1343611888 J * ep ~e@121.97.136.89 1343612081 M * ep Hi, everytime I run vserver-stat, one of my servers loses its context. Is there a fix for this? 1343612176 M * ep There is a workaround by doing echo contextid > /var/run/vservers/vservername but its gets a little annoying as I have to do it almost everyday 1343612700 M * daniel_hozac vserver-stat shouldn't be changing anything. 1343612777 M * ep But whenever I ran vserver-stat, one of the servers loses its context. Restarting the VM won't fix the issue 1343612956 M * ep Based on the FAQ on the website, it says: Just do a : cat /etc/vservers//context > /var/run/vservers/ 1343613158 M * ep I am getting the same error on this post: http://www.mail-archive.com/vserver@list.linux-vserver.org/msg09994.html, but the post was already 5 years ago and I can't find a same issue with a recent one 1343613209 M * ep I am running Debian Squeeze and using the squeeze vserver packages 1343614706 M * daniel_hozac so 3-4 year old packages. 1343614710 M * daniel_hozac you might want to upgrade those. 1343619364 Q * clopez Ping timeout: 480 seconds 1343630095 J * ghislain ~AQUEOS@adsl1.aqueos.com 1343631733 N * Bertl_zZ Bertl 1343631737 M * Bertl morning folks! 1343631914 Q * eyck Remote host closed the connection 1343632898 M * ep morning ;) 1343633938 J * kir ~kir@swsoft-msk-nat.sw.ru 1343636185 M * ghislain hello there 1343636807 Q * jeroen__ Quit: Ex-Chat 1343636850 J * ffrank ~ffrank@g231241042.adsl.alicedsl.de 1343636939 M * ffrank hi. i'm a little confused by the information on http://linux-vserver.org/Downloads. the channel seems to suggest that release candidates of the utils should be preferred over the "current" verison. is this a fair assessment? 1343636997 M * Bertl the "current" version i.e. the most recent "stable" release is many years old 1343637022 M * Bertl so, assuming you do not use a kernel that old, you probably want one of the pre releases 1343637379 M * ffrank will do, thanks 1343637669 M * ep Bertl, does that mean the one on Debian Squeeze is old? 1343637766 M * Bertl no idea, what version is used there? 1343637931 M * ep I think its this one. linux-image-2.6.32-5-vserver-amd64 1343637957 M * ep util-vserver-0.30.216-pre2864 1343637973 M * Bertl well, that's defintiely an older kernel but not ancient, you cannot use that kernel with the latest stable util-vserver in a secure way 1343638048 P * kir PING 1343638048 1343638050 M * ep I don't mind if its old, but I do have a problem with it though. Sometimes when I run vserver-stat, other VMs do not have context on it 1343638098 M * ep doing cat /etc/vservers//context > /var/run/vservers/ would fix it, but it will come back after a couple of days 1343638120 M * ep is it a bug thats already been fixed? 1343638129 M * Bertl well, in general, IIRC, it is suggested to update debian packages to more recent ones, there are also repositories for that 1343638173 M * Bertl the debian versions are often strange mutations of the mainline versions, so it's not uncommon they have their own bugs 1343638263 M * ep so upgrading is probably the best way 1343638277 J * jeroen__ ~jeroen@095-097-051-172.static.chello.nl 1343638292 M * Bertl definitely 1343638612 M * yang ep: however there are no vserver kernel's in more modern releases of Debian, so beware the upgrade...best bet is to compile a custom version of the kernel+vserver patch or use a pre-compiled 3rd party .deb for debian 1343638740 M * ep oh yeah. I've completely forgot about that. 1343638856 M * ghislain beng provide a repository for debian kernel mostly the debian config file with the latest vserver patch you could use those 1343638909 M * ep do you know the URL? 1343639687 J * BenG ~bengreen@cpc29-aztw23-2-0-cust105.18-1.cable.virginmedia.com 1343639733 M * Bertl http://repo.psand.net/info/ 1343639750 M * Bertl http://repo.psand.net/info/index_lenny.html 1343640190 M * ffrank debian has discontinued vserver for a while. policy is to go for lxc in the long run. it's sad. 1343640536 M * Bertl IMHO not the worst of decisions considering the bad package quality 'they' provided 1343640982 J * clopez ~clopez@fanzine.igalia.com 1343641275 M * ffrank Bertl: so would it be a bad idea to base my own utils packages on the one's from debian? any hints on what needs improving? 1343641351 M * Bertl probably it would be a bad idea, as there is no reason for changing the upstream version 1343641374 M * Bertl i.e. upstream util-vserver will work just fine on debian 1343641583 Q * BenG Quit: I Leave 1343641833 M * ffrank hmm, are there upstream deb packages of release candidates? 1343641907 M * Bertl nope, upstream uses sources, but as beng does, it should be rather simple to package them for debian 1343641996 M * ffrank i see. my original question was whether the debian ones used unsound build options or the like, because otherwise the deb-sources should become the basis for one's own up-to-date packages 1343642059 M * Bertl no idea what they used or didn't use, but there is a wiki page collecting some of the brokenness of various debian packages 1343642102 M * Bertl in general, if you plan on doing your personal package, it doesn't matter as long as you're happy :) if you plan on doing a publick package, it would be good to get as close as possible to upstream 1343642185 M * ffrank got it. thanks 1343642250 M * Bertl there is a spec file in the util-vserver sources, and IIRC, there also was something for debian some time ago, might have been removed or might be there but just outdated 1343642282 M * ffrank i shall report my findings ;) 1343642283 M * Bertl in general I think daniel_hozac doesn't mind adding files for building debian packages from the sources (if that is possible with debian) 1343642308 M * Bertl as long as they are not mere hacks but proper solutions 1343643336 Q * ep 1343644434 J * jango2k3 ~jango2k3@pd95bee42.dip0.t-ipconnect.de 1343646301 M * ffrank daniel_hozac: so ctags is a build requirement for the utils? so is iptables? 1343646761 M * ghislain bertl: starting a guest do not launch the S level ? It seems quota are not started and my test show that /etc/rcS.d is not used with vserver 1343647050 M * Bertl which is expected, yes, you can select the runlevel used in sysv init mode in the config, but it will be only one runlevel 1343647057 Q * ensc|w Remote host closed the connection 1343647069 J * ensc|w ~ensc@www.sigma-chemnitz.de 1343647483 Q * jango2k3 Quit: Leaving 1343649427 J * fisted_ ~fisted@xdsl-87-78-229-61.netcologne.de 1343649837 Q * fisted Ping timeout: 480 seconds 1343650191 M * disposable is there a vserver patch for linux 3.2.0 (ubuntu 12.04)? when i try to apply patch-3.2.23-vs2.3.2.12.diff, i get too many rejects. 1343650409 M * Bertl well, for 3.2.0 there probably is an older patch in the same place where the 3.2.23 one is 1343650425 M * Bertl the question is more if ubuntu really uses that outdated kernel 1343650456 M * Bertl (because I doubt that they did not incorporate the fixed done since 3.2) 1343651461 M * ffrank fwiw - they do not. 1343651495 M * ffrank disposable: are you using the linux-source package? 1343652159 Q * FireEgl Read error: Connection reset by peer 1343652825 J * FireEgl FireEgl@2001:470:e5ad:1:5977:3b15:74fe:52e1 1343653395 M * ghislain bertl: you mean that in debian it will only do level 2 and not launch the S one, ok i see. So i need to launch it manualy after the boot 1343653425 M * Bertl usually 'S' is for harware setup and should not be required for services 1343653481 M * Bertl what services do you 'start' in 'S'? 1343653488 M * ghislain yes, but quota are in this level unfortunatly :) 1343653500 M * ghislain but yes all others are hardware clock and such setup 1343653513 M * Bertl so maybe enable quota for runlevel 2 then? 1343653615 M * ghislain yes this is the other option 1343653778 N * s1aden sladen 1343655547 M * disposable Bertl: ffrank: well, the current kernel in Ubuntu 12.04 is linux-image-3.2.0-29-generic. if i do apt-get source linux-image-3.2.0-29-generic and apply patch from 3.2.23, i get 4 rejects (Makefile.rej, fs/proc/base.c.rej, fs/proc/root.c.rej, kernel/pid_namespace.c.rej) which don't look insignificant to me. 1343655593 M * disposable there is no patch for 3.2.0 at vserver.13thfloor.at 1343655726 M * ffrank disposable: well, ubuntu *does* patch the kernel, so patches against upstream won't fit 100%. but 4 rejects is pretty OK, given that kernel.org (and ubuntu presumably) is now 3.2.24 instead of .23 1343655789 M * ffrank i've fitted vserver to patched distro kernels twice, and had at least a dozen merge conflicts to fix each time. I've found git with vimdiff as "mergetool" to be a great toolchain for this purpose. 1343655883 M * disposable ffrank: if the kernel were written in python MAYBE i would give it a go. anyway, how do you know the kernel is in fact 3.2.24? 1343655933 M * ffrank disposable: I don't, it may well be .23 or earlier. I'm going on a limb here, but I'd assume that ubuntu is rather quick to incorporate upstream fixes. 1343655998 M * ffrank small hint: distro patches may include features from the next higher major version. I've had luck peeking at the 2.6.33 patch when porting the 2.6.32 patch to SUSE's sources. 1343656426 M * disposable the patch for linux-3.5 fits with ubuntu's kernel with no problem but unfortunately requires higher version of util-vserver than the one distributed by ubuntu. i'm trying to keep things simple. 1343656649 M * ffrank humm, now that's genuinely surprising 1343656987 M * Bertl disposable: the problem is, there was no 3.2.0 released 1343657001 M * Bertl i.e. mainline released 3.2 and then 3.2.1 1343657016 M * Bertl (so you have to look for 3.2 instead of 3.2.0 :) 1343657054 M * Bertl http://vserver.13thfloor.at/ExperimentalT/patch-3.2-vs2.3.2.5.diff 1343657064 M * Bertl was the patch we released when 3.2 came out 1343657110 M * Bertl but as the patch for 3.5 works for ubuntu's '3.2.0', I presume it is actually a 3.5 kernel 1343657259 M * disposable Bertl: I haven't tried to apply the 3.5 patch against ubuntu's 3.2.0. gimme a moment to try 1343657290 M * Bertl hmm? 1343657312 M * Bertl what did you mean with: the patch for linux-3.5 fits with ubuntu's kernel with no problem? 1343657367 M * disposable i meant with ubuntu's 3.5 kernel (a backport from ubuntu 12.10) 1343657399 M * disposable i now see how ambiguous my statement was 1343657422 M * disposable and i apologise for confusing ffrank and you 1343657467 M * Bertl well, you can grab all the patches since 3.2 and try them one by one 1343657485 M * Bertl you will always get a reject on the Makefile, as it contains the version number of the kernel 1343657494 M * Bertl (i.e. you can ignore that) 1343658368 M * disposable Bertl: it's the other ones I can't ignore http://pastebin.com/JzifiGB7 There should be an official linux-vserver debian source or an ubuntu ppa (now that Christoph Lukas stopped maintaining his one and vserver kernel is missing from debian wheezy) 1343658594 M * Bertl the mainline kernels works perfectly fine on debian/ubuntu I've been told, and there is the psand repository maintaining debian packages as well 1343658964 M * disposable does this slow disappearance from mainline distros mean you're giving up in favour of lxc? 1343659076 M * Bertl we do not maintain or control distro specific packages/kernels/whatever 1343659089 M * Bertl Linux-VServer is and always was distro agnostoc 1343659092 M * Bertl *agnostic 1343659121 M * Bertl i.e. we provide the sources and patches and everybody regardless of the host/guest distro can use them 1343659142 M * Bertl it would be a nightmare to maintain packages for all the distros out there 1343659186 M * Bertl now, for distros like debian dropping Linux-VServer support, you have to ask the debian folks (same goes for ubuntu, although ubuntu usually mimics debian :) 1343659270 M * Bertl as we are using those parts of LXC (kernel side) which are already useable in Linux-VServer we are slowly reducing the Linux-VServer patches as LXC grows 1343659337 M * disposable you need a lobby group to push linux-vserver into libvirt and distros. preferably female and preferably scantily clad. that's bound to attract attention at a dev conference. 1343659338 M * Bertl at some point, LXC will be able to replace Linux-VServer kernel patches and all that remains will be util-vserver (to provide the well known config interface) for those who want it 1343659402 M * Bertl hehe, well, we prefer not to push and we avoid spending time/money/whatever on advertizing :) 1343659450 M * Bertl doesn't mean that we do not appreciate folks advertizing for Linux-VServer or spending their money on advertizing Linux-VServer technologies 1343659966 Q * hparker Ping timeout: 480 seconds 1343660000 M * ffrank disposable: that reject doesn't look too bad. especially the includes are usually trivial, caused by upstream having an include that wasn't in the patch base 1343660635 Q * yang Remote host closed the connection 1343660639 J * yang yang@jazz.linuxshell.org 1343661633 M * ffrank disposable: btw, apparently ubuntu is at 3.2.21, judging from *their* current Makefile patch 1343661771 M * disposable ffrank: thanks. it seems i just need to change the range information lines in the diff file. 1343661806 M * ffrank disposable: that would be strange, because simple offset errors patch can usually handle 1343661919 Q * yang Remote host closed the connection 1343661986 M * disposable well, the first reject needed to be ~50 lines lower 1343662232 J * yang yang@yang.netrep.oftc.net 1343662359 M * ffrank i see. great. 1343663922 M * ghislain bertl: when i try to patch k 3.2.24 i got a: patching file kernel/sched.c 1343663922 M * ghislain Hunk #2 succeeded at 3465 with fuzz 1 (offset -135 lines). 1343663932 M * ghislain the 'with fuzz' means an issue ? 1343663992 M * Bertl depends, basically it means that patch needed to skip one of the context lines to apply the patch 1343664028 M * Bertl as the patches I provide have 3 lines of context, this means that still 2 lines were matched 1343664065 M * ghislain oh so a part was not applied so this is bad (tm) i presume 1343664086 M * Bertl no, the context lines are already there in the file 1343664128 M * Bertl (see man patch) 1343664156 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1343664200 M * ghislain ok, FYI all the 'hunk' are http://paste.linux-vserver.org/22889 so seems pretty good 1343664665 J * bonbons ~bonbons@2001:960:7ab:0:dc21:8aaf:a47:d6b3 1343665659 M * disposable :wq 1343665664 M * disposable sorry 1343665821 Q * ffrank Quit: :wq indeed ;) 1343666399 J * hijacker_ ~hijacker@cable-84-43-134-121.mnet.bg 1343668186 M * ghislain ^^ 1343670282 M * Bertl off for a nap .. bbl 1343670287 N * Bertl Bertl_zZ 1343675772 Q * hijacker_ Quit: Leaving 1343676127 Q * FireEgl Ping timeout: 480 seconds 1343680978 J * petzsch ~markus@p57B67F36.dip.t-dialin.net 1343681281 Q * hparker Ping timeout: 480 seconds 1343681729 Q * clopez Ping timeout: 480 seconds 1343682048 Q * bonbons Quit: Leaving 1343682679 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1343683242 J * clopez ~clopez@131.29.165.83.dynamic.mundo-r.com 1343683457 Q * petzsch Quit: Leaving. 1343684295 Q * imcsk8 Remote host closed the connection 1343686121 Q * PowerKe Ping timeout: 480 seconds