be better than sticking with bugster at this point, but we need to see what the choices are and not just pick bugzilla because it's what everyone else uses [22:04:55] <timeless> alanc: i'll gladly work to improve bugzilla [22:05:03] <timeless> i just need to figure out what you guys want improved [22:05:22] <wesolows> For starters, I'd like it not to be written in perl. [22:05:37] <timeless> jira? [22:05:50] <alanc> having someone willing to fix it where needed will probably be a factor in the decision making - it was for choosing the code management system [22:05:59] <the-decider> jira's rather nice ;) [22:06:05] *** Gropi has quit IRC [22:06:15] <dgtl__> when I try and # fdisk /dev/rdsk/c4d0s2 it comes back saying it cannot open the device. However it's fully accessible in Linux using that fdisk. anyone have any ideas? [22:06:20] <the-decider> though, there's nothing wrong with perl. Probably the one big thing that's bad about jira is that it's written in that nasty java thing. [22:06:34] * the-decider does not drink the java coolaid. [22:06:40] <e^ipi> javaee is nice [22:06:45] *** enzo_ has joined #opensolaris [22:06:47] <wesolows> no, no java [22:07:05] <e^ipi> jira's non-free though [22:07:08] <the-decider> oh, architecturally java and its stack are a wonderful thing... but that's also its downfall. [22:07:14] <e^ipi> well, maybe... it's free for opensource & non-profit [22:07:15] <the-decider> free to open-source projects. [22:07:18] <wesolows> well, you could contemplate a java XMLRPC backend acting as the database connector [22:07:31] <wesolows> the frontend should be ajax instead of cgi for sure [22:07:34] <alanc> if we're going for trendy, it would have to be written in ruby-on-rails with lots of javascript in the web pages [22:07:34] <e^ipi> the-decider: will they choose to interpret opensolaris as a Sun project though? [22:07:40] <e^ipi> if so, it's not free [22:07:48] <timeless> alanc: there's actually one of those around [22:07:49] <wesolows> JSP and J2EE are definitely not the answer [22:07:52] <timeless> i can't remember its name [22:07:54] <timeless> 13bugs? [22:08:03] <wesolows> ruby would be ok too as a backend [22:08:06] <the-decider> On the other hand, I rather like it and have happily purchased a license for it ;) [22:08:30] <the-decider> Some things are worth spending money instead of trying to get something to work yourself. [22:08:45] <wesolows> Is scarab still being worked on? [22:10:05] <timeless> 16bugs.com [22:10:13] <timeless> if you want ajax coolaid with your bugs [22:10:27] <wesolows> I don't so much care about kool-aid as I care about a usable UI. [22:10:28] <elektronkind> Jira is excellent, from my view. It's replacing our ancient Remedy-based system for trouble tickets [22:10:37] *** quasi has joined #opensolaris [22:10:44] <wesolows> If you can put together a responsive, simple UI without AJAX, great. [22:10:52] <elektronkind> it's 2007 and our org is saying hello to a tt system with a web UI [22:10:55] <wesolows> CGI-based bugzilla ain't it though. [22:11:08] <timeless> tt? [22:11:09] <the-decider> stop telling them how much we suck. [22:11:20] <elektronkind> the-decider: it's amusing though! [22:11:36] <elektronkind> tt: trouble ticket [22:11:40] <elektronkind> er timeless [22:11:53] <the-decider> you need to learn to be like the CIO and simply fantasize that you have a different infrastructure and only talk about that one. [22:11:56] <elektronkind> TRAC doens't seem half bad, from a my user-only perspective [22:12:10] <elektronkind> dunno what it's like to admin it, though [22:12:21] <timeless> wesolows: what par of cgi-based puts you off? [22:12:45] *** Bart__M has quit IRC [22:12:49] <wesolows> timeless: page reloading on every action [22:12:59] <bda> trac lacks useful features like dependencies. [22:13:07] <bda> Writing plugins is really easy, though. [22:13:28] <bda> Also, .11 (dev branch) has the new workflow model, which I think includes deps. Not sure, though. [22:13:51] <wesolows> CGI apps just don't feel like applications; they feel like web pages. [22:14:03] <bda> Heh. [22:14:19] <wesolows> If you want the browser to be a UI application platform, DOM manipulation is the only way to go. [22:14:19] <timeless> wesolows: there's at least one client side app that can speak to bugzilla if you prefer that [22:14:21] <the-decider> well, that's more an aspect of interface design than cgi vs. anything [22:14:22] <bda> Go go gadget XML-RPC? [22:14:23] <timeless> deskzilla (?) [22:14:26] <bda> HTTP-RPC? :P [22:14:27] <wesolows> And for better or worse, that means javascript today. [22:14:30] <elektronkind> REST [22:14:40] <bda> :( [22:15:03] <elektronkind> I've written a SOAP service, complete with WSDL... it made my brain hurt really really bad. [22:15:11] <wesolows> SOAP is waaaaay too complex [22:15:18] <elektronkind> REST seems to be the most basic, and is gaining on SOAP [22:15:26] <wesolows> REST is ok, JSON is ok. [22:15:38] <wesolows> But no SOAP. [22:16:05] <quasi> wesolows: ah, so that's what the smell was comming from ;) [22:16:06] <the-decider> well, the trick to writing SOAP is never to write SOAP, but to use frameworks that take care of all of that for you ;) [22:16:28] <elektronkind> SOAP is still handy for highly structure stuff, though. For example, feed PHP a WSDL and life is golden. Just don't be the poor SOB who gets to write the WSDL [22:16:33] <quasi> the-decider: framework or not, it is still bloated [22:17:06] <the-decider> however, that bloat also comes with some interesting features -- such as WS-Security / SAML profiles, etc. [22:17:14] <elektronkind> I wrote a SOAP provider for IMAP accounts recently [22:17:14] <the-decider> none of which you're probably interested in using ;) [22:17:23] <wesolows> Anyway, bugzilla... [22:18:00] <wesolows> I'm sure at least one group will propose it, and that's fine. We'll evaluate that against the requirements along with everything else. [22:18:27] <wesolows> But it would make me sad to have to use bugzilla in anything like its current form. [22:18:38] *** cmihai has quit IRC [22:18:55] <elektronkind> if bugzilla gets proposed, it probably won't be because it's better, it's just what everyone thinks of when they need to reach for the closest problem tracking software they can [22:19:42] <timeless> wesolows: so.... [22:19:58] <timeless> for security bugs, does the system expect the bugs to be encrypted? [22:20:11] <wesolows> I doubt that's a requirement. [22:20:19] <wesolows> Certainly the current system doesn't do so. [22:20:20] <richlowe> elektronkind: it being widespread is one of the few good points. [22:20:33] <sommerfeld> wesolows: for whatever reason, earlier today I stumbled across a picture of a bizarre "hand soap" product: soap, cast in the shape of baby hands. [22:20:45] <sommerfeld> it's unclear whether this is mor eor less disturbing than SOAP [22:20:46] <richlowe> the way security bugs are handled right now is not a model to build from. [22:21:05] <timeless> heh [22:21:15] <alanc> in bugster, the only special treatment for security bugs is blocking them from being visible to the outside systems like b.o.o & sunsolve [22:21:17] <sommerfeld> the short form is that there is no universially agreed upon good way to handle security bugs [22:21:24] <enzo_> list [22:21:25] <timeless> E3. Differentiated, configurable messaging [22:21:25] <timeless> All mail or other messages sent by the DTS to user-supplied [22:21:31] <alanc> anyone inside Sun can see all data on all security bugs [22:21:32] <sommerfeld> so whatever you do you'll get complaints [22:21:34] <timeless> addresses containing defect data must enforce [22:21:35] <timeless> the confidentiality of defects described in E2. [22:21:42] *** jambock has left #opensolaris [22:21:46] <richlowe> sommerfeld: That's true of defect tracking in general. [22:21:47] <timeless> E2 talked about security groups [22:22:09] <richlowe> timeless: Yes, but it (from memory) is worded so as to avoid the more annoying current problems with them. [22:22:16] <richlowe> (everyone in Sun can see them, nobody not in Sun can see them) [22:22:19] <sommerfeld> richlowe: true. but the real bar fights happen over handling of bugs related to security holes [22:22:20] <timeless> note that i don't really have a solution for them [22:22:21] <alanc> sch's DTS requirements actually describe a better system than bugster currently is [22:22:31] *** enzo_ has quit IRC [22:22:32] <wesolows> I don't really believe the problem with security bug handling is related to the DTS, nor do I expect the DTS to solve them. [22:22:46] <timeless> i mean, in theory smime could solve them [22:22:50] <wesolows> It's a philosophical issue related to processes, not tools. [22:23:02] <timeless> since it's the equivalent of e2 [22:23:13] <alanc> the DTS needs to provide a way to make a bug not publically viewable, and then change it to public when the embargo ends [22:23:18] <timeless> but i don't know of anything that can handle that [22:23:48] <sommerfeld> wesolows: if there was consensus on elements of that process, that might affect requirements for the defect tracking system. [22:24:00] <timeless> oh man [22:24:03] <wesolows> timeless: Really? Why is it so hard, when sending mail to an update recipient, to decide what information to send based on who receives it? I don't see what s/mime would do for us. [22:24:03] <timeless> e5 sounds interesting [22:24:16] <alanc> bugster fails that, because it assumes security bugs will be forever secret [22:24:25] <wesolows> sommerfeld: True. I suppose I was implicitly discounting any possibility of reaching consensus. [22:24:33] <timeless> wesolows: most sstems don't have provisions to store s/mime certificates [22:25:08] <timeless> basically, if a bug is security sensitive, you don't want its contents traveling over the open internet [22:25:14] <sommerfeld> wesolows: I don't think it's likely that we'll solve the philosophical problem any time soon. but we may agree on parts of the solution. [22:25:16] <timeless> it's no longer secure [22:25:18] *** Tpenta has quit IRC [22:25:31] <sommerfeld> or, well, agree on what we'll do for now [22:25:34] <wesolows> timeless: I see what you're getting at, but that's not really what the differentiated content stuff is about. [22:25:41] <wesolows> timeless: The two problems are orthogonal. [22:25:44] <timeless> wesolows: ok, [22:25:50] <timeless> i'm interested in problems i haven't solved [22:25:57] <timeless> bugzilla already has differentiated access [22:26:31] * timeless finds e5 very annoying [22:26:33] <wesolows> timeless: Tell you what...what I'd like to see, since you're gung ho about bugzilla, is a requirement-by-requirement analysis of how bugzilla meets the need. [22:26:49] <timeless> wesolows: based on this? [22:26:50] <timeless> ok [22:26:53] <wesolows> timeless: This would actually be very useful, and gives you a leg up since no one else has done that for any other DTS. [22:27:12] <wesolows> It may encourage others to jump in and get this problem solved sooner. [22:27:30] <richlowe> Yeah, I'm still waiting for the mythical bt2 people lurking on tools-discuss to respond to any of that. [22:27:38] <richlowe> I take it they've given up any desire to have their stuff used. [22:27:41] *** nostoi has quit IRC [22:28:10] <wesolows> sommerfeld: I think, realistically, that what will happen is that there will be irreconcilable differences over how security bugs ought to be handled, which no amount of discussion will help. And Sun, acting in an undefined capacity of tyrant of last resort, will dictate the solution (which will be the wrong one). [22:28:38] <richlowe> I'd say it's very clearly up to the consolidations. [22:28:48] <richlowe> or, if the differentiated access is implemented properly, it's done on a bug-by-bug basis. [22:28:57] <richlowe> that's why I was suggesting that requirement solved much of the problem. [22:28:58] <wesolows> But maybe I'm just overly cynical, since my view on security issues is absolutely 100% diametrically opposed to Sun's, and I see absolutely no merit in it. [22:29:17] <richlowe> it gives you the ability to make exceptions to whatever general policy is argued into existance. [22:29:37] <timeless> ok, I presume RM=Release Manager? [22:29:39] <wesolows> richlowe: It's not clear to me how a per-consolidation policy would work. [22:29:40] <timeless> what's IE? [22:29:46] <richlowe> timeless: IE is initial evaluator. [22:29:50] <wesolows> timeless: Responsible Manager, Initial Evaluator [22:30:09] <timeless> do any of you guys have docs.google accounts? [22:30:10] <wesolows> The former is historically a person (but meaningless in an open world), the latter usually an alias with multiple people. [22:30:28] <wesolows> I do not. [22:30:39] <timeless> http://docs.google.com/Doc?id=dhmd4jxt_57m7pjp5 [22:30:42] <timeless> (public url) [22:30:43] * wesolows doesn't give Google his personal information [22:31:55] *** polk__ has quit IRC [22:31:56] *** SirFunk has quit IRC [22:32:04] *** rbrown_ has joined #opensolaris [22:32:26] *** noyb has joined #opensolaris [22:32:29] <wesolows> timeless: as for E5, it's really quite crucial. Random change in tools is maddening and wastes engineering time. [22:32:42] *** polk_ has joined #opensolaris [22:32:57] <wesolows> timeless: Especially in the case of bugzilla, which has had something like 5 or 10 years of life already, there's no excuse for unmanaged random change. [22:32:58] <timeless> wesolows: oh, i know [22:33:18] <timeless> but i wonder how many open systems really satisfy that [22:33:27] <timeless> bugzilla is nowhere near as bad as linux in that respect [22:33:36] <wesolows> If the answer is 0, then we know what we need to do, eh? :-) [22:33:40] <timeless> but it's nowhere near Solaris's stuff [22:33:58] <timeless> anyway, lemme get back to annotating [22:35:44] <_plats> will solaris fail to boot if it cant mount some thing thats in vfstab ? [22:36:50] <_plats> eg a second drive [22:37:05] *** szt has quit IRC [22:37:08] <alanc> IE would probably map to bugzilla's assign-by-default being set to a mailing list [22:37:31] <wesolows> It could, but assigned-to seems more like RE. [22:37:40] <wesolows> IE and RE aren't really the same at all. [22:37:43] <timeless> i tentatively mapped IE to QA contact and RM to assignee [22:37:54] <timeless> eww, are RM and RE distinct? [22:37:55] <wesolows> RE should be assignee, RM should be deleted [22:38:05] <timeless> ok [22:38:06] <timeless> sorry [22:38:07] <wesolows> RM is a Sun management implementation detail [22:38:29] <wesolows> it would be nice if there were fields that could be used to store that, but it wouldn't make me sad if there weren't. [22:38:30] <timeless> oh good, i got the lettering right in my notes just not in here :) [22:38:45] <alanc> at least the way I'm used to using bugzilla, when a bug is filed, it's assigned to a mailing list, so the list sees the new bugs (like an IE), and then when someone takes it, they assign to themselves (at which point it's the same as RE) [22:38:56] * timeless nods [22:39:12] <timeless> i tend to leave a list like creature in the qa contact field [22:39:18] <wesolows> What's the actual intent behind QA contact? Its name sounds more like the SR Contact (ie, a test team member or something like that)? [22:39:22] <timeless> so that people can watch all bugs in a component [22:39:37] <timeless> qa was responsible for triaging bugs to the right assignee [22:39:43] <timeless> and then verifying fixes [22:39:50] *** rbrown has quit IRC [22:39:50] <alanc> yeah - I've always seen "QA contact" used more as "automatic interest list for this product" [22:40:06] <timeless> at least when netscape was aroudn they used it properly [22:40:09] <wesolows> alanc: Which we also have as yet a separate thing in bugster. [22:40:20] <timeless> in open source where you don't have people lasting long enough to keep sucha role [22:40:28] <wesolows> alanc: Though the subcat interest list actually contains the IE, I think. [22:40:35] <timeless> and where people come/go too often and you don't want to spam everyone w/ mass changes... [22:40:53] <timeless> it ends up better to use it not as a real person [22:41:06] <alanc> not sure - would have to ask someone like bubbva who knows bugster internals better [22:41:09] <wesolows> timeless: Open Source doesn't have to mean flighty and unreliable. We expect people to be committed to whatever they're doing here, regardless of who they work for. [22:41:33] <timeless> wesolows: you guys have lots of expectations [22:41:51] <bubbva> you rang? [22:41:55] <timeless> and you have a hell of a lot more stability than the companies i've seen [22:42:02] <timeless> granted nokia isn't as big as sun [22:42:05] <wesolows> timeless: Engineering is not for the faint of heart or the weak of mind. [22:42:10] <timeless> but it isn't anywhere near as stable at least where i'm sitting [22:42:25] <timeless> things might be different in phones, i suppose [22:43:18] <wesolows> Well, Sun is a very large company, but the part that actualyl works on OpenSolaris is a small part of it. [22:43:47] <wesolows> There are other groups that use bugtraq2 as well, but their needs and interests don't necessarily need to be met by whatever new system we come up with. [22:45:21] <alanc> one of the downsides of the OpenSolaris DTS will be that it's disconnected from the rest of Sun - but that's to be expected for a project that's not just Sun [22:45:35] *** SYS64738 has joined #opensolaris [22:45:49] <wesolows> alanc: Well...that is, if the bugtraq2 people don't put in their own successful proposal. [22:46:09] <richlowe> it'd still, most likely, be disconnected. [22:46:14] <richlowe> they'd just share an implementation. [22:46:35] *** baijiutong has quit IRC [22:47:31] *** polk_ has quit IRC [22:47:46] <wesolows> Well, I guess it depends on the details of a proposal that I haven't seen and may not even exist. [22:48:17] *** JWheeler has joined #opensolaris [22:48:24] <richlowe> indeed. [22:48:37] *** polk__ has joined #opensolaris [22:49:04] <bubbva> alanc: does it have to be disconnected? can't we do a bug bridge? two interfaces to one database? I think there are more options here we aren't exploring [22:49:51] <alanc> bubbva: no, as wesolows said - it will be disconnected if the bugster team doesn't want to participate [22:49:55] *** klang has joined #opensolaris [22:49:59] <alanc> as they currently haven't been [22:50:06] <alanc> I was just assuming they'd stay that way [22:50:28] <bubbva> have they been invited? I haven't seen any dialog on the tools-discuss alias (where we told them participation would be) [22:51:02] <alanc> Stephen Hahn sent out the draft list of requirements - no one responded [22:51:29] <richlowe> a couple of times, I think. [22:51:39] * timeless is confused [22:51:43] * timeless goes back to annotating [22:51:45] <bubbva> hrm. maybe it was when I was off getting married, I haven't seen it. [22:51:55] <bubbva> (there's a lot of unread mail in my inbox from that time) [22:52:05] <alanc> it was mid-June [22:52:26] <bubbva> ah, yes - I was off getting married. I'll see if I can find it & kick some ass a bit [22:52:32] <alanc> '[tools-discuss] Draft DTS requirements' [22:52:43] <richlowe> http://src.opensolaris.org/source/xref/website/spec/dts-requirements/ [22:52:56] <wesolows> Also, 2-way bridges are extremely difficult to implement. [22:53:29] <bubbva> DTS? Ah, I think I deleted that thread - I had no idea that was for defect tracking (though now it makes sense, but I'm sure others missed it as well) [22:53:30] <e^ipi> richlowe: you crammed it on opengrok? [22:53:33] <wesolows> And I'd be very concerned about safeguards needed to ensure we don't create an underclass (ie the same problem we have now). [22:54:00] <e^ipi> wesolows: define underclass? [22:54:04] <alanc> no one has put forward any other systems yet though, so the bugster team isn't far behind anyone else [22:54:14] <e^ipi> a hierarchy isn't a bad thing if it makes sense [22:54:16] <richlowe> e^ipi: I didn't. [22:54:21] <richlowe> e^ipi: it auto-happens. [22:54:47] <wesolows> e^ipi: Whether you have useful access depends on whether you work for Sun (and use the SWAN bugtraq interface rather than the bridged external interface). [22:55:10] <e^ipi> eg, people like me shouldn't have unrestricted write access to the tree, that'd be an insane way to run a project. I haven't earned it [22:55:24] <wesolows> e^ipi: If there are two distinct interfaces, then it's possible that deficiencies in the non-SWAN one will languish. [22:55:39] <wesolows> That in turn degenerates into a non-Sun-employees are an underclass situation. [22:56:00] <alanc> e^ipi: sure, but there's a huge difference between access to the bug db and write access to the source code [22:56:10] <wesolows> e^ipi: The important thing there is which criteria are used to determine what kind of access an individual gets. [22:56:20] <alanc> you need it to be easy for people to report bugs, and provide more data as needed [22:56:31] <wesolows> e^ipi: If it's based on contributorship grants, or other openly-governed characteristics, fine. [22:56:39] <wesolows> e^ipi: it it's based on whether you can access the SWAN, that's not fine. [22:56:49] <alanc> bugs.os.o makes it possible to report bugs, barely able to search for existing ones, and impossible to udpate them [22:57:00] <wesolows> And if just anyone can't file and update bugs, what's the use? [22:57:15] <bubbva> remember, bugs.os.o is NOT bugster. nearly all brokenness of that is related to b.o.o. [22:57:20] *** tomww has joined #opensolaris [22:57:55] <wesolows> Absolutely. If everyone had access to a bugster client and a network path to the bugtraq database server, most of the problems would be solved. [22:57:59] <bubbva> I just had a meeting with the bugster people last week, and they were talking about opensolaris stuff - clearly they aren't talking to the right people? I'll ping them again. [22:58:26] *** lisppaste3 has quit IRC [22:59:00] *** derchris has quit IRC [22:59:03] <wesolows> We all know boo is useless; the question I'd like to ask someone who proposes any kind of bridge is how they will avoid the same problems. [22:59:21] *** derchris has joined #opensolaris [22:59:31] <bubbva> right. you can't get much worse, but we should aim to much better [22:59:36] <richlowe> bubbva: in the majority of cases, "The right people" is tools-discuss. [22:59:44] <richlowe> (or possibly website-discuss, depending on context) [23:00:05] <bubbva> I know in the past they were working directly with Stephen Hahn, but we don't need to start that up again. [23:00:35] <CIA-26> bholler: 6524837 PG disabled sticky scheduler on x86 [23:01:51] *** Dink has quit IRC [23:04:23] *** lisppaste3 has joined #opensolaris [23:04:33] *** tsoome has quit IRC [23:07:25] *** _plats has quit IRC [23:07:42] *** klang has quit IRC [23:09:23] *** klang has joined #opensolaris [23:10:31] <bubbva> So, again, the BT2 folks *are* talking to openssolaris folks - just not on the alias. said Charlie: "I fed a response directly to Stephen Hahn." [23:11:42] *** cub has joined #opensolaris [23:12:07] <cub> A user failed to enter the correct LDAP password many times so the account got locked out. How do I clear this? [23:15:44] *** bigjohnto has joined #opensolaris [23:16:11] <cub> bigjohnto; hi [23:17:18] <alanc> bubbva: Stephen isn't passing that on to anyone else, nor should he be seen as the sole point of contact [23:17:39] <alanc> he's just one of several people interested in this, and the first one to write up a requirements spec [23:17:54] <richlowe> as I said, tools and/or website. [23:18:02] <karrotx> i run `/usr/local/sbin/named -c /usr/local/named/etc/named.conf` it starts and then 2 seconds later it segfaults [23:18:10] <karrotx> is this a configuration issue or is it a build issue? [23:18:42] <bubbva> I just said as much to Charlie. I'll give him a call. [23:18:44] <wesolows> karrotx: Are you asking us as engineers, or are you asking the magic 8-ball? [23:19:00] <karrotx> so the answer is; it could be either [23:19:03] <wesolows> karrotx: Because you'll get equally good answers either way. [23:19:08] <cub> where do I set the LDAP password policy and stuffs [23:19:09] <sommerfeld> karrotx: named shouldn't dump core because you feed it a bad config. [23:19:22] <alanc> with no information about how you built or configured it, what could we possibly say? [23:19:24] <karrotx> that's what i was thinking; but i didn't want to rule that out [23:19:25] <sommerfeld> karrotx: but you're clearly running your own build [23:19:35] <karrotx> i'm actually running sunfreeware.com's build [23:19:36] <wesolows> karrotx: I'd (a) look at the core file, and (b) find out whether this happens with the BIND that's part of OpenSolaris. [23:19:54] <wesolows> Why would you use that instead of the one that's actually a tested, supported OS feature? [23:19:58] <sommerfeld> karrotx: but the real answer could be a little of both [23:20:14] <karrotx> wesolows: the real answer is because this is solaris 10; not opensolaris [23:20:20] <karrotx> but i figured you guys would have dealt with bind builds [23:20:25] <wesolows> karrotx: Isn't BIND a part of S10 too? [23:20:40] <karrotx> 8.4.3 is the package; and 9.2.1 is in /usr/sfw [23:20:41] <sommerfeld> wesolows: "it's in there" [23:20:42] <karrotx> but i want 9.4.1 [23:20:53] <wesolows> I've built BIND myself in the past, but I see no reason to do so since the one we ship works perfectly fine. [23:21:14] <wesolows> Oh right, I forgot. It's "old" which means "bad." Sigh. [23:21:28] <karrotx> no; the 8.4.3 package doesn't support views [23:21:41] <wesolows> The 9.2.1 one sure does, though - I know it does because I'm using it. [23:21:42] <karrotx> and i dont have the disk for 9.2.1 /usr/sfw package [23:21:51] <karrotx> disc* [23:22:17] <wesolows> In that case, the only way to debug this is to examine the core. [23:22:54] <karrotx> i did ::status; and it says it segfaulted [23:22:56] <sommerfeld> "disk" is the more conventional spelling these days [23:23:06] <karrotx> i'm not very handy with mdb [23:23:19] *** derchris has quit IRC [23:23:20] <sommerfeld> karrotx: still in magic 8-ball territory.. [23:23:34] *** derchris has joined #opensolaris [23:23:38] <karrotx> hmm [23:23:51] <karrotx> would you like me to show you the output of ::status? [23:24:36] <sommerfeld> no, because that isn't going to provide useful information about the cause of the crash [23:25:11] <wesolows> The stack trace is a good place to start - $C [23:25:25] <wesolows> Was it reading the config file when it crashed, for example? [23:25:44] <sommerfeld> Also spelled ::stack. ::regs and ::dis can also be informative if you know how to read assembler [23:26:53] <karrotx> http://rafb.net/p/CGxdH258.html [23:26:56] <karrotx> looks like it died on libc [23:27:12] <wesolows> No, it did not. [23:27:17] <tomww> karrotx: you could check this in addition to the other things: ldd named and see, if all libs are rsolved properly (none missing, ...) [23:27:33] <wesolows> It created a thread that started in run(), which then called dispatch() and crashed. [23:28:23] <karrotx> tomww: no libs missing [23:28:25] <richlowe> but given it's threadedness, ::walk thread | ::findstack -v would be good to paste. [23:28:25] <wesolows> You will need to find these functions in the BIND source if you wish to debug further - or take sommerfeld's advice and use ::dis and ::regs to try to understand exactly why it crashed (NULL pointer?) [23:28:43] <wesolows> richlowe: That works in userland? [23:28:43] <karrotx> ::dis isn't a command [23:29:28] *** hohum has quit IRC [23:29:36] <richlowe> it sure is. [23:30:11] <sommerfeld> or <pc::dis [23:30:14] *** hohum has joined #OpenSolaris [23:30:15] <karrotx> http://rafb.net/p/rptFEY65.html [23:30:17] <sommerfeld> (see instructions around faulting PC) [23:30:57] *** baijiutong has joined #opensolaris [23:31:15] <wesolows> richlowe: huh, I guess it does. Learn something new every day. [23:33:50] <karrotx> so [23:37:30] *** baijiutong has quit IRC [23:48:05] <elektronkind> ho omg [23:48:06] <elektronkind> http://opensolaris.org/os/project/comstar/ [23:48:17] <elektronkind> COMSTAR rocks my socks [23:49:38] *** jcsmith has quit IRC [23:53:38] * jmcp reads the stmfadm manpage [23:53:40] <jmcp> gack [23:53:49] <jmcp> same stupidly verbose commands [23:54:16] <jmcp> I bet the usage error messages are just as crappy as for mpathadm [23:54:31] <jmcp> that said, comstar is something we've needed for ages [23:57:56] <elektronkind> looks like stmf is modeling itself after mpathadm *spit* [23:58:26] <jmcp> elektronkind: it's written by the same people [23:58:45] <jmcp> if you don't like it, you should really send an email to storage-discuss to say so [23:58:50] * elektronkind plans to make a case for aliases... 'stmf ll ...' instead of 'stmf list-lu ...' [23:59:00] <elektronkind> and so on [23:59:05] <elektronkind> I suppose I should [23:59:12] <elektronkind> I keep forgetting that we can do that now ;) [23:59:36] <jmcp> they'll still ignore you, but at least an objection will have been made in public