[00:00] I figured you were quietly reading. :) [00:02] erichammond: I'm reminded now of sar, too. [00:12] tbc: I've used sar casually but I think it only reports at the same level as the rest of the tools (i.e., user, system, iowait, nice, steal, ...) [00:12] erichammond: that's what I read in the collectl docs just now. Thus collectl may help you more than other tools you've tried. [00:12] Based on my speed reading so far, collectl seems to be similar, unless there is a plugin which does something more. [00:12] ah [00:13] You might at least write to the list of post on the forum to see what that crowd has to say. [00:13] As it turns out, we may have identified the cause through other means. I just noticed that the fork rate increased at the same time on the munin charts. [00:16] is there an x86 version of ubuntu server? [00:19] !x86 [00:27] tbc: We think we've tracked the cause down to the invocation of an external process to calculate MD5 signatures instead of using the CPAN MD5 package. Testing the fix soon. [00:28] glad to hear that [00:33] I have a frustrating problem with my Ubuntu server install. It periodically goes into some sort of suspend state, where it doesn't respond to ping/ssh, etc. and, more strangely, doesn't keep the system time advancing. If I touch a keyboard key it immediately wakes. Could anyone help me figure it out? [00:36] tbc: Thanks for your ideas. We track average page generation time with munin graphs. After our last release, that jumped up a bit (200ms) which we correlated with a jump in system CPU and then to a jump in forks. Then, examining the deltas in the revision control we identified the culprit. [00:37] Excellent! [00:37] I love web development with fast release cycles. [00:40] heh [00:41] ja660k: re. !x86 architectures. Yes. See https://help.ubuntu.com/8.04/installation-guide/i386/hardware-supported.html [00:41] People should hang around a little longer for answers. J [00:41] wweasel: Did you install only server packages, or is X also running? [00:42] tbc, server packages only. [00:42] erichammond: yeah, no problem w/fast releases as long as troubleshooting velocity keeps up. IRC rocks IMO. [00:42] It's on a laptop, but obviously AC connected. [00:43] wweasel: sounds like some random laptop package is trying to save on power. But I haven't encountered that before. Sorry. [00:45] tbc, Yeah :/ It would be nice if I could even know what exactly is happening. Would you know which log file I ought to look through? It's not clear to me. [00:46] wweasel: /var/log/messages [00:46] * tbc is guessing now [00:47] wweasel: Send 'ps -fe' to http://ubuntu.pastebin.com/ and I'll have a look. [00:49] wweasel: make that 'ps aux' [00:49] http://ubuntu.pastebin.com/m7c27bf38 [00:49] * tbc reads [00:49] wait wait, that was ps -fe [00:49] np [00:50] * tbc has old habits [00:50] shall I pastebin ps aux as well, or no worries? [00:50] Yes, let me see that, too. [00:52] http://ubuntu.pastebin.com/m58a456a4 [00:53] hi is there a way to exclude dependencies from packages? i run a hadles ubuntu server without X an a packege depends on X but i don't want to install it [00:54] wweasel: you're running some KDE processes. Those are desktop packages. [00:54] * tbc didn't expect server edition to install KDE [00:54] hrm. That's really odd, I don't think I've installed any KDE packages at all. [00:57] I've seen KDE get sucked in when I've installed an app that uses a GUI. zorzar: you're asking something similar. I'm new to the server edition and may be wrong, but I can say that if one is running a server one should be able to get along w/o X. [00:57] We need a tool that shows all installed packages and what they depend on. I'm sure it exists, but I'm rusty. [00:57] yeah, I mean no X applications [00:57] *nod* [00:57] I just pastebin'd my dpkg --get-selections [00:58] that's all my installed packages [00:58] *chuckle* wweasel - URL, please? [01:01] http://ubuntu.pastebin.com/m1845464c [01:02] reading... [01:03] Like, I'm not all that advanced a linux user, but I've also been using linux for a while. The closest thing I've installed to a desktop app is Deluge. [01:03] but I installed the backend and run a gtk client on my desktop [01:03] Sounds reasonable. [01:03] and even then, that wouldn't have any kde dependencies - it's gtk [01:03] yup === erichammond1 is now known as erichammond [01:04] I wish I could get you the list of installed packages on a clean install and just show the diff with my install. [01:04] np [01:05] We can get that. Still thinking... [01:06] * tbc thinks his KDE suspicion was wrongheaded [01:07] I'm pretty sure there's no desktop components installed [01:07] it's a recent install (1 month) [01:07] so all the packages i've installed are in my recent memory :) [01:08] wweasel: sorry, I have to go soon, but have a look at this list: http://ubuntu.pastebin.com/d580449e0 [01:08] Those are the obviously-named kernel daemons [01:09] I think instead of a package you have some ACPI event putting your system in a less-than-fully-conscious state. [01:09] But that's only a hypothesis. I could be wrong. [01:09] what about them? [01:10] It's plausible [01:10] The last time I was troubleshooting, the guy I was working with suspected it might be ACPI-related [01:11] And suggested I start the server with the kernel parameter acpi=off [01:11] In fact, that's how it's running right now, and no luck...it still suspends. [01:11] ah [01:11] For instance, why would kgameportd be running on a server? I don't think that daemon needs to run. [01:12] Not sure. Do you have a server install? Check if it's running. [01:16] * tbc only has a notebook running Linux atm [01:17] hrm [01:17] Well, it's something power-saving related [01:17] It wouldn't seem to be suspend, because it wakes *instantly* on a keyboard press [01:18] so it's more likely "standby" state. [01:18] Yeah. Waiting for user IO. === dendro-afk is now known as dendrobates [01:18] yeah [01:18] I've never seen anything so crude it shut off so much, though. That's what intrigues me. [01:19] It's an old eMachines laptop from 1994... [01:19] sorry, 2004 [01:19] not 1994. [01:20] big difference :P [01:20] it also has corrupted video ram, making odd distortions on the screen. but that's irrelevant as well. [01:21] Heh. [01:21] I have an old Compaq with a broken display. Boots fine. Nice server admin'd via ssh. But I haven't turned it on in years. [01:23] yeah, that's the idea with my eMachines. less power hungry than a tower, completely unusable as a desktop (even the VGA-out is corrupted by the video ram) [01:23] i'm presently wondering if I should try booting with acpi=off, apm=off [01:24] That sounds like something to try. After all, you can easily reproduce this problem, right? How long does it take? [01:24] For it to go semi-conscious? [01:25] I haven't timed it. I don't think it's consistent either. [01:25] But generally in about 15 min - an hour. [01:25] so easily reproducible [01:25] although it's very weird: "By default, hibernate/standby is disabled on the server edition", according to the internets [01:26] But it's neither suspending nor hibernating. [01:26] no, but it's standby-ing. [01:26] I think [01:26] My first guess was a screensaver. But that's too obvious. [01:26] I had a ubuntu wiki page describing the difference between the three [01:26] Well, it does know to turn off the laptop's screen after a certain period of IO inactivity. [01:27] and does so even though I have acpi disabled [01:27] But screen blank shouldn't be causing the other brain damage. [01:27] So I'm intrigued. [01:27] Could be a bug. [01:27] Indeed. Apparently it's not just a screen blank [01:27] * tbc likes bugs [01:27] bugs are a chance to improve the system [01:28] for everyone [01:28] could be. would probably have to be reproducible on another system though [01:28] Finding them, that is. [01:28] Yeah. But if you can reproduce, it could enough to find & fix. [01:28] Yep [01:31] Alright, well, I'm trying acpi=off apm=off. [01:31] And going to go have dinner. [01:31] Much as it's kind of curious, it's also REALLY frustrating. [01:32] I can imagine. I'll be online for a while yet. [01:33] Alright, I'll be back later! Thanks for all of your help [01:33] Glad to be here. === dendrobates is now known as dendro-afk [02:02] Hello. I'm trying to install ubuntu server form usb.. but it keeps saying Could not find kernel image: Linux boot: [02:02] What could be the error? [02:09] do you get past the grub menu to select the image? [02:10] stimble_; [02:10] No it's just at the start [02:10] SYSLINUX 3.63 Debia-2008.... [02:10] so you don't even see the grub menu? [02:10] No [02:11] have you installed grub to the usb device? [02:11] grub-install /dev/sdb [02:11] or similar [02:11] I'm using unetbootin [02:11] I'm trying to boot of the usb device [02:11] and install ubuntu server [02:11] i dont know anything about unetbootin [02:12] simple well used linux boot usb [02:12] your machine does not support direct netowrk booting? [02:12] that might be a better bet [02:12] I'm burning a new usb now [02:13] Is it easy to install a server with pxe btw? [02:13] unetbootin looks pretty cool tho [02:13] yer its great [02:13] i'd never heard of it [02:13] been usin it for years [02:13] pxe works, yeah [02:13] i have no experience of it [02:13] my dell poweredge 750 searches for pxe at boot [02:13] if set to it i.e [02:14] I've tried hookin up a cd rom reader etc to it but dun work [02:14] last option is to move the hdd to a friends house [02:14] and install it over there [02:14] but, how easy is it with pxe? [02:15] its not too tough [02:15] u just need a tftp server, and dhcp setup [02:15] and grap the netbooot images [02:15] confiugre tftp to server em up [02:15] bah [02:15] and dhcpd to tell the requester to request the netboot image [02:15] settin that up on me netbook is gay [02:15] not easy peasy, but doable [02:16] ha [02:16] tried to hook up the s-ata disc to my ion [02:16] ion asrock 330 [02:16] but apparently they have their own hyped s-ata cables [02:18] kinda bugs me [02:18] i have to setup mythbackend on a crap computer now [02:18] cuz my srv wont boot [02:19] what does 'wont boot' mean/ [02:19] well [02:19] it boots to a concurrent old installation of fedora [02:19] but I can't manage to insatll anythin to it [02:20] or [02:20] how can i put it [02:20] I can't reinstall.. like wipe it all out [02:22] and [02:22] this is ubuntu server channel [02:22] yer I knw [02:22] know [02:22] I want to install ubuntu server [02:22] ok then when you boot the server cd, then it comes times to setup partitions go manual and set mount points and set format = no on them [02:22] it will install over what is there with the things from the server cd [02:23] anything else will be left alone [02:23] yer but ya not followin bogeyd6 [02:23] also, make sure you setup a different user [02:23] well [02:23] this old comp [02:23] will probarly boot just fine [02:23] from ubuntu server from a usb [02:23] but my poweredge 750 [02:23] I just cant get it to boot [02:23] dno how ima solve it [02:23] i dont know how to solve a fedora booting problem, try 3fedora [02:24] hmm this is a new keyboard and the keys still dont work, yet a usb keyboard will [02:25] a fedora booting problem? [02:25] uhm [02:25] an installation problem [02:25] that it currently concures fedora is just a side-note [02:25] not installed by me [02:26] I want to wipe it clean [02:26] as i said earlier [02:26] hmm, my control keys dont work either so no copying pasting [02:27] dat blows [02:27] so you want to wipe the server clean [02:27] yer [02:27] i've tried almost all versions there is of it [02:27] but i mean [02:27] next step is either to bring the hdd to a friend [02:27] or borrow and external cd reader [02:27] dariks boot and nuke [02:28] wipes the hdd completely clean of anything and you can start all over [02:28] well [02:28] i cant boot of anything [02:28] ive tried usb [02:28] but im confused because the ubuntu server cd install will re-do the partitions [02:28] cd [02:28] ok, so its a hardware problem/ [02:28] well [02:28] not really a problem [02:28] or not really a hardware problem [02:28] more lackthereof [02:29] you are making absolutely no sense at all [02:29] the server wont boot a cd, yet its not a cd probme [02:29] did you check the bios to change the boot order to cd, usb, then hdd [02:29] ofcourse, but it's all fuzzy [02:30] the fudge [02:30] ok, im done, good luck [02:30] this is like the fifth usb i've burnt [02:30] that says [02:31] could not find kernel image: linux [02:31] how do you burn a usb [02:32] unetbootin [02:32] jolaren: can you boot another machine from the usb [02:32] it is possible the server wont boot usb [02:32] does it only fail on the poweredge [02:32] lots of computers wont install from an external usb dvd drive [02:33] yeah, lots of older hardware dont support usb boot [02:33] well [02:33] well now im not trying to boot the server [02:34] i'm trying to boot annother computer [02:34] a old one [02:34] a brand new at the time dl 380 wouldnt install windows from external dvd drive either [02:35] stimble_; [02:35] ye [02:35] stimble_; Other machines boots good [02:35] strange [02:35] probarbly bios settings [02:36] well, then that machine just dont support usb boot [02:36] so use cd or pxe [02:38] well [02:38] isnt it strange [02:38] that I get up the loader? [02:39] The [02:39] not strange at all [02:39] SYSLINUX 3.63 Debian-2008... [02:39] Okay.. [02:45] jolaren: the answer is CD or PXE [02:45] stop fighitng it i say [02:46] or take the drive out and install to it from another computer === erichammond1 is now known as erichammond [03:01] stimble_; well i guess ill have to get a cd tomorrow then [03:01] but then i might aswell try to run over to a friend [03:01] or pxe [03:01] just wonder who has sata discs [03:01] pxe seems complicated [03:01] allthought [03:02] a chance to learn:-) [03:02] its not that bad [03:02] not cake, but not terrible [03:04] well but still lad [03:04] reckon if i lay 3 hours of my time doin that [03:04] and it wont work [03:04] reckon my rage === dendro-afk is now known as dendrobates [03:08] tho stimble_ [03:08] would an external cd drive work or should I just go hunt a friend [03:08] no pain, no gain [03:08] usb is your prob im thinking [03:09] so would not count on usb cd [03:09] i would at least try pxe [03:09] there are lots of docs and howtos on it [03:09] yer ive seen plenty [03:09] but their pretty intense [03:09] and in some cases you need floppies? [03:14] picka good one [03:14] and u dont need floopies [03:15] so dont pick one of those [03:26] will be easier to just hope some of my friends [03:26] has a computer with a s-ata drive === dendrobates is now known as dendro-afk [04:09] Hi there iam tring to install a new schema and i am getting this error : ldap_add: Server is unwilling to perform (53) [04:09] additional info: no global superior knowledge [04:09] any idea? [04:45] I'm trying to stream ffmpeg using ffserver, but when a client tries to connect, I get an error message: "Error writing output header." Not much in the way of help from our or other forums. Anybody seen this? [04:52] Roxyhart0: I googled and found this. Does it help? http://www.openldap.org/faq/data/cache/1322.html [05:00] thanks tbc.i read that but still i can't see where is the problem in my cnfiguration as i follow the configuration guide exatly that it say [05:01] how do i generate a text file of all the installed packages on my ubuntu server? [05:01] bizkit: dpkg --get-selections | cut -f1 [05:01] ah, wait... [05:02] dpkg --get-selections | grep -v deinstall$ |cut -f1 [05:02] Roxyhart0: you'll have better luck connecting with LDAP community. I'm not an LDAP expert. [05:03] if i declare dc=mydomain, cd=edu,dc=au. Ahould I declare also dc=edu adfter that? [05:07] * tbc wonders if someone else on this channel can help Roxyhart0 [05:15] tbc: Without (more or less) doing all his LDAP schema creation work for him, probably not. I think at this point Roxyhart0 needs to better understand LDAP and openldap, and then create a working ldif file to add the new schema. See http://www.openldap.org/doc/admin24/guide.html and (for adding/extending schemas) especially Chapter 13 of that guide. [05:16] * tbc didn't think Roxyhart0 had picked the best channel for his question [05:16] That's probably true too. I spent some time yesterday (or the day before?) helping him get openldap installed and working, BTW. [05:17] jmarsden: on this channel? [05:17] who is him? [05:18] i am her [05:18] Roxyhart0: I'm sorry, Rosa. I should have checked your info. [05:18] d'oh [05:19] That was jmarsden's error. :) [05:19] tbc: Yes. Roxyhart0: My bad. Anyway, do read the guide and see if it helps you understand what is going on. [05:20] thanks :D [05:22] So Roxyhart0 do you need help connecting with the OpenLDAP folks? [05:22] I've been meaning to look up a friend who's an LDAP expert. I'm wondering if he's active in the OpenLDAP community. [05:24] really? if he is still in the comunity could be great talk with him [05:32] Roxyhart0: start here: http://ostatic.org/openldap [05:33] do i need to configurate or add to the tree manually cn=schema? [05:36] Roxyhart0: I don't know. Your best course of action is to visit http://ostatic.org/openldap and join their community. It's late and I need to reboot. I'm https://launchpad.net/~tbc0 - I'll be online tomorrow if you still need help connecting with them. [05:37] thanks very much [05:37] yw [06:42] Does anyone know if eucalyptus has support for amazon style elastic map reduce? [07:17] ls [07:17] oops, sorry :) [07:25] anyone ever recovered data from a mdadm raid who's superblocks were screwed n chewed? Ive read a gazillion things googling but nothing seems to work. Im running xfs_repair /dev/md0 right now but nothing yet, it keeps saying "found candidate secondary superblock ... unable to verify superblock, continuing" Is there another way to replace the superblocks? [07:25] Whats the expected amount of time to let this run? There is barely any data on the raid, just a couple hundred megs. [07:40] trying TestDisk now... normally my obsessive backups would save my ass right now but they silently failed 6 important working days ago .. isnt that nice! [07:42] New bug: #516083 in eucalyptus (main) "Missing jar file when Walrus/S3 runs on different host than CLC" [Medium,Fix released] https://launchpad.net/bugs/516083 === viezerd- is now known as viezerd [08:24] what files does ubuntu change to share folders while setting them up in the X system?, coz I've shared some of them in that way and I don't see them in the /etc/samba/smb.conf file [08:26] somebody know a guide to instal postfix, ouried idmap with ldap? [08:43] strange Testdisk found Ext2 superblocks in my XFS drive. That doesnt seem proper ... and may be why xfs_repair wouldnt work. [08:45] somebody had a doc how to instal postfix with ldap support? [08:45] macrocosm: Sounds like maybe it was really ext2 or ext3? In which case, go back to the images of the raw partitions from before you ran xfs_repair, and try again using e2fsck and friends? [08:47] im certain its an XFS drive though, strange [08:48] maybe thats how ive screwed this up ... I must have done something that overwrote the superblocks as ext2 [08:49] macrocosm: Only the superblocks? How would you do that? [08:50] I have no clue.. my apache stoped working and I restarted ... upon return md0 was no longer mounted .. when I tried to mount I got all these errors about the superblock [08:50] I don't think I could write over all XFS superblocks in a filesystem with ext2 ones if I really really tried, except by hand with a sector editor and a LOT of time and patience... that just doesn't sound plausible. [08:51] I cant explain why testDisk would state that the superblocks were ext2, perhaps its confused [08:52] Hmmm... are you confusing RAID superblocks with filesystem superblocks? [08:52] well .. It is a raid0 .. but the drives are formated XFS [08:53] and the testdisk im running is only on that raid .. which has no system files on it [08:54] So there should be RAID superblocks which tell the system that this is a RAID0, parameters of the RAID are X, Y Z.... *and* once you treat the thing as a RAID device, there will then be XFS superblocks on top of that. [08:54] But until you have a RAID device, you can't even begin to see or check for a filesystem... unless I'm VERY confused? [08:55] yea ... the drives are together in the raid... I just cant mount it because of the superblock error [08:55] Does cat /proc/mdstat show that the RAID device md0 is fine and happy? [08:56] yeah [08:56] Personalities : [raid0] [08:56] md0 : active raid0 sdf[0] sdi[3] sdh[2] sdg[1] [08:56] 314572544 blocks 64k chunks [08:56] unused devices: [08:56] OK. So, what does file /dev/md0 output? [08:57] hmm irc doesnt like the output .. do I need to put it inside some special brackets or someasmuch? [08:57] just one line [08:57] dev/md0: block special [08:57] ahh that worked [08:57] any experience with courier-idmap and ldap? [08:58] sorry roxyhart0 ive not used that [08:58] Roxyhart0: You should probably try the courier folks of the ldap folks for that. [08:58] s/of/or/ [08:59] i did, but with ubuntu is a little bit different everything [09:00] Yes, but it you understand what courier and what ldap are doing, that knowledge should apply just fine to Ubuntu, even if the defaults are slightly different. [09:00] It's not like the courier config file format is different or anything like that. [09:02] jmarsden .. is /dev/md0: block special .... a normal output from cmd ... file /dev/md0 ? [09:02] Roxyhart: Sorry - don't use Courier or LDAP. However, Howtoforge often has useful steps for install/setup of this sort of thing. [09:02] Try google : site:www.howtoforge.com +postfix +ldap +courier [09:02] yes howtoforge is awesome [09:03] macrocosm: Yes. It's a valid block device. That's fine. I'm not familiar with XFS enough to know how to proceed in detail; I'd tend to try and find out how to identify an XFS filesystem at a low level (is there a signature in the first block of the disk, or whatever) and make sure the raw data on /dev/md0 "looks like" an XFS filesystem [09:04] If it does, you need to try to repair or rewrite that bad XFS superblock. [09:05] well .. I have been using it for about a month with no problems .. however their was a bug with the latest ubuntu and XFS which caused it to not mount on reboot sometimes... apparently the last time was the last time lol. Ill be good as long as I can get my data off of it... hopefully testdisk will help [09:05] If you know what you are doing with filesystems, see what xfs_db can do for you... [09:06] BTW, next time, don't use RAID0 for data you care about :) [09:06] RAID1 for important data would be a lot more robust. [09:06] macrocosm: there's an xfs_check as well I think. Plus an XFS mailing list - might be worth checking that. [09:07] well .. I would normally not have done that but amazon EBS drives are fully redundant [09:07] supposedly [09:07] and I have obsessive snapshots .. but they silently failed 6 days ago .. when all this transpired I was so happy to pop over and grab a snapshot to restore from but there were no recent ones lol [09:08] gotta love a perfect storm [09:10] macrocosm: So now you get to write a cron job that emails you if there are no snapshots less than 24 hours old... but that doesn't help with the data currently on this array. [09:11] True .. I really wish I had done that to begin with... would have saved me some headach.. but the upside is I have learned much more in the process so thats always good [09:12] I really appreciate your help and tips .. hopefully testdisk will help me get my files off ... ive backed up these dead turd drives so I may try banging them around some first though.. [09:13] Yes. One more idea... try dd if=/dev/md0 bs=1M count=10 >testmd0 ; file testmd0 # this should let file say what kind of fs it saw [09:13] 10+0 records in [09:13] 10+0 records out [09:13] 10485760 bytes (10 MB) copied, 0.205319 s, 51.1 MB/s [09:13] testmd0: x86 boot sector; partition 1: ID=0x83, active, starthead 0, startsector 256, 629143552 sectors [09:14] Ah, that was not a partition that was a "drive" complete with partition table. [09:14] So what does fdisk -l /dev/md0 say? [09:15] Disk /dev/md0: 322.1 GB, 322122285056 bytes [09:15] 2 heads, 4 sectors/track, 78643136 cylinders [09:15] Units = cylinders of 8 * 512 = 4096 bytes [09:15] Disk identifier: 0x00000000 [09:15] Device Boot Start End Blocks Id System [09:15] dev/md0p1 * 33 78642976 314571776 83 Linux [09:16] OK. 83 is the normal ext2/ext3 type byte... but I don't know whether XFS uses it too. [09:17] Hmmm, so what does dd if=/dev/md0p1 bs=1M count=10 >testmd0p1 ; file testmd0p1 say (the last line is the interesting one) [09:17] yes its 83 [09:18] dd: opening `/dev/md0p1': No such file or directory [09:18] testmd0p1: empty [09:19] Oh, hm. So that's not a real device name at all. OK, never mind. It's too late at night for me to figure out the right way to do it... I need to sleep :) [09:20] Oh... maybe dd if=/dev/md0 bs=4K skip=32 count=100 >testmd0p1 ; file testmd0p1 [09:21] 100+0 records in [09:21] 100+0 records out [09:21] 409600 bytes (410 kB) copied, 0.001396 s, 293 MB/s [09:21] testmd0p1: SGI XFS filesystem data (blksz 4096, inosz 256, v2 dirs) [09:21] what is that? a test? [09:21] OK! Got it! [09:22] Yes, it means there really is an XFS filesystem out there at 33 cylinders into your md0 disk. [09:22] So some kind of XFS filesystem recovery tool has a decent chance of getting your data back. [09:23] You can delete the test* files now, by the way :) [09:23]