1118016717 M * Doener` hm, getting mysql to use a lot of rss isn't as easy as i thought with some random db... 1118016817 J * ben_ ~ben@modemcable235.254-80-70.mc.videotron.ca 1118017123 M * Doener` rs: ok, i got mysqld to be 'killed' because of hitting the rss limit. But interestingly mysqld continued to run. I also got a sync timeout on vserver xxx stop, and the message that the context was still running unexpectedly... But the context was gone when i looked at /proc/virtual.... 1118017164 M * rs so you didn't catch the issue 1118017184 M * rs try to fill 3/4 of the memory with mysql and the rest with several other process maybe 1118017209 M * rs like spamd and apache for instance 1118017301 M * Doener` that's what i had... ~25M RSS mysqld, a few apache processes etc... probably because i'm on UP 1118017332 M * Doener` but the fact that the kill of mysqld did not suceed makes me wonder... 1118017445 M * rs yeah could be related 1118017458 M * rs how did you restart the vserver ? 1118017608 M * Doener` vserver xxx stop... 1118017701 M * rs could you try with restart ? 1118017712 M * rs hmm stop should work 1118017719 M * rs it's what my soft do 1118017883 M * Doener` tried with reboot from the inside... now it's hanging at mysql shutdown somehow 1118018235 M * Doener` no success with stop/restart/reboot/reboot -f 1118018251 M * Doener` checking the failing kill now... 1118018595 M * rs interesting :) 1118018780 M * Doener` hm, there's a single location where the text appearing in the syslog appears in the kernel source, and there's no way for a user-space process to survive that AFAICT 1118019194 N * Bertl_oO Bertl 1118019203 M * Bertl back now ... 1118019211 M * Bertl evening rs! 1118019217 M * rs evening bertl! 1118019230 M * rs welcome back 1118019278 M * Doener` evening Bertl 1118019290 M * DaCa hi all 1118019297 M * Doener` hm, the mysqld is really strange... 1118019317 M * Bertl hey Doener`! DaCa! 1118019362 M * Doener` the thread that caused the rss limit to be hit is kind of dead, all threads are in Sl state, i can still connect to the mysqld... 1118019378 M * Doener` | 5 | root | localhost | test | Killed | 353 | *** DEAD *** | create index ads on a(a) | 1118019392 M * Doener` that's from "show processlist"... 1118019564 M * Bertl mysqld refusing to be OOM killed? 1118019572 M * Doener` yep 1118019574 M * Doener` hm, vkill just killed the wrong process... *lol* 1118019627 M * Doener` (on 2.0-pre4, not rc3, i.e. without my patch) 1118019644 J * eXplasm2 explasm@p549FF561.dip.t-dialin.net 1118019697 M * Doener` i did: vkill --xid 100 -s 15 - 11925 (11925 being the pid of mysqld) and sshd got killed... same command again and mysqld got killed... 1118019732 M * Doener` i have absolutely no clue how this could happen... 1118019743 M * Bertl well, why don't you use your patch? ;) 1118019764 M * Doener` hey, if it works, how should i tryhm, vkill just killed the wrong process... *lol* 1118019770 M * Doener` eek, wrong key 1118019821 M * Doener` if the patch works, i wouldn't be able to reproduce the issue (well, maybe i'm unable anyways cause of UP ;) 1118019955 M * rs the rc3 contains the doener patch ? 1118019961 M * Doener` yep 1118020083 Q * eXplasm Ping timeout: 480 seconds 1118020553 M * Bertl well, it doesn't look good for 2.0 util-vserver .... 1118020593 M * Bertl we probably have to hack into the userspace tools once again 1118020615 M * Bertl (or add some more 'legacy' stuff to the kernel) 1118020638 M * Doener` not enough time on enrico's side to do development? 1118020662 M * Bertl well, actually I don't know, he doesn't talk to me anymore ... 1118020696 M * Bertl maybe he is just too busy, maybe he doesn't like to work on it, I really have no idea ... 1118020724 M * Bertl it seems that this is somehow characteristic for this project :( 1118020744 M * Doener` hm? 1118020789 M * Bertl well, it was similar with jack, didn't answer anymore, didn't make releases ... 1118020861 Q * Aiken Quit: Leaving 1118020869 M * shuri :( 1118020871 M * Bertl maybe he is doing a lot of vserver related stuff, I just don't know 1118020889 M * Bertl the last 'release' of the alpha util-vserver is more than a month ago 1118020906 J * Aiken ~james@tooax6-071.dialup.optusnet.com.au 1118020907 M * Bertl the bug reports on savannah do not change ... 1118021458 M * Doener` that actually makes me wonder, as he requested them to be filed there... well, maybe he's just out of time, we can only guess 1118021482 M * Bertl yup 1118021760 M * rs bed time, good night 1118021771 M * Bertl yeah, good idea ... 1118021772 M * Doener` good night rs 1118021775 M * Doener` yep, me too 1118021785 Q * rs Quit: rs 1118021785 M * Doener` good night Bertl 1118021786 M * Bertl off to bed we go ... cya all tomorrow ;) 1118021794 N * Doener` Doener_zZz 1118021795 N * Bertl Bertl_zZ 1118022365 Q * Aiken Remote host closed the connection 1118022614 J * Aiken ~james@tooax6-071.dialup.optusnet.com.au 1118026303 Q * ben_ Ping timeout: 480 seconds 1118031525 Q * eXplasm2 Remote host closed the connection 1118031689 J * eXplasm explasm@p549FF561.dip.t-dialin.net 1118033231 Q * shuri Read error: Connection reset by peer 1118038770 M * eyck hmm, so maybe let's rewrite utils in pascal this time? ;) 1118039208 J * maharaja maharaja@ipax.at 1118042435 J * erwan_ho ~erwan@konilope.dyndns.org 1118043280 Q * erwan_ho Remote host closed the connection 1118043621 J * DaPhreak_ ~phreak@lms.rz.uni-greifswald.de 1118043984 Q * DaPhreak Ping timeout: 480 seconds 1118043990 J * prae ~prae@ezoffice.mandriva.com 1118045815 Q * Aiken Ping timeout: 480 seconds 1118046243 J * gregster ~gregor@greart.de 1118046250 M * gregster hi 1118046368 J * IceTi 125@mw635449.rz.ruhr-uni-bochum.de 1118047345 J * DaPhreak ~phreak@lms.rz.uni-greifswald.de 1118047367 J * rs ~rs@staff.lycos.fr 1118047585 Q * DaPhreak_ Ping timeout: 480 seconds 1118048252 Q * berni jupiter.oftc.net plasma.oftc.net 1118048252 Q * aba jupiter.oftc.net plasma.oftc.net 1118049915 J * berni ~berni@svr01.mucip.net 1118049941 J * aba ~aba@eos.turmzimmer.net 1118050749 Q * aba uranium.oftc.net plasma.oftc.net 1118050749 Q * berni uranium.oftc.net plasma.oftc.net 1118050850 J * aba ~aba@eos.turmzimmer.net 1118050850 J * berni ~berni@svr01.mucip.net 1118053506 Q * rs Quit: rs 1118057407 J * rs ~rs@staff.lycos.fr 1118058078 Q * IceTi Quit: get satisfied! • :: ««« (Gamers.IRC) »»» www.gamersirc.net :: 1118058565 Q * cryo Remote host closed the connection 1118060450 J * cryo ~say@212.86.243.154 1118060768 J * erwan_taf ~erwan@81.80.43.67 1118061023 N * Bertl_zZ Bertl 1118061037 M * Bertl morning folks! 1118061074 J * robig ~Miranda@141.32.5.232 1118061079 M * robig hi 1118061123 M * Bertl hey robig! 1118061237 M * robig where can i see what vserver version i use? 1118061257 M * Bertl easiest way is probably the testme.sh script 1118061264 M * Bertl sec 1118061281 M * Bertl http://vserver.13thfloor.at/Stuff/SCRIPT/testme.sh 1118061369 M * robig what does it do? 1118061382 M * Bertl check basic vserver functionality and report versions 1118061500 M * robig 2.6 kernel releases are still all dev_releases? 1118061723 M * Bertl do you mean mainline 2.6.x.y or the vserver patches? 1118061793 M * robig the script says Linux 2.6.9-vs1.9.3 i686/0.30.203/0.30.203 [Ea]. 1118061793 M * robig quota in the vserver schold be supported, right? 1118061842 M * Bertl it is, but it is not secure ... 1118062036 M * robig that means? 1118062109 M * robig what do you prefer? 1118062329 M * Bertl you need to provide a device for quota ioctl 1118062349 M * Bertl with 1.9.3 the vroot device wasn't available IIRC 1118062357 M * Bertl (or was it?) 1118062431 M * Bertl and without the vroot device, your only option for quota control (if you are really meaning quota) is to provide the real device node, which is not secure 1118062526 M * Bertl 2.0-rc* contains the vroot device and allows a secure setup 1118062555 M * robig i have hrsetup utility 1118062580 M * Bertl hrsetup? 1118062588 M * robig needs a new version a new patched kernel too? 1118062599 M * robig vrsetup :) 1118062608 M * Bertl it's the kernel which matters ;) 1118062618 M * Bertl does your vrsetup work with 1.9.3? 1118062636 M * Bertl i.e. are you able to configure vroot devices? 1118062720 M * robig how can i check this out? 1118062822 M * robig ive to go now... cu later :) 1118062833 M * Bertl k (cat /proc/devices) 1118062845 P * robig 1118062945 J * miller7 ~none@adsl16-49static.access.acn.gr 1118062953 M * miller7 hello 1118062972 M * miller7 Bertl, are you there? 1118063002 M * Bertl yup, greetings! 1118063011 M * miller7 :-) 1118063015 M * miller7 Can I ask you something? 1118063022 M * Bertl I guess so ;) 1118063024 M * miller7 I moved a vserver OS from one box to another 1118063030 Q * mcp Read error: Connection reset by peer 1118063031 M * miller7 I have some problems 1118063036 M * miller7 I try to chxid the files 1118063037 M * miller7 but I get 1118063043 M * miller7 Inappropriate ioctl for device 1118063052 M * miller7 do you know what this might be? 1118063053 J * mcp ~hightower@wolk-project.de 1118063067 M * Bertl miller7: probably no tagxid support on that mount? 1118063077 M * miller7 Hmm... I don't know... 1118063092 M * miller7 if the previous machine had it but this one not, will it affect the guest OS? 1118063098 M * miller7 for example not being able to write? 1118063141 M * Bertl depends on how you did copy/move the guest 1118063147 M * miller7 tar 1118063157 M * miller7 tar xf and tar cf 1118063173 M * Bertl from an xid aware kernel or not? 1118063182 M * miller7 I am not sure 1118063184 M * miller7 could be 1118063225 M * miller7 is there any quick fix I can do? 1118063232 M * Bertl well, assumed that you used uid/gid mangling xid on the original host (default) then doing the tar from a non xid aware kernel will give 'funny' ids ;) 1118063254 M * Bertl you should check for that, i.e. uid/gid above 65535 1118063283 M * Bertl except for that, it would be good to state the issue you encounter ;) 1118063283 M * miller7 how do I do that? 1118063299 M * miller7 I'm in the middle of crisis right now because the old server is dead 1118063300 M * miller7 :) 1118063349 M * Bertl ah, so you copied it on the new server (the tar) 1118063358 M * miller7 Yes 1118063365 M * Bertl and this server uses what kernel? 1118063381 M * miller7 2.4.27-1.29rc2 1118063394 M * Bertl the new one? 1118063397 M * miller7 Yes 1118063411 M * miller7 it's not a new installation, it's just a backup box 1118063455 M * Bertl well, probably the best option is to mount both partitions with tagxid/tagctx and do the copy again 1118063511 M * miller7 that's not possible since the old server is dead 1118063519 M * miller7 I have the tar only 1118063530 M * miller7 anything I can do with the tar alone? 1118063614 M * Bertl well, trickier, but possible ... 1118063629 M * Bertl first unmount the partition, and remount without tagxid 1118063772 M * miller7 where can I see the tagxid? I can't find such a thing in /etc/fstab 1118063796 M * Bertl the vservers are on a separate partition, I hope? 1118063806 M * miller7 not on the backup one 1118063808 M * miller7 :( 1118063814 M * Bertl then you lost :/ 1118063822 M * miller7 ok :) 1118063845 M * Bertl I would suggest: 1118063858 M * Bertl - setup a new backup system with recent 2.4 kernels 1118063868 M * Bertl - make the /vservers a separate partition 1118063877 M * Bertl - mount it without tagxid 1118063882 M * Bertl - restore the tar(s) 1118063892 M * Bertl - unmount it, mount it again with tagxid 1118063910 M * Bertl - check that the funny uid/gids are gone and xid tagging is present 1118063919 M * Bertl - start the guest and be happy ;) 1118063932 M * miller7 I don't see any funny uid/gid 1118063944 M * miller7 I see !!ERR!! 1118063953 M * miller7 ./lsxid /vservers/vs1 1118063959 M * Bertl yes, because the partition is not xid mounted 1118063963 M * miller7 ah ok 1118063995 M * miller7 got u now 1118064044 M * miller7 This is a very strange thing though, isn't it? 1118064061 M * Bertl no, not really ;) 1118064069 M * miller7 :-) 1118064098 M * miller7 Can you technically explain the error? I mean, why can't I reset the flags somehow in the files or something? 1118064165 M * Bertl because the 'flags' aren't there, if the partition is not mounted with tagxid 1118064196 M * miller7 how come tar puts the flags then? or it does not? 1118064224 M * Bertl no, tar doesn't look at the xid information 1118064228 M * miller7 ic 1118064241 M * miller7 And why the guest OS does not allow file creation then? 1118064244 M * Bertl but, as your filesystem was not mount tagxid (when making the tar) it did copy the uid/gid 1118064246 M * miller7 that's my problem I think 1118064274 M * Bertl and uid/gid are wrong now, because they contain the xid (in a mangled version) 1118064284 M * miller7 ic 1118064296 M * miller7 if I reset /tmp uid/gid? 1118064301 M * miller7 will it work? 1118064323 M * Bertl probably not, many files will be 'wrong' 1118064332 M * Bertl different groups, different users ... 1118064341 M * miller7 ic 1118064530 Q * albeiro Ping timeout: 480 seconds 1118064547 J * albeiro ~albeiro@procyon.romke.net 1118064704 M * miller7 thank you very much Bert! 1118064718 N * Doener_zZz Doener 1118064729 M * Doener morning 1118064817 P * miller7 1118064824 M * Bertl Doener: morning! 1118065786 Q * MattAR Ping timeout: 480 seconds 1118065856 Q * rs Quit: rs 1118067021 J * rs ~rs@staff.lycos.fr 1118067918 P * erwan_taf Leaving 1118068458 M * DaPhreak lo Bertl, Doener, rs ... 1118068465 M * Doener hey DaPhreak 1118068488 M * Bertl morning DaPhreak! 1118068531 M * DaPhreak Bertl/Doener: got a question about the xid-tagging stuff .. is there a way to _change_ the xid of a file !? well i basically have problems exchanging files (currently it's only tar-archives) between host -> vps 1118068556 M * DaPhreak the vps simply doesn't sees the file ... which smells strong like a xid-related problem 1118068609 M * DaPhreak i tried already the lsxid .. but only get !!ERR!! 1118068653 M * Bertl means that xid tagging is disabled 1118068888 M * DaPhreak Bertl: any other ideas (since it _is_ enabled or does xfs again have problems with the xid-tagging?) 1118068899 M * DaPhreak see: http://phpfi.com/64775 1118068939 M * Bertl cat /proc/mounts on the host? 1118069028 M * DaPhreak or is noatime the devil ? 1118069095 M * Bertl you are missing the tagxid mount option 1118069138 M * DaPhreak and again i'm getting enlighted ;) 1118069181 M * DaPhreak thanks a lot Bertl :) 1118069193 M * Bertl you're welcome! 1118069199 M * Doener Bertl: does the VXF_FORK_RSS check in copy_process() make sense for CLONE_THREAD or !CLONE_VM? 1118069233 M * Bertl probably not ... 1118069250 M * Bertl but it only strikes when already over limit 1118069261 M * Bertl so blocking any fork attempt sounds reasonable, no? 1118069262 M * DaPhreak ah by the way: -rc3(grsec) runs rock solid since nearly 5 days .. 1118069289 M * Bertl good to hear ... 1118069441 M * Doener Bertl: does it? it checks if it would be possible to increase the rss value by the process' current rss value... 1118069487 M * Doener say limit is 1000, usage is 800, process A uses 300. now process A does a clone with CLONE_THREAD, this would fail if VXF_FORK_RSS is set, right? 1118069666 M * Bertl hmm, right ... would be better to have two checks there 1118069681 M * Bertl one for the fork itself (which checks if over limit) 1118069697 M * Bertl and another one, which only happens on !CLONE_VM 1118069716 M * Bertl (and checks for the total amount) 1118069762 J * simon_ ~simon@vodsl-2029.vo.lu 1118069765 M * simon_ hi 1118069800 M * simon_ my vserver host crashed today, and since, when i mount the vserver partitions (ext3), mount reports that they are ext2, allthough they were ext3 1118069824 M * simon_ checking with tune2fs -l, still says the fs has a journal...but i can't mount it as ext3 1118069826 M * simon_ any ideas? 1118069849 M * Bertl probably some ext3 bug 1118069856 M * Bertl which kernel version is that? 1118069916 M * simon_ 2.6.11.5-vs1.9.5 1118069973 M * Bertl you should e2fsck the partition, then remove the journal inode 1118069978 M * Bertl and recreate it again 1118069994 M * simon_ i already did all kinda checks... several times... 1118070008 M * simon_ how can i remove the journal ? couldn't find that opt. 1118070080 M * Bertl tune2fs -O ^has_journal /dev/hdXX 1118070118 M * Doener Bertl: in your last message to ml you mention util-vserver 0.30.208. Just a typo or did i miss a new location for alpha tools? 1118070138 M * Doener or maybe some prediction of the future? ;) 1118070145 M * Bertl no, I just copied enricos method of future references ... 1118070155 M * Doener ok 1118070176 M * Bertl I know that I would not suggest to use 0.30.207 for production on 2.0* 1118070191 M * Bertl (too many unfixed issues) 1118070345 M * simon_ i recreated the journal..but still am unable to mount it with -t ext3 .. :-/ 1118070359 M * simon_ should i assume that it _IS_ ext3, and ignore that... or ? 1118070400 M * Bertl did you fsck it afterwards? 1118070408 M * simon_ no 1118070410 M * Bertl (after removing the journal) 1118070413 M * simon_ no 1118070418 M * Bertl please do so ... 1118070422 M * simon_ i'll redo everything then... 1118070477 M * Bertl also check that you do not have a strange option in your fstab 1118070498 M * Bertl some 'mount' commands (debian iirc) are trying to be too smart ;) 1118070514 M * simon_ only -o grpquota,usrquota 1118070521 M * simon_ (using gentoo) 1118070537 M * Bertl do you ahve quota support enabled? 1118070550 M * Bertl also check with dmesg for kernel output 1118070553 M * simon_ in the kernel? yes 1118070589 M * simon_ e2fsck keeps reporting errors, on mount-unmount... 1118070611 M * Bertl mount-umount? 1118070617 M * simon_ yes 1118070622 M * Bertl what's that? 1118070643 M * Bertl and how is e2fsck involved with the mount/umount? 1118070661 M * simon_ i do a e2fsck... it cleans the fs 1118070681 M * simon_ then i mount it... unmount it... run e2fsck again, and it reports again errors 1118070693 M * simon_ strange 1118070697 M * Bertl well, then you have disk corruption here ... 1118070799 M * simon_ mount -t ext3 -o grpquota,usrquota /dev/main/server /vservers/server 1118070812 M * simon_ mount: /dev/main/server already mounted or /vservers/server busy 1118070829 M * simon_ without "-t ext3" it works... 1118070844 M * simon_ i removed the journal... check force... created a new one..check force 1118070846 M * simon_ hmm 1118070847 M * Bertl hum, sure you haven't mounted it somewhere else too? 1118070854 M * simon_ yes, sure 1118070859 M * Bertl ext2 can be mounted more than one time ... 1118070877 M * simon_ mount says it's not mounted anywhere else.. 1118070886 M * simon_ where else can i check? 1118070893 M * Bertl /proc/mounts 1118070901 M * Bertl mount doesn't know anything ... 1118070910 M * simon_ /proc/mounts... same 1118070915 M * Bertl (it just outputs the mtab) 1118070926 M * simon_ i know... the link is correct though.. :-) 1118070939 M * Bertl ugh, you use a linked mtab? 1118070962 M * simon_ hdd corruption is improbable... it's a soft raid1 and the kernel doesn't report anything strange.. 1118070986 M * simon_ thought so, but apparently not :-)... anyway.. 1118070998 M * Bertl well, mounting and unmounting does not corrupt ext 'usually' 1118071017 J * ruuth VooDoo@topas.informatik.uni-ulm.de 1118071023 M * Bertl welcome ruuth! 1118071039 M * ruuth Hi Bertl ;) 1118071087 M * simon_ i had a kernel panic this morning.... had to power cycle.. 1118071098 M * Bertl did you record it? 1118071118 M * simon_ wasn't able to :-( 1118071126 M * simon_ i found this in dmesg: EXT2-fs warning (device dm-5): ext2_fill_super: mounting ext3 filesystem as ext2 1118071148 M * simon_ an before: EXT2-fs: dm-5: couldn't mount because of unsupported optional features (4). 1118071149 M * ruuth Question: I want to make my guest-vservers as small as possible ... what do I have to do? ... can I share the portage-tree of the root-server with the guests (gentoo) ... if YES then HOW :)? Other suggetions? 1118071169 M * TheSeer mount -o bind 1118071174 M * TheSeer for shareing 1118071205 M * Bertl simon_: it might sound stupid, but did you verify that you booted the correct kernel? 1118071234 M * Bertl ruuth: unification ... 1118071272 M * DaPhreak Bertl: not so good on gentoo (if you use different compile-settings per vserver) 1118071309 M * simon_ i checked my kernel config for ext3 support, just to be sure :-)...it does have ext3 support ... 1118071350 M * Bertl DaPhreak: well, in the worst case it's the full size (binaries, libs) for each of them 1118071364 M * Bertl simon_: ext3 quota too? 1118071404 M * simon_ for ext3 there is only global quota support.. yes it's enabled 1118071432 M * ruuth Bertl: more info please ... I'm still a newbie to vservers :| 1118071444 M * simon_ (mounting wihtou the quota options doesn't change anything) 1118071546 M * Bertl http://www.freesoftwaremagazine.com/free_issues/issue_05/focus-linux_vserver/ 1118071551 M * Bertl (just stumbled over it ;) 1118071555 M * DaPhreak ruuth: well, you have to imagine that unification is to link matching things like executeables and libraries of different vservers together so that (lets say you run 10 vservers) you'll only need the files 1 time and not 10 .. 1118071611 M * ruuth DaPhreak: I think I got taht - but how do I do it exactly - "hands on" ... is there a HowTo? 1118071635 M * Bertl latest tools provide an utility called vhashify 1118071665 M * simon_ bertl anything else you think could be the problem, or something i could check ? 1118071670 M * ruuth DaPhreak: Which OS ist best with vservers? What's the difference between vserver and (for example) xen? 1118071683 M * simon_ Bertl, i doubt, upgrading the kernel would change anything... 1118071701 M * DaPhreak ruuth: read the focus-linux_vserver paper Bertl pointed on .. 1118071704 M * Bertl ruuth: define vservers? 1118071704 M * DaPhreak ;) 1118071734 M * Bertl simon_: well, no idea but I would try ... 1118071750 M * Bertl simon_: in general, I would backup as much as possible right now ... 1118071834 M * simon_ :-/... hmmpf k 1118071866 M * Bertl your system doesn't look very healthy right now ... 1118071993 M * simon_ (kernel doesn't report anything unusual regarding the disks though) 1118072003 M * simon_ thnx for your help Bertl ! 1118072057 M * Bertl you're welcome! 1118072093 M * ruuth Bertl: Host-System and Guests ... at the moment I use Gentoo for both .. but I have seen vservers on debian ... and it's much faster to setup there ... 1118072119 M * ruuth Bertl: Is that what you asked for ;)? 1118072137 M * Bertl well, no, but it answers my question pretty well ;) 1118072177 M * Bertl ruuth: you should use s/OS/Distro/ otherwise folks might answer, Linux of Course! 1118072252 M * ruuth Bertl: Ok ... good point ;) ... DISTRO DISTRO DISTRO ... I think I got it now *smile* 1118072490 M * Bertl it really depends ... 1118072530 M * Bertl I'm happy with my distro of choice, debian works fine (if you don't use the outdated packages), gentoo seems a little more complicated as there is no real tool support yet ... 1118072560 M * Bertl fedora is the devel platform for the tools, I provide mandrake packages ... 1118072595 M * Bertl some folks do slackware/ubuntu/whatever stuff and it seems to work reasonably well too 1118072639 M * Bertl Doener: what do you think, should we make the startup/shutdown helper conditional on non legacy? and maybe make them async for this case too? 1118072779 M * Doener let me see if i got that right: no startup/shutdown helper for legacy vservers, and the other vshelper calls are async for legacy vservers, while non-legacy vservers have all helpers and those are run sync, right? 1118072822 M * Bertl yep 1118072835 M * Bertl should at least reduce the issues observed for now 1118072859 M * Bertl and I'm not sure if we need the startup/shutdown at all for ngnet 1118072891 M * Bertl I lean more to using a netup/netdown for that (which is tied to a network context startup/shutdown) 1118072979 M * Doener what's the difference between sync/async? 1118073003 M * Bertl with async, the kernel returns the syscall pretty soon 1118073015 M * Bertl often before the helper did finish 1118073039 M * Bertl with sync, the reboot syscall is 'hanging' until the helper returns 1118073151 M * Doener hm ok, i should learn to ask the right questions ;) what's the reason for a sync call of the vshelper? 1118073251 M * Bertl well, it allows 'synchronization' ;) and besides that, I'm not sure that the async isn't problematic regarding the passed arguments 1118073293 M * maharaja re 1118073301 Q * simon_ Quit: Leaving 1118073312 M * Bertl wb maja! 1118073315 M * maharaja bertl: do you know anything about an nmbd patch to work inside the vservers? 1118073328 M * maharaja bertl: nmbd does not find any interfaces 1118073340 M * maharaja thou samba is working without any problem 1118073396 M * maharaja mhm 1118073416 M * maharaja maybe i should specify the netmask *trying* 1118073431 M * Doener Bertl: i have no objections against the change 1118073463 M * Doener Bertl: btw, do you know how the comment on the top of the fsf article is meant? "(will be free on the 15th of July, 2005 )" 1118073465 M * Bertl good, any pending aptches from your side? (e.g. fork stuff) 1118073483 M * Doener no, nothing here 1118073514 M * Bertl ad free; no, but sounds funny ;) 1118073519 M * Doener i'm fighting with the code to figure out why mysqld does not die after oom kill 1118073541 M * maharaja bertl: do you consider the following scenario as a bug of vserver or of samba: 1118073559 M * Bertl (samba of course ;) 1118073564 J * Doener` ~doener@p54875F01.dip.t-dialin.net 1118073571 M * maharaja bertl: you have to specify an interface in smb.conf (e.g. interfaces = 10.1.1.54/24) 1118073599 M * maharaja bertl: this example is working. omitting the netmask (/24) causes a malfunction in nmbd 1118073629 M * Bertl malfunction means? 1118073633 M * maharaja bertl: like: "can't determine netmask for x.x.x.x" 1118073651 M * maharaja and then it states that it is waiting for interfaces to appear 1118073657 M * Bertl probably it scans the interfaces and gets confused by unusual looking interfaces 1118073674 M * maharaja it does, as far as i know 1118073687 M * maharaja kernel is 2.6.11.6-grsec-vs1.9.5 1118073689 M * Bertl so as I said, a bug in samba ;) 1118073713 M * maharaja ok, thnx 1118073718 M * maharaja ill write a mail to the samba mailinglist 1118073796 M * Bertl maybe you should try to figure out _what_ causes the scan to fail first, but that's up to you ... 1118073861 M * DaPhreak hmm .. maharaja samba needs two interfaces .. 1118073870 M * maharaja the first error is "no local interfaces found" 1118073879 M * maharaja uhm - "no network interfaces found" 1118073885 M * maharaja well - if i get time ill do that 1118073894 M * maharaja but right now i have to fix something for tomorrow 1118073897 Q * berni Remote host closed the connection 1118073900 M * DaPhreak first the ip it listens on .. and at second the broadcast of the subnet that ip is on .. 1118073926 M * DaPhreak s/it on/its on/ 1118073929 M * maharaja daphreak: you can do that by using the /xxx notation 1118073932 M * maharaja like 10.1.1.54/24 1118073955 M * DaPhreak well maharaja i'm talking of your vserver config, not the samba one ;) 1118073960 M * maharaja if you "forget" to specify a subnetmask, samba tries to guess your mask by askin the system 1118073973 M * maharaja mhm 1118073976 M * maharaja ? 1118073977 M * maharaja :) 1118073984 M * DaPhreak yupp .. second .. 1118074002 Q * Doener Ping timeout: 480 seconds 1118074056 M * DaPhreak at first, samba needs some _special_ caps .. (NET_BROADCAST, NET_RAW and NET_ADMIN) (i know unsecure ..) 1118074145 M * Bertl Doener`: did you find some time to look at the broken out patches? 1118074147 M * DaPhreak then it needs two interfaces (0 and 1 here) .. to the first, you assign the ip where the samba should listen on, and the second one, you put in the broadcast as _interface ip_ but also a nodev-stub file .. 1118074200 J * simon_ ~simon@vodsl-2029.vo.lu 1118074237 M * DaPhreak Bertl: there you have the mess of enrico's future references ;) (ml...) 1118074241 M * simon_ are there any special kernel-patches i should apply for amd64 (kernel 2.6.11.11 vs2.0-rc3) ? 1118074271 M * DaPhreak should work out of the box ... no idea since i don't own one ;) 1118074281 M * simon_ k 1118074346 M * Bertl DaPhreak: nah, piece of cake! 1118074365 M * Bertl simon_: no, should work fine 1118074424 M * maharaja daphreak: i do not understand that :) is that vserver related or samba related (2 ips) 1118074486 M * DaPhreak maharaja: well more samba than vserver ... samba has some (or most) problems if it (or he) doesn't sees the broadcast .. 1118074489 M * Bertl Doener`: is your 2.6.11-rc5 patch up to date? 1118074963 J * Bertl_ ~herbert@janus.mc.tuwien.ac.at 1118075082 Q * Bertl Ping timeout: 480 seconds 1118075089 N * Bertl_ Bertl 1118075214 M * Doener` it's at rc2... should be up2date in a few minutes ;) 1118075350 Q * prae Quit: Client exiting 1118075388 Q * rs Quit: rs 1118075396 M * Doener` http://www.13thfloor.at/~doener/vserver/patches/patch-2.6.12-rc5-vs2.0-rc3.diff 1118075461 M * maharaja daphreak: as you seem pretty experienced with samba, did you ever try samba + ldap + pdc + domain admins? 1118075519 M * DaPhreak maharaja: well i messed around with samba and vserver (as far as Bertl showed pointed me to) ;) never tried that combination 1118075540 M * maharaja i c 1118075568 M * maharaja mhm - seems likei ts working right now 1118075568 P * simon_ Leaving 1118075570 M * maharaja strange.... :) 1118075598 J * sim0n ~simon@vodsl-2029.vo.lu 1118075678 M * badiane hi guys 1118075695 M * badiane is it possible to run hostapd in the vserver 1118075758 M * DaPhreak hostapd as in "HostAP wireless daemon" ? 1118075798 M * badiane yep 1118075831 M * badiane the modules are loaded upon boot in the host 1118075846 M * Bertl what does it do? 1118075851 M * badiane I wanted to run the daemon in the vserver 1118075862 M * badiane it acts like an access point 1118075892 N * Loki|muh_ Loki|muh 1118075894 M * badiane allows for radius auth, eap, and all sorts of other things 1118075904 M * DaPhreak hostapd - user space daemon for access points, including, e.g., IEEE 802.1X/WPA Authenticator for number of Linux and BSD drivers, integrated EAP authenticator, and RADIUS authentication server 1118075904 J * shuri sjnesjd@64.235.209.226 1118075909 M * shuri Ola 1118075917 M * DaPhreak lo shuri 1118075933 M * sim0n does the latest util-vserver fix the dietlib-utilvserver-amd64 issues ? (maybe stupid question..) 1118075976 M * badiane yep that's the one 1118075999 M * badiane I set the debug options 1118076006 M * Bertl sim0n: if you have a recent dietlibc yes 1118076006 M * badiane hold on 1118076067 M * sim0n 0.28 ... i'll try to get a newer one 1118076107 M * badiane I can't get any debugging informaton 1118076130 M * badiane Starting advanced IEEE 802.11 management: hostapd...failed. 1118076293 M * Doener` brb 1118076294 Q * Doener` Quit: Leaving 1118076838 J * Doener ~doener@p54875F01.dip.t-dialin.net 1118076854 M * Doener Linux doener 2.6.12-rc5-vs2.0-rc3 #1 Mon Jun 6 18:41:58 CEST 2005 i686 AMD Athlon(tm) XP 2600+ AuthenticAMD GNU/Linux 1118076856 M * Doener :) 1118076872 M * Doener ARGH! 1118076879 M * TheSeer argh? 1118076886 M * Doener .12-rc6 1118076893 M * Bertl *G* 1118076894 M * TheSeer hehe 1118076915 M * TheSeer gna.. i start to hate NFS... 1118076916 M * Doener i checked just before updating the patch 1118076931 M * Doener murphy, some day I'll getcha! 1118077101 M * Bertl TheSeer: hmm? 1118077596 M * sim0n i dunno why, but dietlibc 0.29 won't build so libraries... ? 1118077615 M * sim0n i just run make && make install 1118077632 M * Bertl well, we probably don't need them anyway ;) 1118077652 M * Doener sim0n: on x86_64? 1118077690 M * sim0n Doener, yes 1118077781 M * Doener last time i checked, debian didn't ship so libs for x86_64, so i guess that's ok (not building those libs). util-vserver tools are linked static anyways AFAIK 1118077804 M * sim0n ok.... 1118077810 M * sim0n found something in the FAQ 1118077812 M * sim0n Q: Where are the shared libaries? make install didn't install them! 1118077812 M * sim0n A: You have to explicitly build them with "make dyn". Since they are 1118077812 M * sim0n experimental and only supported on a small subset of the platforms, 1118077812 M * sim0n that is not default. Also, I recommend you only use shared libraries 1118077812 M * sim0n if you really know what you are doing. For example, you can't just 1118077813 M * sim0n use your system shared libraries, because they have a dependency on 1118077815 M * sim0n glibc in them, so the program will crash. And you have to explicitly 1118077817 M * sim0n compile the code with -fPIC or -fpic. You can then use them by 1118077821 M * sim0n substituting "diet-dyn" for "diet" on the command line. 1118077823 M * sim0n sorry for spamming... :-) 1118077895 M * sim0n i was wondering about this, because in 0.28, it did make those so's automatically 1118078052 J * rs ~rs@mon75-8-82-230-181-39.fbx.proxad.net 1118078053 M * TheSeer Bertl: i do backup via NFS to a backup server... and it keeps timing out for whatever f****** reason ;) 1118078081 M * TheSeer Jun 6 18:47:38 brain kernel: nfs: server toxic.local.salesemotion.net not responding, still trying 1118078093 M * Bertl TheSeer: wrong mount options ;) 1118078103 M * TheSeer Jun 6 18:48:32 brain kernel: nfs: server toxic.local.salesemotion.net OK 1118078117 M * TheSeer yeah.. figured that one ;) messing with the options all day ;> 1118078130 M * Bertl the proper mount options are: tcp,vers=3,rsize=32768,wsize=32768,async,rw ;) 1118078150 M * TheSeer hmm.. i didn't use vers=3 .. 1118078155 M * TheSeer let's see.. 1118078157 M * Bertl if you prefer to be able to unmount it if the filer goes down, add hard,intr ;) 1118078168 M * TheSeer i have that ;> 1118078188 M * sim0n i'm getting some new errors when compiling util-vserver-0.30.207 against dietlibc-0.29 1118078196 M * sim0n /usr/include/asm/../asm-x86_64/pda.h:26: error: `CONFIG_X86_L1_CACHE_SHIFT' undeclared here (not in a function) 1118078196 M * sim0n /usr/include/asm/../asm-x86_64/pda.h:26: error: requested alignment is not a constant 1118078203 M * sim0n /usr/include/asm/../asm-x86_64/processor.h:65: error: requested alignment is not a constant 1118078203 M * sim0n /usr/include/asm/../asm-x86_64/processor.h:228: error: requested alignment is not a constant 1118078208 M * Bertl TheSeer: and do not send cards saying 'Thank You', just donate something *G* 1118078269 M * TheSeer :-P 1118078286 M * TheSeer toxic.local.salesemotion.net:/storage on /mnt/backup type nfs (rw,rsize=32768,wsize=32768,timeo=14,tcp,retry=720,hard,intr,nfsvers=3,addr=192.168.1.10) 1118078293 M * TheSeer let's see if that's more stable now... 1118078581 M * sim0n any idea what the problem might be ? 1118078631 M * Bertl sim0n: looks like you have broken headers 1118078650 M * Bertl (i.e. they look like headers from the kernel to me) 1118078666 M * Bertl (but they should be kernel headers from the glibc) 1118078798 Q * romke Quit: reboot 1118078830 M * pusling is there some kind of vnetstat ? 1118078864 M * brc bertl! 1118078882 M * Bertl brc! 1118078926 M * brc ltns 1118078937 M * brc i have a new doubt :) 1118078950 M * brc if tagxid is enabled, i should see it in /proc/mounts. is that right ? 1118078959 M * brc cause i enabled in a new machine, and it is not being shown there. 1118079045 M * sim0n Bertl, those are my kernel headers... dunno why, but util-vserver uses them 1118079118 M * sim0n Kernelheaders: /lib/modules/2.6.11.11-vs2.0-rc3/build/include 1118079129 M * sim0n (output from the config script) 1118079151 M * Bertl yes, but they seem to be at ... /usr/include/asm/ too (see messages ;) 1118079178 M * Bertl pusling: not that we know of. any ideas what it would do? 1118079210 M * pusling Bertl: something like enter all vserves and do a netstat and order it nicely ? 1118079229 M * sim0n Bertl, but it's normal that i have different headers there... those are the ones my glibc is linked against ... hmmm 1118079287 M * Doener Bertl: rss ulimit is -m, right? 1118079292 M * Bertl sim0n: yes indeed ;) 1118079318 M * Doener nvm, man bash says that it is -m 1118079330 M * Bertl yup, just checked ... 1118079344 M * sim0n Bertl, and i only get those errors when i link against dietlibc.... no problem when linking against glibc... 1118079389 M * brc Bertl: Must tagxid appear in /proc/mounts if it is enabled ? 1118079408 M * Bertl sim0n: probably your dietlibc isn't compatible with your distro? 1118079424 M * Bertl brc: on 2.6 kernels, yes! 1118079429 Q * albeiro Ping timeout: 481 seconds 1118079430 M * Doener hmmm.... 1118079447 M * Doener max memory size (kbytes, -m) 8192 1118079452 M * Doener 11639 doener 17 0 67064 64m 284 S 0.0 6.4 0:00.18 oom-test 1118079473 M * sim0n Bertl, it is with my distro... on my other x86 server, it works just fine. it rather is a x86_64 problem :-/ 1118079475 M * brc Bertl: weird i just enabled it but it seems not to be enabled 1118079476 M * Doener top says i got 64m rss... 1118079528 M * Bertl brc: how? 1118079551 M * Bertl sim0n: does your distro not provide a recent dietlibc? 1118079553 M * brc i've just done the same i did on the other machines. 1118079562 M * brc rebooted it, and it doesnt show tagxid on /proc/mounts 1118079563 M * Bertl brc: which is what? 1118079586 M * brc which is adding the rootflags=tagxid to lilo/grub 1118079589 M * sim0n Bertl, (using gentoo) i merged dietlibc 0.29... the most recent i could find... 1118079623 M * Bertl brc: ah, so you are ignoring my warnings ... never do it for the root fs ;) 1118079625 M * brc Bertl: the only difference from the other systems is that i have UDEV on this one. 1118079636 M * Doener in a vserver, with hard rss rlimit, the process gets killed 1118079642 M * brc Bertl: yes, remember we tried it like 1 year ago, and it was fine ? 1118079647 M * brc has anyone reported problems ? 1118079669 M * Bertl no, but probably because I keep scaring folks away from doing so ;) 1118079688 M * Bertl in your case, the rootflags probably apply to the initrd ;) 1118079707 M * Bertl (which in turn, happily ignores them for your real rootfs) 1118079768 J * berni ~berni@svr01.mucip.net 1118079776 M * brc what happily ? 1118079781 M * brc bertl, but i have tagxid on / running for a long time 1118079786 M * brc on 2 boxes 1118079853 M * Bertl welcome berni! 1118079855 M * brc almost 1 year 1118079861 M * brc so initrd is ignoring it 1118079868 M * brc do you think that i might have problems 1118079872 M * Bertl brc: well, doesn't change the fact that it is ignored right now? 1118079895 M * brc Bertl: it is not being ignored on the 2 other boxes 1118079897 M * brc just on this new one 1118079926 M * sim0n Bertl, how unsecure is it if i run some private vservers using util/glibc for now ? 1118079948 M * sim0n or do you have an idea what i could do to get the utils build correctly? 1118079990 M * Bertl brc: check the fstab on the system (i.e. add the tagxid entry to the root) and rebuild the initrd (but shht! don't tell anybody!) 1118080012 M * Bertl sim0n: what distro did you say? 1118080013 M * brc i am starting to get worried 1118080018 M * sim0n Bertl, gentoo 1118080027 M * Doener Bertl: got a testcase for the oom killer not working for multi-threaded processes in a vserver 1118080047 J * erwan_ho ~erwan@konilope.dyndns.org 1118080058 M * brc Bertl: i have tagxid enabled, and all thefiles to the vservers are chctxed to their context. but i am not using vdlimits YET. i am gonna start using it now, on this new box. 1118080063 M * Doener as the ulimit does not seem to work as i expect it, i can't tell whether the killer works on the host 1118080067 M * brc do you think that this is the momment where i should start getting real problema ? 1118080069 M * brc do you think that this is the momment where i should start getting real problems ? 1118080143 M * Bertl Doener: hmm, you can still allocate memory? 1118080161 M * Bertl Doener: I mean 'instantiated memory' 1118080186 M * Bertl brc: might be, maybe not ... you'll have to test 1118080213 M * Bertl sim0n: does gentoo provide a dietlibc >0.28 ? 1118080235 M * sim0n Bertl, no, it doesn't. but i did an ebuild for 0.29 1118080236 M * DaPhreak nope .. 1118080254 M * Doener after oom kill? no idea, only one thread allocates memory, and the other one is not killed... i don't think that's supposed to happen, is it? 1118080274 M * Doener what's 'instantiated memory'? 1118080304 M * Bertl Doener: well, that might be a mainline kernel bug (regarding OOM) but maybe it is intentional 1118080333 M * Bertl Doener: well, 'reserving' memory alone will just raise the VM/AS limit, not the RSS limits ... 1118080339 M * Bertl s/limit/usage/ 1118080360 M * Bertl sim0n: well, you could try my patches ontop of 0.28 ... 1118080374 M * Bertl they worked fine with mandrake and rh2 1118080380 M * Bertl s/rh2/fc2/ 1118080380 M * ruuth hi! when I want to start my vserver I get "Setting system clock ..." -> failed to set system colck to hardware clock [!!] ... Error Problem starting needed services .. bootmisc was not started ... syslog-ng vixie-cron either ... what have i done wrong? ... vserver 2.0rc3 on gentoo 1118080394 M * Doener i malloc() small chunks of memory, so that the rss value gets increased... in a vserver with hard rss rlimit, the process is oom killed, on the host, with ulimit -HS -m8192, it is not 1118080408 M * sim0n Bertl, yeah... where can i find them ? 1118080408 M * Bertl ruuth: your guest is executing stuff it is not supposed to execute 1118080414 M * DaPhreak ruuth: bug Hollow about this ;) 1118080448 M * Bertl sim0n: http://vserver.13thfloor.at/Stuff/DIETLIBC/ 1118080461 M * Bertl first apply the two patches, then replace the syscall.S 1118080467 M * sim0n thnx....i'll try them 1118080478 M * DaPhreak btw Bertl: seems like the make install rule of dietlibc is a bit screwed .. ?! or is it just me ? 1118080501 M * DaPhreak (its 0.29) 1118080508 M * ruuth Hollow: hi! when I want to start my vserver I get "Setting system clock ..." -> failed to set system colck to hardware clock [!!] ... Error Problem starting needed services .. bootmisc was not started ... syslog-ng vixie-cron either ... what have i done wrong? ... vserver 2.0rc3 on gentoo 1118080514 M * Bertl DaPhreak: well, didn't test 0.29 yet ... 1118080558 J * nayco ~nayco@lns-vlq-47-nan-82-252-254-83.adsl.proxad.net 1118080581 M * DaPhreak got only three files installed after make && make install: (diet, elftrunc, dnsd) 1118080586 M * nayco Hello, all ! 1118080627 M * Bertl hello one! 1118080651 M * nayco ;-) 1118080669 M * ruuth Bertl: Do you have any Idea? I followed the "manunal" (http://gentoo.home.xnull.de/doc/en/vshandbook/index.xml?part=2&chap=3) 1118080702 M * Bertl ruuth: you ahve to disable all services which are hardware related (or make them fail silently) 1118080715 M * TheSeer Bertl: Jun 6 19:32:45 brain kernel: nfs: server toxic.local.salesemotion.net not responding, still trying 1118080725 M * TheSeer Jun 6 19:47:12 brain kernel: nfs: server toxic.local.salesemotion.net OK 1118080735 M * TheSeer somehow that didn't fix it yet ;/ 1118080735 M * Bertl then your network/router is bad 1118080745 M * TheSeer well.. everything else works fine.. 1118080759 M * TheSeer im rather suppsecting the nfs-server box 1118080768 M * Bertl kernels? 1118080770 M * TheSeer or its noname nic / driver to be exact 1118080794 M * TheSeer 2.6.11-1.14_FC3 <- client 1118080801 M * Bertl urgh! 1118080814 M * daniel_hozac isn't 27 the most recent errata kernel? 1118080815 M * TheSeer 2.6.9-1.11_FC2 <- server 1118080836 M * TheSeer daniel_hozac: dunno.. could check though ;) 1118080975 M * sim0n ruuth, have you followed the instructions regarding baselayout ??... and did afterwards run etc-update ? 1118081001 M * ruuth sim0n: yes - I think so .. 1118081007 M * daniel_hozac my firewall suggests it is. 1118081014 M * sim0n ruuth, I, don't think so :-) 1118081018 M * TheSeer daniel_hozac: correct.. .27 is the latest.. 1118081024 M * sim0n ruuth, at least you forgot to run etc-update... 1118081041 M * sim0n ruuth, i did the exact same thing today too, and it works great here :-) 1118081059 M * sim0n i followed this manual though: http://dev.gentoo.org/~hollow/vserver/guide/ 1118081137 M * ruuth sim0n: wait a sec ... I'll have a look :) 1118081208 M * sim0n ruuth, i'm sure, if you follow that one, it will work ;-) 1118081209 M * DaPhreak well sim0n used the old one and ruuth used the new one ;) but the new one is still work in progress :) 1118081254 M * ruuth DaPhreak: ah! 1118081270 M * ruuth I used the "long howto" ... 1118081305 M * DaPhreak yupp .. hopefully the new baselayout/bash/stuff will go stable pretty soon :) 1118081305 M * ruuth DaPhreak: What should I do? 1118081318 M * sim0n DaPhreak, euh? dunno... i just know that it works... and the baselayout i used was recently updated... so it can't be that "old" :-) 1118081325 M * DaPhreak hmm ... do as sim0n said ;) use the old one .. that definitly works 1118081329 M * brc bertl 1118081333 M * brc still there ? 1118081334 M * Bertl hmm? 1118081391 M * sim0n oops... recently = 6.1.05 ... :-)... thought it was 6.5.05 or so... maybe not that recent... but anyway, it works 1118081541 M * brc Bertl: adding tagxid to fstab wont work 1118081547 M * brc did it now, and i just remember that it didnt work 1118081555 M * brc that's why i appended it on rootflags 1118081572 M * Bertl brc: did you rebuild the initrd? 1118081575 M * brc is thre something like an "append" for initrd ? 1118081584 M * brc cause there is no fstab on my initrd 1118081587 M * brc sorry doesnt no much about initrd 1118081587 M * brc :) 1118081642 M * ruuth sim0n: can I update the baselayout or something? 1118081687 M * ruuth sim0n: Help! *smile* 1118081724 M * sim0n ruuth, do exactly as described on that page... put the ebuild in the portage overlay etc... the update your baselayout (you should see a new flag: "vserver"), then run etc-update... and that's all 1118081779 M * ruuth sim0n: Ok ... thx ... tomorrow ;) ... now "CSI" :) 1118081797 M * sim0n ruuth, i'll paste the important parts in private... hold on a sec. 1118081845 M * ruuth sim0n: Many thanks! 1118081867 M * sim0n there you go... :-)... no problem! 1118081913 M * sim0n Bertl, should i replace the libcompat/syscall.S file with your file?... because i get this strange error: 1118081921 M * sim0n libcompat/syscall.S: Assembler messages: 1118081921 M * sim0n libcompat/syscall.S:93: Error: .err encountered 1118081921 M * Bertl yep 1118082007 M * sim0n uhm, what can i do about that error? 1118082021 M * Bertl did you replace the file? 1118082141 M * sim0n yes i did 1118082162 M * Bertl well, then your distro seems to be very different from the ones I checked it with ... 1118082177 M * Bertl probably it define threaded behaviour or so ... 1118082210 M * Bertl but the tools should not use this interface anyway 1118082214 M * Bertl so just remove the .err 1118082222 M * Bertl (and hope for the best ;) 1118082247 M * sim0n i use ntpl... maybe that's the prob? 1118082262 M * Bertl probably ... 1118082291 M * sim0n "hope for the best".... hmm... should everything work without crashing my system ?... i can't "hope", this machine has to keep running ;-P 1118082313 M * Bertl sim0n: well, then better get a working dietlibc for your distro/arch ;) 1118082325 M * Bertl i.e. bug your distro maintainers ... 1118082352 Q * rs Quit: rs 1118082358 M * sim0n ok, i will... thnx 1118082360 M * Bertl Doener: in my qemu test setup, I seem to be unable to trigger the reboot helper ... 1118082383 M * sim0n uhm... is it safe to run some private vservers with glibc til then ? 1118082402 M * Bertl modulo the resolver issues, probably yes 1118082449 M * Doener Bertl: hm, so reboot -f does nothing? 1118082529 M * Bertl nevermind, didn't set it properly in the sysfs :/ 1118082584 M * sim0n bertl, ? 1118082599 M * Bertl sim0n? 1118082613 M * DaPhreak hmmm is tagxid unsupported on xfs ?! 1118082621 M * Bertl should be ... 1118082622 M * DaPhreak Jun 6 20:39:03 quasimodo XFS: unknown mount option [tagxid]. 1118082630 M * sim0n should be safe?... 1118082635 M * sim0n hmm ok... thnx! 1118082645 M * DaPhreak running 2.6.11.11-vs2.0-rc3 1118082652 M * Bertl DaPhreak: hmm, sec checking ... 1118082663 M * DaPhreak yeah .. 1118082803 M * Bertl hmm, yeah seems that was lost ... 1118082883 M * Bertl brc: which kernel/filesystem for root on that machine? 1118083054 M * Bertl DaPhreak: if you can tell me where xfs hides the flag values nowadays ... I might be able to add it ;) 1118083091 M * DaPhreak heh, gonna do a search tommorow .. maybe i find something ;) 1118083128 M * Bertl ah, found it ... 1118083157 M * DaPhreak great :) 1118083170 M * Bertl they are smart, therefor they do not use the flag parser but instead code it as if/else tree doing 20 strcmps :/ 1118083238 M * Bertl of course, this is just because of the legacy code and backwards compatibility ... 1118083260 M * Bertl xfs is crap (at least the way it is in the kernel right now) 1118083511 M * brc bertl 1118083512 M * Bertl DaPhreak: you know that you have to test it? 1118083590 M * DaPhreak sure :) 1118083606 J * romke ~romke@procyon.romke.net 1118083638 Q * mcp Read error: Connection reset by peer 1118083647 M * Bertl brc 1118083648 M * sim0n Bertl, dietlibc compiles if i apply your dietlibc-gethostbyname-fix01.diff patch and a syscall patch from http://bugs.gentoo.org/show_bug.cgi?id=90316 ... i'll try to get util-vserver compiled as well... 1118083670 Q * Loki|muh Read error: Connection reset by peer 1118083671 J * Loki|muh loki@satanix.de 1118083710 M * brc Bertl: should i do a mount /dev/hdaX / -o tagxid on initrd's linuxrc ? 1118083715 J * mcp ~hightower@wolk-project.de 1118083753 M * Bertl brc: yep, that would probably help (if it does something similar right now) 1118083757 M * Bertl wb mcp! 1118083932 M * Bertl evening romke! 1118083960 Q * berni Ping timeout: 480 seconds 1118084153 Q * Beave Ping timeout: 480 seconds 1118084440 J * berni ~berni@svr01.mucip.net 1118085795 M * jkl anyone running a samba server within a vserver? 1118085811 M * jkl smbd and nmbd are having trouble finding local interfaces ... 1118085871 M * daniel_hozac WORKSFORME. 1118085876 M * jkl haha 1118085881 M * Bertl hmm, we had that today ;) 1118085891 M * jkl yeah, so uhhh 1118085899 M * Bertl jkl: check the irc logs ;) 1118085902 M * jkl it would help if i wasnt already running samba on the host 1118085932 A * jkl smacks himself 1118085974 M * Bertl well, with proper config, I guess you can run more than one, no? 1118086011 M * jkl absolutely 1118086055 M * jkl well, i dont know if it's able to bind to separate interface aliases 1118086077 M * jkl because then it would have worked while it was still running on the host 1118086083 M * Bertl me neither ... you have to check that ... 1118086117 M * Bertl but telling from the stuff folks reported here, samba seems to be highlander software ... 1118086124 M * Bertl (at least to some extend) 1118086181 M * jkl well, im not too worried 1118086193 M * jkl the whole point is to run samba on the vserver and not the host =) 1118086256 M * sim0n vservers seem to run fine now on my x86_64 (using glibc) 1118086295 M * sim0n but, uname -m gives x86_64, but it shouldn't (it should return x86...x86 guest) 1118086317 M * sim0n any idea how i could change that? 1118086440 M * Bertl yup, vuname (or in the config) 1118086649 M * sim0n what config should i change? (couldn't find it on the flower page...) 1118086687 M * DaPhreak should be in uts/ somewhere 1118086713 M * sim0n machine 1118086713 M * sim0n 1118086713 M * sim0n The machine-type of the vserver ? 1118086722 M * sim0n should i put x86 in that file? 1118086727 M * sim0n guess so ? :-) 1118086766 M * DaPhreak try it ;) 1118086990 M * sim0n yeah! that worked! THNX! :-) 1118087002 M * Bertl /dev/discs/disc1/part1 /vservers xfs rw,tagxid 0 0 1118087023 M * DaPhreak works ?! :) 1118087039 M * Bertl well, hey, you have to do the testing ;) 1118087051 M * DaPhreak yeah .. if i've got the patch ;) 1118087057 M * Bertl expect a patch in a few minutes 1118087083 M * DaPhreak yeah 1118087203 M * Bertl http://vserver.13thfloor.at/Experimental/FOR-2.0/delta-xfstag-fix01.diff 1118087288 M * DaPhreak make'ing now :) 1118087305 M * Doener hm... so it's true that apple is moving to x86... 1118087314 M * DaPhreak yeah ... sadly 1118087323 M * Bertl well, in theory ... 1118087544 M * DaPhreak hopefully they do this moving _only_ in theory .. their stocks already told them "_don't do this_ dumbass" ;) 1118087593 M * Bertl ahhh, now I figured it ... 1118087615 M * brc bertl 1118087617 M * brc /dev/root / ext3 rw,tagxid 0 0 1118087618 M * brc :) 1118087626 M * Bertl Doener: async usermodehelper is only partially async ... 1118087645 M * Bertl brc: congrats! what was the secret? 1118087692 M * brc i was creating a "linuxrc" on the initrc. the correct was to create a "init". just added the "mount" command there 1118087702 M * brc now it is time to play 1118087702 M * brc eheh 1118087754 M * Bertl Doener: so what about letting the user decide (per context) if the reboot helper should be sync or async? 1118087761 M * sim0n Doener, ???? apple moving to x86 ??? where did you get that from ? do you have news-links or so ? 1118087867 M * eyck sim0n: http://www4.macnn.com/macnn/wwdc/05/ 1118087867 M * DaPhreak sim0n: http://news.com.com/Apple+to+Intel+Some+advantage,+lots+of+risk/2100-1006_3-5716696.html 1118087879 M * sim0n thnx 1118087956 J * albeiro ~albeiro@albeiro.usercloak.oftc.net 1118087970 M * Bertl evening albeiro! 1118087978 M * albeiro hello Bertl :) 1118087993 M * albeiro i hate networking problems with admins not knowing what it is all about ;) 1118088022 M * Bertl what is it all about? 1118088048 M * Doener 42 1118088056 M * eyck it's all about pentiums baby... 1118088056 Q * locksy Ping timeout: 480 seconds 1118088057 M * Bertl ah, Doener is back ;) 1118088072 M * eyck yeah, with ultimate answers and stuff 1118088083 M * albeiro Bertl: switches are hanging without any good reason, same with server 1118088154 M * Doener Bertl: hm... why would the user want to choose that? 1118088199 M * Bertl okay, let me put it this way: which would be preferable for reboot/restart? async or sync? 1118088217 M * DaPhreak Doener: yeah thats the answer to the almighty question of the universe ;P but whats the question ?! 1118088255 M * Doener DaPhreak: the question is what it is all about... and as i don't know the question, i said the answer ;) 1118088257 M * Bertl Doener: I know that startup/shutdown requires sync ... 1118088305 M * Bertl and I also know that I need async to satisfy the legacy stuff 1118088362 M * Doener hm, what about having startup/shutdown sync and the rest async? 1118088421 M * Doener hey, sarge actually got stable... 1118088459 M * shuri :) 1118088594 M * albeiro anyway i will never understand apple decision 1118088608 M * albeiro just to intrduce small offtopic here ;) 1118088642 Q * sim0n Quit: Leaving 1118088644 A * Pazzo opens a bottle of vine - let's party! And don't stop 'til Sarge on Hurd becomes stable ;-) 1118088671 M * Bertl *ouch* that 'Hurds' ... 1118088749 M * Pazzo :-) Hi bErTl, hi dOEnEr *hicks* 1118088806 A * Pazzo is going to be drunken as it already happened some hours ago that sarge became stable - and still no hurd :'-p 1118089056 M * DaPhreak heh poor Pazzo ;) 1118089162 M * Pazzo hi DaPhreak ;) 1118089410 M * DaPhreak Bertl: /dev/hda7 on /var/vservers type xfs (rw,noatime,tagxid) 1118089834 M * Bertl good! working as expected? regarding lsxid/chxid umount/mount inbetween? 1118089929 M * DaPhreak what should be in the output of $( lsxid ) ?! 1118090347 M * DaPhreak Bertl: well the ls inside the vserver doesn't show the file anyway :) 1118090361 M * DaPhreak Bertl: well the ls inside the lsxid www/tmp/vserver doesn't show the file anyway :) 1118090364 M * DaPhreak www www/tmp/ 1118090367 M * DaPhreak 0 www/tmp/wordpress-1.5.12.tar.bz2 1118090397 M * DaPhreak vserver www exec ls /tmp/ returns nothing 1118090618 M * Bertl tmp is a mounted filesystem ;) 1118090639 M * Bertl nobody would expect it to auto-copy files you put there on the host ... 1118090719 M * DaPhreak aaah :) works now ;) 1118091326 M * DaPhreak *grrr* but now does the reboot -f buzzing me ;( about the missing /dev/initctl 1118091416 M * Doener reboot -f? that shouldn't care about /dev/initctl 1118091428 M * DaPhreak yeah .. _should_ 1118091430 M * DaPhreak ;) 1118091477 M * Bertl well, if it does, rip it out and replace it by a smaller one ;) 1118091496 M * DaPhreak heh .. was another script ;) .. 1118091501 M * DaPhreak well time for bed now :) 1118091512 M * Bertl night then! 1118091545 M * DaPhreak yeah .. n8 Bertl ;) well we'll probably see us tomorrow morning ;) 1118091711 M * Doener night DaPhreak 1118092540 J * yarihm ~yarihm@80-218-6-236.dclient.hispeed.ch 1118092636 M * Bertl evening yarihm! 1118092663 M * Bertl Doener: so what was your conclusion on sync vs async helper? 1118092685 M * Doener 22:06:01 Doener hm, what about having startup/shutdown sync and the rest async? 1118092709 M * Doener plus no startup/shutdown for legacy "contexts" 1118093264 M * yarihm hi Bertl 1118093356 M * Bertl Doener: hmm, missed that one ... 1118093367 M * Doener np 1118093377 M * Bertl Doener: but that doesn't answer my question: 1118093392 M * Bertl 22:03 < Bertl> okay, let me put it this way: which would be preferable for 1118093392 M * Bertl reboot/restart? async or sync? 1118093415 M * Hollow hey folks 1118093433 M * Bertl Doener: because actually, I think it _isn't_ the best solution to make it async 1118093510 M * Hollow Bertl: any time estimate until going stable? 1118093520 M * Doener can a context disappear while the reboot helper is running at all if the helper is running sync? 1118093554 M * Doener maybe i got that wrong, but the helper is started from a process running in the vserver, right? 1118093585 M * Doener and if that one is sync, the context won't disappear until the helper is finished... 1118093607 M * Bertl Hollow: as soon as all the test reports are evaluated ;) 1118093633 M * Hollow ok... ;) 1118093639 M * Bertl (the huge pile of replies to my 'please test the rcs' mail) 1118093665 M * Doener 0 or less? 1118093716 M * Bertl definitely less 1118093761 M * Bertl but I'm very confiden, a few hours after we release 2.0 folks will come here and complain that this and that isn't working ... 1118093932 M * aba Release is done. At least for me :) 1118093950 J * rs ~rs@imhotep.rhapsodyk.net 1118093992 M * Bertl aba yup ;) 1118094001 M * Bertl evening rs! 1118094012 M * rs evening 1118094020 M * Doener hum hum... in the last few days there've been some folks running 1.9.5, which is the latest development version on http://13thfloor.at/vserver/ ... maybe we should use that site for more stuff (for the non-ml/irc folks) 1118094021 M * aba (and we're running on top of our network capabilities :) 1118094055 M * Bertl Doener: yep, I plan to exted the 13thfloor pages to contain test/pre releases too 1118094074 M * Doener k 1118094092 M * Bertl will probably happen right after LT2005 1118094130 M * Bertl ad helper: no and that might be an important difference ... 1118094146 M * Bertl for example, if you want to 'wait' for the context exit, how would you do that? 1118094170 M * Bertl with an async helper, the context might already be dead, or even restarted again ... 1118094189 M * Bertl (that's one of the things which make it racey ;) 1118094300 M * Doener and the sync helper does not keep the context alive? 1118094302 M * Bertl the start/stop have to be sync, and they will move into non legacy, probably even into ngnet ... 1118094528 M * Bertl Doener: yes, it does, that's the whole point ... 1118094538 M * Bertl you know for sure, that the reboot command doesn't finish before your helper was run 1118094559 M * Bertl you can then (as in my hack example) start a thread which is waiting for the context termination, and exit the helper 1118094590 M * Doener d'oh! 1118094596 M * Bertl the syscall will return, reboot will 'finish' and hopefully got a signal (kill) before 1118094605 M * Doener threading in user space... 1118094617 M * Doener i obviously totally forgot about that... 1118094649 M * Doener my thinking was: helper gets called, context won't go away, helper waits for the context to go away... 1118094651 M * Bertl we can do the following: 1118094671 M * Bertl make the hreboot async for legacy, sync for new api 1118094677 M * Doener but of course it works if some other process does the waiting... 1118094682 M * Bertl move the startup/shutdown into NGNET 1118094702 M * Bertl i.e. make it part of 2.0 but depend on the NGNET config 1118094776 Q * nayco Quit: Bonne nuit ! 1118094820 J * locksy ~locksy@mrtg.sisgroup.com.au 1118094825 M * Bertl okay, anyways ... off for now .. back later ... 1118094832 N * Bertl Bertl_oO 1118094832 M * Doener cya! 1118095208 M * brc bertl 1118095209 M * brc gone 1118095209 M * brc :( 1118095239 M * mugwump Hey have we seen much happen on the util-vserver front? 1118095744 M * brc i am trying to use vdlimit 1118095749 M * brc can someone help me ? 1118095796 M * mugwump I've always been an LVM person myself for that sort of thing.. 1118095884 M * brc what is lvm ? 1118095927 J * Aiken ~james@tooax6-003.dialup.optusnet.com.au 1118095961 M * jkl what is vdlimit? 1118096054 M * brc limit hard disk per vserver 1118096085 M * jkl ah sweet 1118096106 M * jkl i think LVM may stand for linux virtual machine? 1118096108 M * jkl i am not sure 1118096248 M * Doener brc, jkl: LVM is Logical Volume Management 1118096273 M * brc Doener: ok 1118096384 M * Doener with that you can create a logical volume for each vserver and thus provide the vserver with a limited amount of hdd space 1118096411 M * Doener of course this does not work with unification, as hard links can not be made across filesystems 1118096616 M * brc hm 1118096619 M * brc but 1118096631 M * brc if the user uses "df", will he see his real limit ? 1118096657 M * Doener he will see how much space he has on 'his' logical volume 1118096671 M * brc cool 1118096674 M * Doener basically you can treat the lv like a partition 1118096692 M * Doener but resizing a lv is easier ;) 1118096706 M * daniel_hozac well, not easier than using one partition and using vdlimit ;) 1118096839 M * Doener and vdlimit allows really hot resizing ;) 1118096858 M * Doener brc: http://www.tldp.org/HOWTO/LVM-HOWTO/ -- chapter 1 and 2 might be interesting for you 1118097025 Q * Aiken Quit: Leaving 1118097098 M * brc gonna check it out 1118097103 M * brc as soon as i finish the vdlimit stuff 1118097117 M * brc my doubt is if i can ru vdlimit twice, without removing it before 1118097122 M * brc understood what imean ? :) 1118097181 M * daniel_hozac which vdlimit are you using? the one in util-vserver or Bertl's? 1118097186 J * Aiken ~james@tooax6-003.dialup.optusnet.com.au 1118097189 M * brc hm 1118097191 M * brc an old one 1118097195 M * brc util-vserver 1118097650 N * Bertl_oO Bertl 1118097653 M * Bertl back now ... 1118098051 Q * erwan_ho Remote host closed the connection 1118098068 M * Doener wb Bertl 1118098110 M * Doener I agree on your proposal for the helper 1118098172 M * Bertl the last one? with async/sync on legacy? 1118098191 M * Bertl and startup/shutdown moved to ngnet? 1118098226 Q * yarihm Quit: Leaving 1118098233 M * Doener yep, sounds quite reasonable 1118098262 M * Bertl okay, then let's do it this way, I'll prepare a patch 1118098355 M * Doener i'll port rc3 to rc6 now... 1118098407 M * Bertl k 1118099001 M * Doener seems like an easy one... just a few offsets 1118099489 Q * eyck Ping timeout: 480 seconds 1118100134 J * eyck eyck@81.219.64.71 1118100536 M * Doener brb 1118100538 Q * Doener Quit: Leaving 1118101347 J * Doener ~doener@p54874FB5.dip.t-dialin.net 1118101362 M * Bertl wb Doener! 1118101380 M * Doener Linux doener 2.6.12-rc6-vs2.0-rc3 #1 Tue Jun 7 01:07:37 CEST 2005 i686 AMD Athlon(tm) XP 2600+ AuthenticAMD GNU/Linux :) 1118101392 M * Bertl congrats! 1118101430 M * Doener thx 1118101489 M * Doener http://www.13thfloor.at/~doener/vserver/patches/patch-2.6.12-rc6-vs2.0-rc3.diff 1118101579 M * Bertl tx