1308701090 J * julius ~julius@217.20.127.15 1308701123 Q * dowdle Remote host closed the connection 1308701157 Q * nou Ping timeout: 480 seconds 1308701212 Q * julius_ Ping timeout: 480 seconds 1308701237 Q * derjohn_mob Ping timeout: 480 seconds 1308701438 J * derjohn_mob aj@88.128.224.190 1308702187 J * Chat8356 axlgmj@app6.chatmosphere.org 1308702312 Q * daniel_hozac_ Ping timeout: 480 seconds 1308702789 Q * Chat8356 Quit: Chatmosphere for iPhone & BlackberryPART #vserver Chatmosphere for iPhone & Blackberry 1308703269 J * el_bb_ ~el_btd@109.125.107.76.dynamic.cablesurf.de 1308703271 J * nou Chaton@2001:6f8:328:bbc:6666:6667:: 1308703740 M * nicholi why is it enough to set the barrier flag on just say /vservers instead of each individual guest filesystem in /vservers? 1308703786 M * nicholi also i'm a bit confused about the documentation in wiki, and i think it might be wrong (http://linux-vserver.org/Secure_chroot_Barrier) 1308703803 M * Bertl you do not even need the barrier nowadays, it's more an additional security measure 1308703804 M * nicholi at the bottom it gives an example to set the barrier for each guest: cd -P /etc/vservers//vdir; setattr --barrier .. 1308703812 M * nicholi ahh ok 1308703825 M * Bertl but you put it just above the guest root 1308703839 M * Bertl to prevent any changes to the filesystem just above that 1308703872 M * nicholi right, but that wouldn't protect guests from each other 1308703877 M * nicholi right ? 1308703889 M * Bertl it does 1308703944 M * Bertl the only escape (in the original implementation) was to modify the filesystem above the guest filesystem to remove the barrier itself 1308703973 M * Bertl besides the fact that it isn't required anymore, the barrier blocks any attempts to get rid of the barrier :) 1308704056 M * nicholi i get what you are saying, but if it prevents changes to the filesystem above it...and sibling directories count as part of the filesystem "above" ? 1308704081 M * Bertl it prevents changes and access to the directory containing the guest root 1308704090 M * nicholi ahh ok 1308705361 J * daniel_hozac ~daniel@c-923071d5.08-230-73746f22.cust.bredbandsbolaget.se 1308705376 M * Bertl wb daniel_hozac! 1308708285 M * Bertl off to bed now ... have a good one everyone! 1308708292 N * Bertl Bertl_zZ 1308710714 Q * el_bb_ Quit: goodbye 1308711102 J * mib_fo4y2i 180a322a@ircip2.mibbit.com 1308711132 M * mib_fo4y2i hello? 1308713516 Q * mib_fo4y2i Quit: http://www.mibbit.com ajax IRC Client 1308716823 J * sannes ~ace@cm-84.209.81.224.getinternet.no 1308718960 Q * thierryp Remote host closed the connection 1308721946 J * thierryp ~thierry@zanzibar.inria.fr 1308722765 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1308723641 Q * derjohn_mob Ping timeout: 480 seconds 1308724577 J * ghislain ~AQUEOS@adsl2.aqueos.com 1308725141 J * petzsch ~markus@dslb-088-075-124-160.pools.arcor-ip.net 1308725455 J * derjohn_mob ~aj@213.238.45.2 1308728447 Q * arekm Ping timeout: 480 seconds 1308728465 Q * hijacker_ Ping timeout: 480 seconds 1308730102 Q * petzsch Read error: Connection reset by peer 1308731207 Q * derjohn_mob Ping timeout: 480 seconds 1308731217 J * derjohn_mob ~aj@p4FFD256E.dip.t-dialin.net 1308731618 J * kir ~kir@swsoft-msk-nat.sw.ru 1308733567 J * petzsch ~markus@dslb-088-075-124-160.pools.arcor-ip.net 1308735582 M * g_en__ if my util-vserver build fails like this, what i'm missing? http://pastebin.com/9L1rG7jR sorry ahead for building such old software 1308735608 N * g_en__ glen 1308735663 M * glen linker? diet? linux-libc-headers? 1308735744 N * Bertl_zZ Bertl 1308735750 M * Bertl morning folks! 1308735816 M * Bertl what kernel/patch/util-vserver? 1308735839 M * glen i'm just building util-vserver 1308735849 M * glen util-vserver-0.30.215 1308735869 M * Bertl without a kernel? 1308735877 M * glen kernel will be pld linux's 2.6.27.x 1308735898 M * glen but i'm stuck with userspace compile 1308735902 M * Bertl that sounds like a great combination 1308735944 M * glen i'm with older pld release, and besides the new pld combination is not that stable 1308735953 M * Bertl what is the purpose of building an util-vserver version which is more than three years old with a kernel which is known to be broken (but a lot newer)? 1308735987 M * glen >=2.6.27 requires udev dependency which building is more than trouble 1308736007 M * glen what is broken in 2.6.27? 1308736016 M * glen it's one of the "ltm" branches 1308736040 M * Bertl why not take the stable 2.6.22 kernel then? 1308736081 M * glen if i build without dielibc, i get different but somewhat similiar error: http://pastebin.com/N0WLLsVR 1308736135 M * glen i could take 2.6.22 too, but i have problem building util-vserver 1308736138 M * Bertl well, you definitely do not want to build without dietlibc ... but OTOH, you do not seem to care for a working Linux-VServer setup either 1308736231 M * glen what does dietlibc does there? glibc is broken? i tought difference is just "heaviness" of glibc 1308736255 M * Bertl no, glibc is broken in that way that it dynamically loads stuff 1308736278 M * Bertl which basically allows a guest to mess with host libraries and also leads to wrong behaviour 1308736305 M * Bertl if you really think 2.6.27 is the best your 'old' distro can handle, I'd go for 2.6.22 but a recent util-vserver 1308736354 M * glen is recent util compatible with kernels down to that old ones? 1308736366 M * Bertl yes, it should be 1308736382 M * Bertl although we haven't tested with 2.6.22 for some time now 1308736443 M * glen found 2.6.25 kernel, and seems to boot up host and guest 1308736500 M * glen util-vserver-0.30.214-5.amd64 being userspace 1308736529 M * Bertl are you doing a history project? 1308736589 M * glen no, just latest available in pld isn't that stable 1308736604 M * glen so want to use something that is known to work (on our other hosts) 1308736613 M * Bertl what does that mean, pld stopped working with 2.6.27? 1308736629 M * glen no no 1308736636 M * Bertl so? 1308736659 M * glen known combination is pld th, which has: util-vserver-0.30.216-1.pre2955.3.x86_64 kernel-2.6.38.6-8.x86_64 1308736681 M * glen but it is not stable, some problems which are not ok 1308736694 M * Bertl so updating that with a recent 2.6.38 kernel/patch and util-vserver and everything should be fine 1308736727 M * glen i have one lock up situation, but i can't really play with that to make bugreport 1308736734 M * Bertl if you actually _have_ problems with that combination, you'll even get immediate help with it here 1308736758 M * glen the thread starts here: http://lists.pld-linux.org/mailman/pipermail/pld-devel-en/2011-May/022173.html 1308736816 M * glen damn, one post is missing in that list archive 1308736875 M * glen anyway, that combination i posted there: 1308736875 M * glen 2.6.37.6-2 1308736876 M * glen util-vserver-0.30.216-1.pre2955.3.x86_64 1308736911 M * glen had problem that if one guest got 20+ conncurrent http connections, apache inside that guest locked up for 5 minutes 1308736962 M * glen so i booted back to 2.6.36.2-1, to avoid the lockup (and downgraded sudo too to avoid 32bit guest bug) 1308737032 M * Bertl so pld uses a sudo (32 bit) which doesn't work on a 64bit kernel? 1308737149 M * glen that is solved now, don't pay attention 1308737157 M * Bertl how was it solved? 1308737171 M * glen ok, you didn't read the posts 1308737186 M * glen kernel upgrade was needed whcih had updated patch 1308737203 M * glen supposedly fixed by this: 1308737204 M * glen http://cvs.pld-linux.org/cgi-bin/cvsweb.cgi/packages/kernel/kernel.spec.diff?r1=1.874;r2=1.875;f=h 1308737222 M * Bertl okay, so back to my suggestion, using a 2.6.38.8 kernel with Linux-VServer patch and everything would be fine, no? 1308737234 M * glen i.e patch update from patch-2.6.37-vs2.3.0.37-rc2.diff to patch-2.6.37-vs2.3.0.37-rc3.diff 1308737249 M * Bertl instead you are going back to even older kernels 1308737250 M * glen i don't think i'm able to build current util-vserver 1308737270 M * Bertl how so? 1308737306 M * glen those link errors i pasted? 1308737350 M * glen oh now. HEAD comppiled: Wrote: /home/glen/rpm/pld/packages/RPMS/util-vserver-0.30.216-1.pre2967.1.i686.rpm 1308737372 M * Bertl why do you compile an old version _and_ for i686? 1308737373 M * glen is that utils fresh enough? 1308737399 M * glen i686 is just because i test builded there :) 1308737425 M * Bertl simply grab the pre2981 and be done 1308737441 M * glen doing that already 1308737464 M * Bertl but yes, the 2967.1 (whatever that .1 means) is definitely better for a more recent kernel 1308737508 M * glen .1 is pld subversion (pld needs own bumps too if rebuilding same snap to perform upgrades, typical in package management) 1308737666 M * glen but with 2.6.38 i will have problems with/without udev 1308737699 M * glen as i have udev-124 and >=2.6.27 do not work with that one. afaik 1308737709 M * glen and 2.6.27 supposedly didn't work without udev at all 1308737748 M * Bertl so how does the 2.6.37.6-2 or 2.6.36.2-1 work? 1308737773 M * glen that is newer distro line, where is newer udev 1308737796 M * Bertl but your aim is to run an ancient one instead, yes? 1308737834 M * glen more or less, i think keeping old distro line and updating kernel+utils is less painful and known to work that upgrading whole distro line 1308737851 M * Bertl note that if udev is really your only concern, the recent kernels support udev on tmpfs which is populated by the kernel 1308737885 M * glen i don't think that approach is tested (or even enabled) in pld 1308737892 M * Bertl so similar to the devfs, it should work without udev support from userspace, at least for statically detected devices 1308737945 M * Bertl well, as pld folks are not time travellers either, I totally agree, but the setups you are trying to construct aren't tested by anyone either ... 1308738030 M * glen blah. arekm (likely) has put some construct that don't work for older distro line, need to play with those too :( 1308738033 M * glen xz >= 1:4.999.7 is needed by kernel-2.6.38.8-1.src 1308738036 M * glen hostname is needed by kernel-2.6.38.8-1.src 1308738038 M * glen rpm-build >= 4.5-24 is needed by kernel-2.6.38.8-1.src 1308738073 M * Bertl there is no hostname in pld before 2.6.38 ? 1308738083 M * glen it was in different package 1308738366 M * Bertl and why exactly are you trying to get the 'old line' working? 1308738397 M * Bertl does the new pld line not work with your hardware? 1308738435 M * glen i can't just upgrade os right now 1308738451 M * glen that is simply put 1308738477 M * glen i'm trying to compile 2.6.38 1308738483 M * Bertl okay, and the old os isn't supported anymore by pld? 1308738575 M * glen something like that, if i'm able to find combination, then that could be used 1308738837 J * hijacker ~hijacker@87-126-142-51.btc-net.bg 1308738846 M * Bertl well, what about updating udev, which shouldn't be a big deal IMHO, and then using a recent kernel/patch/util-vserver? 1308738953 M * glen don't know which udev version is good enough 1308738982 M * glen by good i mean one that works with picked kernel version, altho if i use that pld th uses, that could be considered "working" 1308738983 M * Bertl I'd say, the first one booting up with 2.6.38.8 will do 1308740575 M * glen eh. gcc 3.3 is too old too to compile that kernel 1308740575 M * glen /home/users/glen/rpm/BUILD.amd64-linux/kernel-nogrsecurity-2.6.38.8/linux-2.6.38/drivers/staging/line6/pcm.c:137: error: implicit declaration of function `__sync_fetch_and_and' 1308740584 M * glen (lacks atomic ops) 1308740591 M * glen but i'll build kernel in pld th, and install in pld ac 1308740622 M * glen error: util-vserver-0.30.216-1.pre2981.1.amd64: req util-linux >= 2.13 not found 1308740696 M * Bertl sounds like an rpm dependancy 1308740708 M * glen yes, that dep is added due ionice 1308740747 M * glen perhaps it's optional? 1308740790 M * Bertl maybe, if not or if you want it, just compile it and install it (will be detected) 1308740875 M * glen compiling util-linux update itlself major build dependency hell 1308740956 M * Bertl http://paste.linux-vserver.org/19995 1308740976 M * Bertl (doesn't look like many dependancies to me :) 1308741004 M * glen nah, _building_ it is 1308741037 M * Bertl so grab a new rpm, unpack that file and see if it works 1308741060 M * Bertl util-vserver doesn't care where it comes from :) 1308743116 P * kir Leaving. 1308743654 M * glen ok. got some combination installed, reboot time 1308743760 M * glen util-vserver-0.30.216-1.pre2981.1.amd64 kernel-2.6.38.8-1.x86_64 1308744978 J * ryker ~Adium@c-76-16-115-27.hsd1.in.comcast.net 1308745118 J * ghislain1 ~AQUEOS@adsl2.aqueos.com 1308745405 Q * ghislain Ping timeout: 480 seconds 1308748736 M * WMP Bertl: new server: 1308748738 M * WMP vserver mint start 1308748739 M * WMP RTNETLINK answers: File exists 1308748741 M * WMP /usr/lib/util-vserver/vserver.functions: line 1594: /dev/cgroup/mint/tasks: No such file or directory 1308748854 M * Bertl the first error means that you are trying to bring up an IP which most likely is already there 1308748921 M * Bertl the second one means that you messed up the cgroups once again 1308749003 M * WMP Bertl: what i missed? 1308749028 M * Bertl no idea, what did you do? 1308749062 M * WMP emmm, i want to make new vserver 1308749087 M * Bertl so what was the command line? what's your kernel and util-vserver version? 1308749100 M * WMP 2.6.38.8-vs2.3.0.37-rc17-vserver 1308749121 M * WMP util-vserver: 0.30.216-pre2981; 1308749122 M * Bertl no idea what the -vserver is, but the rest looks like a good choice 1308749137 M * WMP yes, becouse before restart work good 1308749149 M * Bertl so you did reboot, yes? 1308749159 M * WMP yes 1308749163 M * Bertl were the runlevel scripts executed? 1308749166 M * WMP and after reboot all dont work 1308749174 M * Bertl i.e. util-vserver and vprocunhide? 1308749185 M * WMP maybe i shoud now run /etc/init.d/util-vserver 1308749192 M * WMP only vprocunhide 1308749201 M * Bertl no, they should have been run already on bootup 1308749213 M * WMP but isn't ;) 1308749219 M * Bertl fix this and reboot the machine to verify that they are executed 1308749231 M * WMP now i cant reboot 1308749262 M * WMP is posisble to run this new Vserver? 1308749269 M * WMP guest* 1308749284 M * Bertl depends on what 'this new guest' ist 1308749286 M * Bertl *is 1308749300 M * WMP emmm... 1308749321 M * WMP vserver mint build -m clone ... 1308749323 M * Bertl make sure that the runlevel scripts are executed at boot 1308749339 M * Bertl if they aren't for whatever reason, execute them by hand 1308749346 M * Bertl before you try anything 1308749391 M * WMP run now /etc/init.d/util-vserver is safe to guest? 1308749566 M * Bertl guests won't be very happy without it 1308749583 M * WMP ugh 1308749590 M * WMP so i must make reboot... 1308749590 M * Bertl it doesn't affect the guests on disk, but you shouldn't have a guest running without executing that first 1308749630 M * Bertl as I said yesterday and probably the day before, make sure that the runlevel scripts are executed on bootup :) 1308749713 M * WMP it isn't, i must add this script to runlevel 1308749732 M * Bertl so do that?! 1308749782 M * WMP but i cant reboot now 1308749791 M * Bertl then wait till you can 1308749841 M * WMP Bertl: http://wklej.org/id/550888 - ok? 1308749892 M * Bertl looks good 1308749955 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1308750255 M * karasz hello, where can I find _examples_ of cpu throtling/scheduling? 1308750527 M * Bertl older or newer kernels? 1308750564 M * karasz 2.6.27.43-vs2.3.0.36.8-dist is older or newer? 1308750581 M * karasz util-vserver 0.30.216-pre2864 1308750587 M * Bertl somewhere inbetween, mainly unsupported ... 1308750604 M * karasz and cpu is Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz 1308750655 M * Bertl I'd strongly advise to upgrade to 2.6.38.8 in your case 1308750714 M * karasz i can upgrade to 2.6.35.12 1308750732 M * karasz or maybe a newer 2.6.35 1308750746 M * Bertl why's that? 1308750811 M * karasz 2.6.35 is LTS afaik, and that is what we have in our SDE 1308750822 M * karasz for the moment at least 1308750839 M * Bertl SDE being? 1308750880 M * karasz system development environment http://git.opensde.net 1308750905 M * Bertl okay, well, and you are interested in cpu limits, hard limits maybe as well? 1308750956 M * karasz i am unfamiliar with the terms, what i want to achieve is to not let guests hog all the cpu 1308750991 M * Bertl with 'other' processes running or even when the cpu would be otherwise idle? 1308750992 M * karasz as far as i understood that is not cpu.hard and soft because those mean cpu times 1308750996 M * karasz but i might be wrong 1308751057 M * karasz if cpu is idle they can use it, but i think it would be wise to leave a small percentage of cpu free at all times (at most 2-3%) 1308751116 M * Bertl if that is what you want, you definitely need hard limits, together with the long-term requirement, that would make 2.6.32.41 your best choice 1308751281 M * karasz so that means that i need to revamp the existing/functioning server to achieve that? 1308751306 M * karasz revamp=rebuild+reinstall+reinstall existing vserserv guests 1308751331 M * Bertl yep, 2.6.32.41 + Linux-VServer supports hard scheduler limits via cgroups 1308751338 M * karasz ic 1308751343 M * karasz thx for the info 1308751363 M * Bertl and that's what you want, as you figured already, the rlimits are not what you need to limit 1308751382 M * karasz yeah, thought that after reading the docs 1308751395 J * d3vic3 ~quimzed@41.225.4.218 1308751402 M * karasz and 2.6.38 prefered to 2.6.32, right? 1308751417 M * Bertl yep, although 2.6.38 doesn't feature hard limits 1308751438 M * Bertl mainline has been working on that for more than a year, but there is nothing included yet 1308751448 M * Bertl a patch to add that pops up every now and then 1308751551 Q * pmjdebruijn Quit: leaving 1308751702 M * karasz patch-2.6.32.41-vs2.3.0.36.29.7.diff < for 2.6.32.41 right? 1308751797 M * Bertl yep 1308751811 J * pmjdebruijn ~pascal@overlord.pcode.nl 1308751828 Q * pmjdebruijn 1308751856 J * pmjdebruijn ~pascal@overlord.pcode.nl 1308753467 J * d3vic3_ ~quimzed@41.225.4.218 1308753585 Q * d3vic3 Ping timeout: 480 seconds 1308753859 M * Bertl nap attack ... bbl 1308753863 N * Bertl Bertl_zZ 1308753987 Q * FireEgl Ping timeout: 480 seconds 1308754198 P * d3vic3_ 1308754606 J * FireEgl ~FireEgl@173-16-9-3.client.mchsi.com 1308755313 J * dowdle ~dowdle@scott.coe.montana.edu 1308757040 J * BenG_ ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1308757049 Q * BenG Quit: I Leave 1308757101 Q * BenG_ 1308758164 Q * thierryp Remote host closed the connection 1308760102 J * bonbons ~bonbons@2001:960:7ab:0:90c4:7b10:93c9:b1a2 1308760589 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1308761043 N * Bertl_zZ Bertl 1308761046 M * Bertl back now ... 1308761476 Q * derjohn_mob Ping timeout: 480 seconds 1308762251 J * arekm ~arekm@ixion.pld-linux.org 1308762375 M * nicholi Bertl: from your last convo, what is the most recent kernel with hard cpu scheduling ? 2.6.32.41 ? 1308763915 M * Bertl yep 1308764122 Q * tomreyn__ Quit: tomreyn__ 1308765361 M * nicholi unify is magical... 300 instances of ubuntu in about 25-30gb :) 1308765395 J * Piet ~Piet__@659AACMF8.tor-irc.dnsbl.oftc.net 1308765520 Q * bonbons Quit: Leaving 1308769536 M * WMP Bertl: after reboot i have: RTNETLINK answers: File exists 1308769543 M * WMP but all other work good 1308769576 J * imcsk8 ~ichavero@148.229.1.11 1308771319 M * Bertl which means that you are using an IP which is already used 1308771503 M * WMP lsof dont show connection on this ip 1308772257 M * karasz Bertl: recommended util-vserver for 2.6.32.41? 1308772289 M * karasz latest as in util-vserver-0.30.216-pre2981? or util-vserver-0.30.216-pre2883 will do fine? 1308772352 Q * ghislain1 Quit: Leaving. 1308772688 Q * hijacker Quit: Leaving 1308773251 Q * hparker Quit: Quit 1308774577 M * WMP Bertl: how to clone guest over 2 other hosts? :) 1308774647 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1308776006 Q * sannes Remote host closed the connection 1308776519 M * nicholi karasz: my guess would be for anything thats 2.6.32+ always get latest util-vserver (as in straight from svn) 1308776550 M * nicholi which itself hasn't been touched in awhile it looks like 1308777968 Q * petzsch Quit: Leaving. 1308778129 J * derjohn_mob aj@80.187.243.57 1308778363 Q * BenG Quit: I Leave 1308779623 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1308780360 M * [Guy] I'm trying to do something weird with network contexts and it doesn't work :) 1308780363 N * [Guy] Guy- 1308780376 M * Guy- I'm starting dnscache with ncontext --create --silentexist --nid 4 -- nattribute --set -- naddress --add --ip 127.1.1.1/8 --ip 172.18.10.254/32 -- 1308780396 M * Guy- it binds its client sockets to 127.1.1.1 and tries to contact 127.0.0.53 1308780407 M * Guy- these packets don't show up on the loopback interface in tcpdump 1308780410 M * Guy- what could be wrong? 1308780428 M * daniel_hozac you want to use lback 1308780494 M * Guy- where? 1308780540 M * Guy- naddress doesn't advertise an lback switch 1308780632 M * Guy- ah, in the source I see that there is one 1308780647 M * Guy- how do I use it? --lback 127.1.1.1 instead of --ip 127.1.1.1/8? 1308780720 M * Guy- yes, that worked, thanks 1308781167 Q * DoberMann Ping timeout: 480 seconds 1308781347 M * WMP daniel_hozac: hmmm, in quest on ubuntu i have: start: Unable to connect to Upstart: Failed to connect to socket /com/ubuntu/upstart: Connection refused 1308781532 M * daniel_hozac as expected unless you use plain init 1308781565 M * WMP hmmm ;) 1308781567 M * WMP ok 1308782660 J * DoberMann ~james@2a01:e35:8b44:84c0::2 1308783445 Q * _are_ Quit: No Ping reply in 180 seconds. 1308783466 J * _are_ ~quassel@vs01.lug-s.org 1308785236 N * ensc Guest5781 1308785246 J * ensc ~irc-ensc@p5DF2F4D8.dip.t-dialin.net 1308785655 Q * Guest5781 Ping timeout: 480 seconds 1308786431 M * ser where should i search for vserver kernels for fedora 15? 1308786437 J * bernard 518d54d9@ircip3.mibbit.com 1308786443 M * bernard hi 1308786459 M * bernard i cant get irc channels to run in ubuntu 1308786495 P * bernard 1308786501 M * ser bernard: do they work properly in debian?