[00:00] maybe! [00:00] shame i know jack about the kernel or coding lol [00:00] # cat /proc/sys/kernel/sem [00:00] 250 32000 32 128 [00:00] let me have a look what those values mean ... [00:03] hmm, no description there, try echo "500 64000 64 256" > /proc/sys/kernel/sem [00:05] ahh found something, yes, this should/could help ... [00:05] values are SEMMSL SEMMNS SEMOPM SEMMNI [00:06] the SEMMNI/SEMMNS is the limit in your case ... [00:07] still there nesh? [00:10] yeah [00:15] does it help/work/fail? [00:26] yuppers [00:26] thanks [00:27] @nesh the iproot message is still there? or already gone? [00:28] gone [00:28] fine .. was probably because the first start failed ... [00:28] hope so :) [00:29] it's probably the stables release since ctx-1 ... [00:29] it's probably the most stable release since ctx-1 ... [00:42] fbc (~fbc@ppp-172-137.26-151.libero.it) left irc: Quit: I'm hungry [01:58] uhg [01:58] i had a hard drive failure and lost most of my VPS's :( [01:58] on one server [01:58] hmm, are you going to blame me for this? [01:59] sorry, that wasn't nice ... :( [02:01] do you have an up to date backup? [02:02] what about software raid to protect against such failures? [02:03] mp RAID no backups no blame u! [02:04] software raid is an easy option, maybe you should consider that now ... it just needs another disk ... [02:04] s/easy/cheap/ [02:04] using raid5 (software) can also increase the responsiveness of a vserver host ... [02:05] yay software raid! [02:05] no more drive bays [02:05] BAWB shut it! [02:06] Dan I should migrate to compaqs :( [02:06] sounds good to me [02:06] hi dan! hmm did I put another finger in his wound? [02:06] i can mount the old disk and when i starta copy [02:06] Bertl: nobody wants "should have" when it's obvious :( [02:06] it time out read errors on me [02:07] imma kill you dan [02:07] Action: Nesh wonders how expensive Data Revory costs [02:07] but I like being alive [02:07] data recovery is expensive [02:07] and my clients are not even mad at me! [02:07] hmm, I didn't use 'should have', did I? ;) [02:07] cause i told em you should have paid for tape backup heh [02:08] Bertl: yeah, you didn't [02:08] Nesh: heh [02:08] maybe I should have done so? *G* [02:08] hehe [02:08] its all good in the hood [02:09] Dan cp a File with ext3 fs on to FreeBSD and then back to linux should not corrupt it right? [02:10] heh, no, it shouldn't [02:10] md5sum is your friend ... [02:10] hmm must have no started the CP in time [02:10] dude i started the CP when i started getting errors like i have time for that! [02:10] [root@host03:/home/dmistry]$ file GAMBIT [02:10] GAMBIT: Linux rev 1.0 ext3 filesystem data (needs journal recovery) (errors) [02:10] fsck.ext3 GAMBIT is a mess ;) [02:11] heh [02:11] when drives fail, they fail spectacuarly [02:12] yay for spectacuarly [02:12] serving (~serving@213.186.189.93) left irc: Ping timeout: 485 seconds [02:12] i wonder how much of a break i can get on the DL380's all RAID'd up [02:13] hmm, maybe I should now bring that anecdote about the lab machines which just behaved as they where named ... [02:13] maybe your server just had to sacrifice that disk ... ;) [02:14] yeah yeah [02:15] Action: Nesh recovers journel's [02:16] journals are no good against data corruption [02:16] shh lemme try man [02:16] :( [02:16] give me some hope! [02:16] Dan you know whats funny? [02:17] well, probably journal replay can't hurt ... [02:17] The only fuckin VPS that i managed to copy off without problems was OKTAYS!!!@!@!@!@ [02:17] the only NON PAYING person on that box! [02:18] haha [02:18] i wonder why klogd is taking up 86% cpu [02:18] everything else had the paid for bit [02:18] :) [02:18] hehe [02:45] GAMBIT: ***** FILE SYSTEM WAS MODIFIED ***** [02:45] GAMBIT: 29526/786432 files (3.9% non-contiguous), 210122/1572864 blocks [02:46] hmm i am scared to try and mount it [02:46] hmm, didn't you say, it's customer data? [02:46] yeah [02:47] guess that makes you a good man then ... [02:47] nah /home is missing lol [02:48] which is the docroot for sites [02:48] lol [02:48] /usr is missing too. [02:48] hmm, everything on one partition? [02:48] there is a shitload of stuff in lost-found [02:48] yeah [02:49] GAMBIT is a 6gig FS i formatted as ext3 and installed vps in [02:49] probably the second level is okay, have a peek into the dirs in lost&found [02:50] yeah the stuff in #34343 etc looks fine [02:50] have no idea how to recover any of it tho [02:50] well probably /home has become some #12345 .. so identifying the vps should be easy then ... [02:51] its only one VPS [02:51] each VPS has its own loopback fs [02:52] ahh, I remember you are the one with the loopback stuff ... [02:52] yeah [02:52] it works unless you have a crash [02:52] lol [02:53] so this _is_ a loopback file or you are looking for one? [02:53] this is the LOOPBACK file i mounted [02:54] okay ... so you have to figure out what dir was home, to regain the data, right? [02:54] right [02:54] do you remember some dir or file below /home/ ? [02:54] i have hundreds of #1231 dirs in lost-found tho [02:54] something like /home/nesh ... [02:55] perllocal.pod [02:55] was in / [02:55] if so, just do ls -d /lost+found/*/ (when there was /home/) [02:57] sitesls -d ./lost+found/*/sites [02:57] ls: ./lost+found/*/sites: No such file or directory [02:57] sites was /home/sites [02:58] hmm, in this case, the corruption must have been worse ... [02:58] Yuupp [02:59] i wonder if there is a way to limit the speed a cp or scp happens [02:59] so it does not thrash the old disk maybe then i can copy [03:00] find /olddisk -type f [03:00] gah, stupid extra big enter key [03:00] find /olddisk -type f | while read file; do sleep 1; cp $file newlocation; done [03:00] or something like that :) [03:00] there is a tool called dd-rescue or something like that ... [03:00] oh yeah, the dd with rate limiting [03:01] keep talking boys [03:02] http://www.garloff.de/kurt/linux/ddrescue/ [03:03] hmm [03:04] Bawb what does your sleep do [03:04] is that just not to thrash the disk [03:05] it's no good for large files [03:05] oh so lemme write this command [03:06] find /mnt -name GAMBIT | while read file ; do sleep 1;cp $file /home/admin; done [03:06] how does that look :) [03:06] GAMBIT is the large file [03:06] and what should that accomplish? [03:06] i dunno man [03:07] me neither ... [03:07] did Dan say that was good for large files? [03:07] I guess this was some misunderstanding ... [03:07] Nesh: no, I said it wasn't good for large files [03:07] i always mis understand dan too smart for me [03:07] something like a rate-limited dd would be [03:08] so dd-rescue? [03:08] yeah [03:08] hmm, nope, why do you want to reate-limit it ? [03:08] man you guys are killing me [03:08] heh [03:09] obviously I've missed something .. maybe you can explain why you want to limit the rate? [03:09] or throughput ... [03:09] Bertl: to try to keep the disk from failing [03:09] hmm, low data transfer helps there? [03:09] yeah [03:09] ide or scsi? [03:09] ok fella's [03:10] SCSI [03:10] i have compiled dd_rescue [03:10] Bertl: it might [03:10] help a brother out. [03:10] [root@host01:/mnt]$ ls [03:10] GAMBIT PROCYBER SMARKET.1 gambit/ petra/ pymecomp/ wheelspecs/ [03:10] PETRA PYMECOMP WHEELSPECS lost+found/ procyber/ smarket/ [03:10] drive mounted [03:10] so far no IO error [03:10] hmm, don't see how this could help, but dd_rescue will 'try' hard to recover what is still available ... from a disk partition or file ... [03:11] won't work on many files or directories though ... [03:11] Nesh: were you able to transfer all the data over in the previous copy? [03:11] no it kept IO time out on me [03:11] only one I got was Oktays [03:11] ah [03:12] must the disk still mounts [03:12] so i though slowing down the cp might help [03:12] it might help keep the disk cooler [03:12] but dd_rescue does not have a rate_limit feature [03:13] nope [03:13] so why say yes! [03:13] I thought it did :) [03:13] Nesh> i always mis understand dan too smart for me [03:13] something like a rate-limited dd would be [03:13] so dd-rescue? [03:13] yeah [03:13] got me all excited [03:13] (cooling the disk) it might also solidify the grease (or how is this called), and destroy the disk ... [03:13] :( now i am still screwed [03:14] Nesh: you are in a maze of twisty passages [03:14] i know man [03:14] there are many people pointing different directions, and no clear map [03:15] :( goodand rate limited dd shows zip [03:15] what you need is flair ... [03:15] no i need a rate limited copy thing! [03:16] hmm, maybe we can diskuss this a little, then? [03:16] discuss I mean ... [03:16] sure [03:16] okay, you _think_ you need something rate limited, because you get I/O errors ... right? [03:16] correct [03:16] and you _think_ that limiting the speed would reduce this errors, right? [03:17] yes sir [03:17] okay, on what do you base this assumption? [03:17] cause i thikn thrashing the disk with fast IO causes time outs [03:18] and its worth a shot at this point. [03:19] so, and there is a tool called dd_rescue, which copies at full speed until an error occurs, then backs off, reduces the block size and tries to reread the faulty space several times ... until it gives up, and replaces the faulty space with zeros ... [03:19] doesn't that more sound like that what you actually want? [03:19] ok [03:19] i have installed dd_rescue [03:19] only one drawback, dd_rescue doesn't work on _many_ files ... [03:20] let me try [03:20] it just works on a partition or something like your loopback file ... [03:21] dd_rescue: (warning): SMARKET (33272.5k): Input/output error! [03:21] hehe [03:25] OH MY GOD ONE COPIED!@!@!@ [03:25] shit lemme test it [03:27] copying rescued file now [03:27] wish me luck fella's [03:27] Action: Bertl wishes nesh luck ... [03:27] k brb smoke while it copies [03:31] vert [03:31] Last message repeated 1 time(s). [03:31] er bert [03:31] yes? [03:31] i should run a fsck.ext3 on the file before i mount it right [03:31] once it finishes cp [03:31] well, depends ... you could try a mount ro without fsck ... [03:32] this will playback the journal .. but not modify the structures ... [03:32] good idea [03:32] 30secs left on copy [03:33] how do you copy it? [03:33] scp [03:33] no option? [03:33] shit it mounted clean [03:33] no options should i? [03:33] probably -c blowfish -C would speed up the copy ... [03:34] ahh [03:34] -C would only help if it was I/O bound [03:34] network I/O bound, that is [03:34] well, it probably is .. [03:38] YAYAYAYYAYAYAAY [03:38] that one started!!@!@! [03:38] lemme try another! [03:38] :) [03:39] hmm, today I added a PayPal link to my site 8-) [03:39] lol [03:39] still have many more to work man [03:39] that is where the concept of 'script' comes to mind ... [03:39] i do not script [03:40] i do things manually [03:40] :0 [03:40] well, then maybe some manual command like ... [03:40] for n in *; do echo $n; dd_rescue if=$n of= ... ; done [03:41] i only have 3 more [03:41] plus this one is erroring [03:41] like a son of a bitch [03:43] i guess i shall let it run [03:44] Additional sense indicates Unrecovered read error [03:44] I/O error: dev 08:11, sector 82511648 [03:44] scsi0: ERROR on channel 0, id 1, lun 0, CDB: Read (10) 00 04 eb 1b d7 00 01 00 00 [03:45] these are the errors that it gives. [03:45] sounds okay for a broken disk ... [03:45] yeah the other one went fast this is alot of errors [04:03] ok i dd-rescued one more [04:03] lets see what happens with this one [04:03] -C -c blowfish ;) [04:04] scp GAMBIT -c blowfish -C root@host03:/vservers/ [04:04] serving (~serving@213.186.190.223) joined #vserver. [04:04] hmm, a real join? serving? [04:05] mm [04:10] BERT [04:10] YAY! [04:10] second one recovered [04:10] did I mention the paypal link yet *G* ... [04:10] working on the 3rd one now [04:10] well you have 3 more to go! [04:10] better keep wishing me luck! [04:10] ;) [04:12] hehe [04:20] 3rd ONE RECOVERED [04:20] @!@!@!@!@!@ [04:21] hmm, I'm not going to mention the paypal link this time ... [04:21] lol [04:28] scp the 4th one now [04:35] 4th one worked! [04:37] hmm, I detect a lack of enthusiasm ... [04:37] 4th one worked!!@!@!@!@! [04:38] ahh .. good ... [04:41] the last one is just too messed up [04:42] hangs right away [04:42] :( [04:42] welll i guess only one not working is good [04:42] the dd? or the mount? [04:42] dd [04:42] well give it some time ... [04:42] no it drops to the prompt [04:42] hmm, what options? [04:42] no options [04:43] never used any on the other ones heh [04:43] is there a help? [04:43] yup [04:43] show me! [04:43] USAGE: dd_rescue [options] infile outfile [04:43] Options: -s ipos start position in input file (default=0), [04:43] -S opos start position in output file (def=ipos); [04:43] -b softbs block size for copy operation (def=16384), [04:43] -B hardbs fallback block size in case of errs (def=512); [04:43] -e maxerr exit after maxerr errors (def=0=infinite); [04:43] -m maxxfer maximum amount of data to be transfered (def=0=inf); [04:43] -l logfile name of a file to log errors and summary to (def=""); [04:43] -r reverse direction copy (def=forward); [04:43] -t truncate output file (def=no); [04:43] -w abort on Write errors (def=no); [04:43] -a spArse file writing (def=no), [04:43] -A Always write blocks, zeroed if err (def=no); [04:43] -i interactive: ask before overwriting data (def=no); [04:43] -f force: skip some sanity checks (def=no); [04:43] -q quiet operation, [04:43] -v verbose operation; [04:43] -V display version and exit; [04:43] -h display this help and exit. [04:44] hmm, what is the error message? [04:44] Summary for WHEELSPECS -> /home/admin/WHEELSPECS: [04:44] dd_rescue: (info): ipos: 80450.0k, opos: 80450.0k, xferd: 80450.0k [04:44] errs: 0, errxfer: 0.0k, succxfer: 80450.0k [04:44] +curr.rate: 112kB/s, avg.rate: 27273kB/s, avg.load: 32.2% [04:44] oh wait shit [04:44] the file on the /mnt is messed up [04:45] it shows file type tar [04:45] lol [04:45] it looks like this file was copied _without_ any error at all ... [04:46] yeah [04:46] i think the file is messed up [04:46] why do you think so? [04:46] i could have screwed it up eailer in the day when trying to recover [04:46] the last one was the one the experimented on [04:46] heh [04:47] and what experiments did you do? [04:47] cpio's tar's [04:47] i think i some point i may have done a tar over the file name [04:47] okay, then try tar tf [04:48] or tar tzf if you used compression ... [04:48] it lists fs structure [04:48] :( [04:48] but not for the vps i blew it up [04:48] hahhhahaaha [04:48] sucks to be that client! [04:49] i have one more option [04:49] which i am trying to put in to play now. [04:49] i managed to get a SCP of eailer and back that one up [04:49] file WHEELSPECS [04:49] WHEELSPECS: data [04:49] uhg [04:49] thats not good should be ext3 [04:57] bert [04:57] aby suggestions? [04:58] hmm, what shall I suggest, I don't know what you've done ... [04:58] maybe you can explain to me, what you might? have done? [04:58] file WHEELSPECS [04:58] WHEELSPECS: data [04:58] instead of ext3 [04:58] this is from another backup from this am [04:59] yeah, could be anything dd if=/dev/zero of=WHEELSPECS bs=1M count=100 gives that for example [04:59] you did an scp of the loopbackfile, while mounted to this machine? [04:59] yeah [04:59] :( [05:00] okay, and you copied the entire file? [05:00] yup [05:00] no compression, no archives, no cpio ..? [05:00] nope [05:00] straight scp [05:00] no options [05:00] well that sounds not too bad ... [05:00] mmm [05:00] make a copy with cp -va (of this file) [05:01] cp -va WHEELSPECS WHEELSPECS.save [05:01] doing [05:02] done [05:02] this is ext2/ext3? [05:02] (or should be) [05:02] ext3 [05:02] e2fsck -fn WHEELSPECS [05:03] e2fsck 1.27 (8-Mar-2002) [05:03] Couldn't find ext2 superblock, trying backup blocks... [05:03] e2fsck: Bad magic number in super-block while trying to open WHEELSPECS [05:03] 1.27? you are still using that old buggy version? ... never mind ... [05:04] okay, you are sure that you didn't do anything you just forgot to mention? [05:04] dunno [05:05] put it on a loopback (losetup /dev/loop/ WHEELSPECS) and try the e2fsck again ... [05:05] ok [05:05] then with -fn -b 8193 [05:06] mount it first [05:06] or run it on the /dev [05:06] ? [05:06] nope no mount ... [05:06] losetup /dev/loop12 WHEELSPECS [05:06] so [05:06] e2fsck -fn -b 8193 WHEELSPECS [05:06] ? [05:06] e2fsck -fn /dev/loop12 [05:06] e2fsck -fn -b 8193 /dev/loop12 [05:06] e2fsck -fn -b 16384 /dev/loop12 [05:07] e2fsck -fn -b 32768 /dev/loop12 [05:07] nope :( [05:07] none of them [05:07] i guess its a lost cause [05:07] luckily [05:08] i managed to pull the archive file from the VPS [05:08] okay xxd WHEELSPECS | head -1000 and put it somewhere on the web ... [05:08] before it died [05:09] http://www.dineshmistry.com/blah.txt [05:10] that is funny, but definitely no filesystem ... [05:10] lol [05:10] nice! [05:10] okay, you are sure that the file size is correct? [05:10] www.f-o-g.com.tar.gz www.jay101.com.tar.gz www.turbospecs.com.tar.gz www.wheelspecs.com.tar.gz [05:10] www.airsoftprofiles.com.tar.gz www.huntersairsoft.com.tar.gz www.krendiharmon.com.tar.gz www.webslumlord.com.tar.gz [05:10] i managed to get the tar files of the sites!! :) [05:10] yeah its the right size [05:11] okay, then try the following: [05:11] ok [05:11] mke2fs -S /dev/loop12 [05:11] done [05:12] now e2fsck -fn /dev/loop12 [05:12] /dev/loop12: ********** WARNING: Filesystem still has errors ********** [05:12] Floating point exception [05:12] now e2fsck -f /dev/loop12 [05:13] ok finished [05:13] ***** FILE SYSTEM WAS MODIFIED ***** [05:13] 11/786432 files (0.0% non-contiguous), 24690/1572864 blocks [05:13] okay try to mount it ... [05:14] [root@host03:/home/dmistry]$ mount -t ext3 -o usrquota,grpquota /dev/loop12 /vservers/tmp/ [05:14] mount: wrong fs type, bad option, bad superblock on /dev/loop12, [05:14] or too many mounted file systems [05:14] ext2 as type ... [05:15] /dev/loop12 5.9G 8.0k 5.6G 1% /vservers/tmp [05:15] heh nothing in there [05:15] well, probably the file is almost empty ... [05:16] ahh yeah [05:16] prolly [05:16] you can make another copy of the WHEELSPECS.save and try with the following sequence ... [05:16] ok [05:16] mke2fs -jS /dev/loop13 [05:16] e2fsck -f /dev/loop13 [05:17] dev12 [05:18] uhg [05:18] no space left left [05:18] eh i give up [05:18] i need a beer [05:19] Bert thanks for all your help man [05:19] no big deal ... I was doing a backup in the meantime ;) [05:19] you da man [05:20] that dd_rescue saved my ass [05:22] bert [05:22] wanna help me with a perl module problem now haha [05:22] :) [05:23] sure ... just let me mention that paypal link again ... [05:23] ok [05:23] ;) [05:23] you can keep mentioning :) [05:24] trying to install the GD perl module via cpan [05:24] okay, what perl module problem could there be, which isn't easily solved via cpan? [05:25] cpan> install GD [05:25] -fPIC -I/usr/lib/perl5/5.6.1/i386-linux/CORE -DHAVE_JPEG GD.c [05:25] GD.xs: In function `newDynamicCtx': [05:25] GD.xs:430: structure has no member named `gd_free' [05:25] GD.xs: In function `gd_cloneDim': [05:25] GD.xs:450: structure has no member named `alpha' [05:25] Last message repeated 1 time(s). [05:25] YADA YDA YDA [05:25] that is a graphics package, right? [05:25] GD.xs:1881: `gdFTEX_LINESPACE' undeclared (first use in this function) [05:25] GD.xs:1885: `gdFTEX_CHARMAP' undeclared (first use in this function) [05:25] GD.xs:1887: `gdFTEX_Unicode' undeclared (first use in this function) [05:25] GD.xs:1889: `gdFTEX_Shift_JIS' undeclared (first use in this function) [05:25] GD.xs:1891: `gdFTEX_Big5' undeclared (first use in this function) [05:25] GD.xs:1895: warning: assignment makes pointer from integer without a cast [05:25] make: *** [GD.o] Error 1 [05:25] /usr/bin/make -- NOT OK [05:25] Yeah [05:25] gd-1.8.4-4 gd-devel-1.8.4-4 are installed [05:26] probably you have the wrong (old, outdated) libgd installed? [05:26] hmm maybe [05:26] mandrake lists gd-2.0.15 ... [05:26] lemme try and upgrade [05:26] this is redhat7.3 [05:29] yeah, perl GD requires at least 2.0.12 (gdlib) [05:41] loger joined #vserver. [05:54] mmm [05:54] no rpm tho [05:55] do you know how to rebuild an rpm? [05:55] I mean build an rpm from a source rpm? [05:55] rpm -bb ? [05:56] rpm -bb file.SPEC ? [05:56] hmm rpmbuil --rebuild ... and rpm -ba .spec ... [05:56] http://vserver.13thfloor.at/Stuff/e2fsprogs-1.34-1mdk.src.rpm [05:57] thx [05:57] try to rebuild that one, if it works ... just remove the 1mdk ... [05:57] and replace it by -1 or something like that ... [05:59] whats that patch do? [06:00] the autoconf patch? [06:00] yeah [06:00] e2fsprogs-1.19-mountlabel3.patch [06:00] thats yo no? [06:01] what does 'yo no' mean? [06:01] your patch [06:01] fsck -version [06:01] fsck 1.34 (25-Jul-2003) [06:01] better? [06:01] hmm nope, not my patch ... [06:02] ahh ok [06:02] though you added stuff ;) [06:02] but maybe you might need that one on rh .. if you use the labeling feature ... [06:03] yeah i did a wget e2fsprogs-1.34.tar.gz [06:03] then a rpm -ta e2fsprogs-1.34.tar.gz [06:03] thats ok right [06:03] well, depends ... sometimes yes, sometimes no .. [06:04] uh well [06:04] thats what I did and i installed the .rpm already :( [06:04] no problem, as long as you do not reboot ... [06:04] really? [06:05] RH7.3 you said? [06:05] yeah [06:05] the thing compiled clean [06:05] it semms there is a e2fsprogs-1.34-1.test.rh73.src.rpm somewhere ... [06:06] why do you think rebooting will break me tho? [06:06] should i downgrade? [06:06] it could be .. you obviously have this patch in RPM/SOURCES, right? [06:06] e2fsprogs-1.19-mountlabel3.patch [06:07] correct [06:07] to me, it sounds like this would add labels to the mount points, right? [06:07] wait [06:07] so your configuration might use :ROOT instead of /dev/hda1 ... [06:07] that e2fsprogs-1.19-mountlabel3.patch was from YOUR rpm from mandrake [06:08] i never used that [06:08] i downloaded the source from sourceforge [06:08] and did a rpm -ta on it [06:08] rpm -ql -p /usr/src/RPM/SRPMS/e2fsprogs-1.34-1mdk.src.rpm [06:08] e2fsprogs-1.23-autoconf.patch.bz2 [06:08] e2fsprogs-1.34.tar.bz2 [06:08] e2fsprogs.spec [06:08] --- so not mine ... [06:08] well right i mean that link you posted :P [06:09] does not contain the e2fsprogs-1.19-mountlabel3.patch ;) [06:09] Cool [06:09] so i am ok [06:09] hmm, why do you think so? [06:10] obviously you rebuilt some redhat rpm on that machine, which probably used this patch ... [06:10] but i downloaded from http://twtelecom.dl.sourceforge.net/sourceforge/e2fsprogs/e2fsprogs-1.34.tar.gz [06:11] which does _not_ contain that patch ... [06:11] correct [06:11] _not_ [06:11] as does not my rpm ... [06:11] oh [06:11] but probably your redhat RPMs _should_ apply this patch ;) [06:11] or soemthing similar for the newer tools ... [06:12] LABEL=/ / ext3 defaults 1 1 [06:12] that what you mean [06:12] yup, that won't work now ;) [06:12] /dev/sda1 / ext3 defaults 1 1 [06:12] fixed [06:13] lemme fix grub too [06:13] kernel /boot/vmlinuz-2.4.23-vserver-1.22 ro root=/dev/sda1 max_loop=64 [06:14] that should all be ok [06:14] hopefully ... [06:14] guess the next time I reboot i shall find out [06:14] for sure ... [06:15] heh [06:15] well, I'll go to bed now ... cu 2morrow ... [06:16] and enjoy the vs1.22 release ... [06:16] Nick change: Bertl -> Bertl_zZ [08:27] say_ (~say@212.86.243.154) joined #vserver. [08:28] say (~say@212.86.243.154) left irc: Read error: Connection reset by peer [13:51] Doener` (~doener@p5082D52B.dip.t-dialin.net) joined #vserver. [13:59] Doener (~doener@pD95883AC.dip.t-dialin.net) left irc: Ping timeout: 512 seconds [14:49] serving (~serving@213.186.190.223) left irc: Read error: Connection reset by peer [15:00] Doener` (~doener@p5082D52B.dip.t-dialin.net) got netsplit. [15:02] Doener` (~doener@p5082D52B.dip.t-dialin.net) returned to #vserver. [15:08] mhepp (~mhepp@r72s22p13.home.nbox.cz) joined #vserver. [15:18] hmm. [15:18] anyone alive? [15:19] At times, yes. [15:29] Doener` (~doener@p5082D52B.dip.t-dialin.net) got netsplit. [15:30] Prove it [15:30] Nick change: ccooke_ -> ccooke [15:31] I'd rather not. [15:32] Doener` (~doener@p5082D52B.dip.t-dialin.net) returned to #vserver. [15:38] ah ha! [15:39] so you admit you don't exist [15:46] No. [16:11] damn [16:11] I thought I had you, there [16:59] Nick change: Bertl_zZ -> Bertl [17:00] the dead ones walk again ... ;) [17:00] hi all! [17:04] afternoon [17:04] bertl: can you confirm one thing for me? [17:04] it's a *silly* thing, but... [17:04] maybe, lets hear .. [17:05] the vserver code doesn't have *anything* to do with PCI, right? [17:05] I mean, nothing called in, say, PCI bus scanning is going to touch a piece of code modified by the vserver patch? [17:05] right? [17:05] or SCSI? [17:05] right ... [17:07] (2.4.23 +vserver) will not boot on most of our web servers, because it doesn't detect the second PCI bus... which the scsi/raid controller is on. [17:07] and 2.4.23 does, with the _same_ config? [17:07] this is the problem I'm having :-) [17:08] see, I have only one spare webserver... [17:08] and it's in the office [17:08] and it's fscking overheating [17:08] this is Frustrating [17:08] so far I've confirmed that a *clean*, *unpatched* 2.4.22 works. [17:08] as yet, the fscker won't run long enough to compile a 2.4.23 kernel... [17:09] huh? [17:09] irmin (~irmin@lake.ptc.spbu.ru) joined #vserver. [17:15] bbrb, just a few minutes ... [17:24] irmin (~irmin@lake.ptc.spbu.ru) left #vserver. [17:25] @ccooke so your system has hardware problems then? [17:25] no... [17:26] what is the problem with compiling a 2.4.23 vanilla kernel then? [17:26] but the only machine I can *test* on is trapped in a hot room, not a nice, cool datacentre [17:26] it'#ll get sorted, it's just *slow* [17:27] hmm, sorry, 'the fscker won't run long enough to compile a 2.4.23 kernel' didn't sound so ... [17:27] that is frustration talking [17:27] I get a bit further each time... [17:29] between 2.4.22 and 2.4.23 there change a lot ... regarding ACPI and bios ... [17:31] yeah [17:31] that's what I'm hoping it is [17:31] well, I hope it's not, actually, but... [17:33] well, you have a lot of options to specify the bus layout on bootup, so if detection is the issue (not misconfiguration ;) then this should be solved easily ... [17:34] could you do an lspci -vvv (of a misdetected machine) and put this somewhere on the web [17:34] ? [17:44] I can detect the second pci bus with lspci -H1 -M -v [17:45] or could [17:45] since I don't have direct access to the machine now :-) [17:46] hmm, sounds like business policy, or what is a webserver without conectivity good for at all? [17:48] well. [17:48] the web servers are in the data centre [17:48] I am in the office [17:48] I have one spare machine here, which I'm using - slowly. [17:48] and you forgot, ssh isn't on your side ;) ... [17:49] ssh is useless in this case [17:49] well, whatever you say ... [17:49] I'd actually need to be there to work on it [17:49] the servers all boot from CD [17:50] this is wonderful... as long as I have a kernel that works [17:50] if 2.4.22 'works' for you, why not use it? [17:50] pflanze (~chris@cc-linux4.ethz.ch) joined #vserver. [17:51] hello [17:51] hi chris! [17:51] does udp still not work with the ip binding stuff? [17:51] if I limit inetd to eth0,eth1,lo, talk won't work anymore. [17:52] bertl: bugfixes. In linux *and* in some patches. [17:52] @chris hmm, you are sure it is udp, which doesn't work? [18:02] shuri (~ipv6@ipv6.electronicbox.net) got netsplit. [18:03] shuri (~ipv6@ipv6.electronicbox.net) returned to #vserver. [18:06] fbc (~fbc@ppp-240-135.26-151.libero.it) joined #vserver. [18:07] hi fabio! [18:07] hi! [18:08] would it be possible to have a "localhost like" interface in a vserver? I was thinking on something like an alias on lo [18:08] jack is working on that ... might be in the next devel release ... [18:09] good! [18:09] I'm very pleased with disk limit patch, it works nicely [18:09] nevertheless you can use 127.0.0.x aliases for now, but do not complain if they aren't secure or woring ... [18:10] my biggest "fear" was about security :) However, I don't explicitly need this feature now, so I'll happily wait [18:11] As you suggested, with few more parameters in vserver conf files and a slightly modified vserver script now I got disk limited vservers without much effort [18:12] great! [18:12] just make sure to share your findings with the community ;) [18:12] oh, I'd like to do so, but I think my modifications are really basic :) [18:13] what would you appreciate more, some basic info, you can laugh about and use it anyway, or no info at all ;) [18:13] I added CQ_INODES_TOTAL and CQ_BLOCKS_TOTAL to the conf files and I calculate INODES_USED and BLOCKS_USED with du, find, cut and sed [18:14] Ok, I understood the concept.. I'll post some info on the ML :) [18:14] thanks ... [18:14] JonB (~jon@129.142.112.33) joined #vserver. [18:14] hi jon! [18:15] hey Bertl [18:23] okay, have to leave now ... cu l8er ... [18:24] Nick change: Bertl -> Bertl_oO [18:28] fbc (~fbc@ppp-240-135.26-151.libero.it) left irc: Quit: I'm hungry [20:06] dam big snow storm here [20:16] shuri: then stay inside ? [20:21] yes [20:21] :P [20:22] manaha (~carlos@213.37.96.229) joined #vserver. [20:24] manaha (~carlos@213.37.96.229) left irc: Quit: Leaving [20:54] Alex (alex@alex.city17.org) left irc: Ping timeout: 499 seconds [23:24] BobR (~georg@MAIL.13thfloor.at) joined #vserver. [23:34] ensc (~ircensc@ultra.csn.tu-chemnitz.de) joined #vserver. [23:34] hi [23:39] hey ensc [00:00] --- Tue Dec 16 2003