1269302898 Q * ntrs_ Ping timeout: 480 seconds 1269303349 Q * bonbons Quit: Leaving 1269307604 J * Romster ~romster@202.168.100.149.dynamic.rev.eftel.com 1269309535 M * Bertl off to bed now .. have a good one everyone! 1269309542 N * Bertl Bertl_zZ 1269311608 Q * MeCooL Ping timeout: 480 seconds 1269316849 J * SauLus_ ~SauLus@c135068.adsl.hansenet.de 1269317257 Q * SauLus Ping timeout: 480 seconds 1269317257 N * SauLus_ SauLus 1269322213 J * FireEgl FireEgl@173-16-9-10.client.mchsi.com 1269322467 J * thierryp ~thierry@home.parmentelat.net 1269323319 J * sharkjaw ~gab@90.149.121.45 1269325236 J * imcsk8 ~ichavero@evdomip-5-41.iusacell.net 1269326556 J * ntrs ~ntrs@77.29.4.212 1269326822 Q * derjohn_mob Ping timeout: 480 seconds 1269328298 J * hijacker ~hijacker@213.91.163.5 1269328372 J * ncopa ~ncopa@180.40.189.109.customer.cdi.no 1269328758 J * ghislain ~AQUEOS@adsl2.aqueos.com 1269328813 Q * thierryp Remote host closed the connection 1269329722 J * derjohn_mob ~aj@139.12.1.252 1269329834 Q * imcsk8 Quit: This computer has gone to sleep 1269332306 J * thierryp ~thierry@zankai.inria.fr 1269332520 J * dna ~dna@31-216-103-86.dynamic.dsl.tng.de 1269332890 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1269334888 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269335553 J * yarihm ~yarihm@80-219-150-83.dclient.hispeed.ch 1269336477 Q * petzsch Quit: Leaving. 1269336604 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1269337677 Q * m_o_d Ping timeout: 480 seconds 1269338419 J * kir ~kir@swsoft-msk-nat.sw.ru 1269338948 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269339969 Q * bonbons Quit: Leaving 1269340080 M * BenG is there a way to give a guest the impression that's it's memory is limited? 1269340114 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1269340248 M * petzsch http://linux-vserver.org/util-vserver:Cgroups#using_cgroup_to_enforce_memory_limits : When you wish the guests to see only their limited memory pool, be sure to include VIRT_MEM in your cflags config file. 1269340453 M * BenG cheers petzsch, so the feature is only available with vs2.3.0.36.29+ 1269340499 M * petzsch prior memory limiting is done like described here: http://linux-vserver.org/Memory_Limits 1269340515 M * petzsch the cflag to see only the virtual memory in the guest is the same 1269340528 M * BenG oh, marvelous 1269340722 M * BenG Unknown ccap 'VIRT_MEM' 1269340743 M * BenG I don't get why that's happening 1269340828 M * BenG I just did "echo VIRT_MEM >> ccapabilities" 1269340838 M * BenG in the appropraite folder, but now the guest won't start 1269340874 J * djMedrzec ~filip@filip.math.uni.lodz.pl 1269340877 M * djMedrzec hi :) 1269340900 M * djMedrzec is there a way to check in a script whether I am running inside vserver? like some entry in /proc or something? 1269340914 M * petzsch hmm... i allways put them in /etc/vservers/vserver-name/cflags (old: flags) 1269340918 M * ghislain geb: will look into the test for a link in case for a next cversion of the .deb 1269340966 M * petzsch djMedrzec: /proc/ is very likely a lot less entries than on a real server 1269340972 M * petzsch what's uname -a saying? 1269341073 M * BenG djMedrzec, oh yes, that does the trick 1269341077 M * BenG thanks 1269341382 M * djMedrzec on host: Linux tefnet 2.6.31-17-tefserver-core2 #54+tefnet1 SMP Thu Jan 7 19:32:11 CET 2010 x86_64 GNU/Linux, on vserver: Linux test.tefnet.pl 2.6.31-17-tefserver-core2 #54+tefnet1 SMP Thu Jan 7 19:32:11 CET 2010 x86_64 GNU/Linu 1269341405 M * djMedrzec no difference except the hostname 1269341477 M * djMedrzec I want to have conditional expressions in my startup scripts so that I don't do fsck if I'm inside a vserver 1269341535 J * SlackLnx ~SlackWare@a85-139-11-123.cpe.netcabo.pt 1269341606 M * ncopa hi 1269341624 M * ncopa thanks for providing 2.6.32.x patches, even if 2.6.33 is out 1269341626 M * ncopa appreciate 1269341721 A * SlackLnx boas 0/ 1269341918 M * BenG petzsch, do you know how to set VIRT_MEM as a default? would I put it in /etc/vservers/.defaults/ somewhere? 1269341985 M * BenG hmmm, doesn't seem to work 1269342036 M * petzsch not listed on the great flower page... duno if there's a workaround for it 1269342065 M * petzsch ncopa: also happy about the 2.6.32.x patch as 2.6.32.x is supposed to get LTS from kernel.org 1269343200 M * djMedrzec invserver() { test ! -e /proc/cmdline; } 1269343204 M * djMedrzec is there anything better? 1269343226 Q * BenG Quit: I Leave 1269343567 N * Bertl_zZ Bertl 1269343571 M * Bertl morning folks! 1269343605 M * Bertl djMedrzec: there is no 100% way to detect that you are in a guest, except for doing a syscall atm 1269343696 M * Bertl djMedrzec: but assuming that the guest are _not_ specifically protected or disguised, you can test for the xid in /proc/self/status 1269343802 M * djMedrzec Bertl: on a virtual one it's != 0? 1269343903 M * Bertl correct, inside a guest you see VxID: 1301 (for example) unless it is 'disabled' with a special flag 1269343926 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1269343953 M * djMedrzec Bertl: thanks :) 1269345273 Q * petzsch Quit: Leaving. 1269346242 Q * bonbons Read error: Connection reset by peer 1269346259 Q * BenG Quit: I Leave 1269346388 P * kir Leaving. 1269346435 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1269348080 J * ntrs_ ~ntrs@77.28.24.130 1269348516 Q * ntrs Ping timeout: 480 seconds 1269349162 J * BenG ~bengreen@cpc2-aztw22-2-0-cust521.aztw.cable.virginmedia.com 1269349289 Q * BenG 1269351352 Q * bonbons Read error: Connection reset by peer 1269351567 Q * yarihm Quit: Leaving 1269351900 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1269353850 J * barismetin ~barismeti@zanzibar.inria.fr 1269354196 Q * dna Quit: Verlassend 1269354690 Q * SlackLnx Read error: Connection reset by peer 1269354714 J * SlackLnx ~SlackWare@a85-139-11-123.cpe.netcabo.pt 1269354899 M * Bertl off for now ... bbl 1269354903 N * Bertl Bertl_oO 1269355223 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269355256 J * dowdle ~dowdle@scott.coe.montana.edu 1269355990 Q * ghislain Quit: Leaving. 1269355999 J * ghislain ~AQUEOS@adsl2.aqueos.com 1269357314 Q * petzsch Quit: Leaving. 1269357509 Q * swen Remote host closed the connection 1269357592 M * derjohn_mob Hi, I compiled a 2.6.33.1 + 2.3.0.36.30.3 (or so :)) and now I get a "vspace: clone(): Invalid argument" when trying to start vservers 1269358291 J * MeCooL mecool@94.129.169.67 1269358398 Q * niki Quit: Leaving 1269358610 M * derjohn_mob Do the tool cause that behavior or the kernel patch? 1269359891 Q * MeCooL Ping timeout: 480 seconds 1269359927 N * Bertl_oO Bertl 1269359935 M * Bertl derjohn_mob: util-vserver version? 1269360000 Q * sharkjaw Remote host closed the connection 1269360255 M * derjohn_mob debian squeeze ( vserver 0.30.216-pre2864 -- manages the state of vservers ) 1269360526 M * Bertl I'd try a newer one 1269360564 Q * ktwilight Ping timeout: 480 seconds 1269360627 M * derjohn_mob Bertl, foo :/ even in Experimental there are no newer ones. 1269360658 Q * ncopa Quit: Ex-Chat 1269360696 M * Bertl you might work around that by unselecting certain space features 1269360750 M * derjohn_mob is this the place for the freshest tarballz ? http://ftp.linux-vserver.org/pub/utils/util-vserver/ 1269360762 M * derjohn_mob or do I need to pull from SVN ? 1269360795 M * derjohn_mob and: Which features should I try to disable ? 1269361027 M * Bertl http://people.linux-vserver.org/~dhozac/t/uv-testing/ 1269361033 M * Bertl namespace features :) 1269361034 M * derjohn_mob Bertl, I did not change any vserver option when moving from 2.6.32.x to 2.6.33. 1269361053 M * derjohn_mob so, I have to disable mainline kernel's stuff ? 1269361058 M * derjohn_mob (space stuff) 1269361109 J * MeCooL mecool@94.128.18.204 1269361143 M * derjohn_mob CONFIG_NAMESPACES=y <== that one ? 1269361316 M * Bertl well, you do not want to completely disable this, but some of the newer ones maybe 1269361392 J * ktwilight ~keliew@91.180.59.107 1269361684 M * derjohn_mob Bertl, it the only namespace thingy i can find in my .config 1269361690 M * derjohn_mob besides that: vserver 0.30.216-pre2882 -- manages the state of vservers 1269361702 M * derjohn_mob it still does not clown .... 1269361713 M * derjohn_mob eh, clone ;) 1269361727 M * Bertl upload your kernel .config 1269361883 M * derjohn_mob ok 1269361898 M * derjohn_mob maybe make oldconfig did make bad foo ? 1269362245 M * derjohn_mob *lol* Your post could not be processed. It looks like spam. 1269362560 M * derjohn_mob Bertl, http://volatile.derjohn.de/config-linux-2.6.33-vserver2.3.0.36.30-nfct-300hz-volpre-derjohn.de-amd64 1269362574 M * derjohn_mob (the pastebin's don't like my content) 1269362584 M * derjohn_mob *pastebins 1269362755 M * Bertl could you upload 'vserver-info - SYSINFO' as well? 1269362904 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269363203 M * derjohn_mob ok 1269363255 M * derjohn_mob Bertl, http://paste.linux-vserver.org/14850 1269363304 M * derjohn_mob Bertl, I wonder what I do differently from other deployers of the newest Experimental Patch ... 1269363459 M * Bertl no idea, actually ... 1269363468 M * Bertl does testme.sh work fine? 1269363486 M * derjohn_mob *lol* .... I didnt try that one for a long time ... let me check ! 1269363526 M * derjohn_mob completely succeeded - no fails. 1269363588 M * Bertl okay, and you get that error when you do what exactly? 1269363597 M * derjohn_mob my .config for 2.6.32.3 worked fine, I used a make oldconfig to "convert" it to a 2.6.33 config 1269363630 M * derjohn_mob after that, the error appeard. vserver-stat does not throw an error, but vserver build and vserver enter do. 1269363655 M * derjohn_mob I recognized the error when i tried to start a vserver. 1269363658 M * Bertl okay, so 'building' a new guest gives you that error, yes? 1269363680 M * derjohn_mob yes, I cannot build a fresh one (tried with -m debootstrap) 1269363725 M * Bertl okay, give me a few minutes to update one of the test machines to the latest 2.6.33 patch 1269363739 M * derjohn_mob I currently run the 2.6.33.1 on my notebook, on my vserver-testserver 2.6.33 doesnt run at all, as there are problems with the forcedeth driver. 1269363766 M * derjohn_mob I enabled staging drivers and run the kernel as 300Hz - but this shouldnt matter 1269363790 M * derjohn_mob Bertl, ok, thanks! 1269363810 M * derjohn_mob I'll translocate now, will be back in soon .. 1269363825 M * Bertl k, cya 1269363895 Q * petzsch Quit: Leaving. 1269364300 Q * derjohn_mob Ping timeout: 480 seconds 1269365258 J * fzylogic ~fzylogic@dsl081-243-128.sfo1.dsl.speakeasy.net 1269365481 M * fzylogic Bertl: when CONFIG_CGROUP_MEM_RES_CTLR_SWAP is disabled, the memory calculations assume that old-style limits are in place and display full system swap as available to the guests when in fact all that's in place is the cgroup hard memory limit 1269365570 M * Bertl yeah, we probably need to put in some defaults there 1269365715 J * dna ~dna@31-216-103-86.dynamic.dsl.tng.de 1269365794 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269365844 Q * SlackLnx Read error: Connection reset by peer 1269365864 J * SlackLnx ~SlackWare@a85-139-11-123.cpe.netcabo.pt 1269365879 M * fzylogic I wasn't sure how to tell the difference between having no limit and having an unlimited limit 1269365897 M * fzylogic but seeing as how the plan's to remove those old limits altogether... 1269365928 M * fzylogic turns out that CONFIG_CGROUP_MEM_RES_CTLR_SWAP makes it a royal pain to change limits around on the fly without restarting a guest 1269365951 Q * SlackLnx Read error: Connection reset by peer 1269365968 J * SlackLnx ~SlackWare@a85-139-11-123.cpe.netcabo.pt 1269365973 M * Bertl really? 1269365990 M * fzylogic you can't lower the total limit below the amount of memory in use 1269366009 M * fzylogic so if you have the swap limit, it lets you lower the physical ram limit and the remainder gets swapped out 1269366010 J * derjohn_mob ~aj@f048099174.adsl.alicedsl.de 1269366024 M * Bertl well, that sounds fine to me, no? 1269366049 M * fzylogic not when you have 40 customers who don't know any better on each host :) 1269366049 M * Bertl if you want to 'lower' the memory limit below the current usage, you have to kill something in the guest 1269366067 M * fzylogic precisely, I'd much rather have it trigger OOM and then lower the limit 1269366085 M * fzylogic all of the swapping seriously slows down the host 1269366125 M * Bertl not quite unexpected 1269366152 M * Bertl you could submit a feature request to mainline ... 1269366164 M * fzylogic no, not at all. but far from optimal when you allow any customer to change their limits freely 1269366169 M * fzylogic yes, I was considering it 1269366177 M * fzylogic I just discovered the issue late yesterday 1269366195 M * fzylogic and for now I'm more than happy to just remove that option from the kernel 1269366218 M * fzylogic assuming the memory display was correct, that is :) 1269366241 M * Bertl well, we can fix that up if you volunteer to do the testing ... 1269366253 M * fzylogic absolutely 1269366482 M * Bertl soo, what behaviour makes sense when memory limits are on, but the swap limits are off? 1269366508 M * Bertl we probably want to show no swap at all, yes? 1269366560 M * fzylogic correct 1269366592 M * fzylogic used/free/available should all be hard-coded to zero when there's no soft/swap limit in place 1269366726 M * Bertl okay, makes sense, give me a second to test that 1269366990 M * Bertl fzylogic: that should do the trick: http://vserver.13thfloor.at/ExperimentalT/delta-memcg-fix01.diff 1269367036 M * fzylogic building 1269367073 J * imcsk8 ~ichavero@189.155.124.174 1269367273 Q * ntrs_ Ping timeout: 480 seconds 1269367289 Q * barismetin Quit: Leaving... 1269367357 Q * imcsk8 1269367520 Q * petzsch Quit: Leaving. 1269367552 Q * thierryp Quit: ciao folks 1269368817 Q * balbir Ping timeout: 480 seconds 1269369254 M * fzylogic so far, so good 1269369442 J * balbir ~balbir@122.172.14.145 1269371165 J * petzsch ~markus@p4FF4799E.dip.t-dialin.net 1269371170 M * derjohn_mob Bertl, any new in reproducing the problem/upgrading yout testserver ? 1269371269 M * Bertl not yet, build finished a few minutes ago 1269371402 M * derjohn_mob kk 1269371467 M * Bertl rebooting now ... 1269372053 M * Bertl guest build is running now ... 1269372649 Q * dna Ping timeout: 480 seconds 1269372786 M * Bertl derjohn_mob: guest install and startup worked fine 1269372822 M * Bertl this was now with 0.30.216-pre2864 1269372828 M * Bertl will now test with latest version 1269373207 M * Bertl building now with 2882 1269373490 J * dna ~dna@31-216-103-86.dynamic.dsl.tng.de 1269373720 M * Bertl derjohn_mob: works fine too 1269373755 M * Bertl derjohn_mob: ah, you enabled CONFIG_CGROUP_NS=y 1269373793 M * Bertl this requires that you either tell the kernel that it shouldn't unshare this one (config) or you need to disable it in the kernel config 1269373809 M * Bertl btw, the same is true for 2.6.32 1269373900 J * ntrs ~ntrs@77.28.24.130 1269374407 M * derjohn_mob Bertl, how do I tell the kernel ? (dynamically?) .. but I'd better disable that feature 1269374416 M * derjohn_mob I will cause confusion in out company ... 1269374426 M * derjohn_mob *out 1269374440 M * Bertl recent kernels allow you to set an unshare mask 1269374458 M * Bertl (not sure there is util-vserver support for that yet) 1269374629 M * derjohn_mob well, I dont need that feature ... 1269374650 M * derjohn_mob which technology does use those namespaces? LXC ? 1269374684 M * Bertl I presume so 1269374799 M * derjohn_mob Is LXC generally a surrogate for vserver ? Or is the approach too different? 1269374851 M * Bertl it is what mainline builds to replace Linux-VServer and similar in the future 1269374898 M * Bertl according to sam vilain it works perfectly fine already ... but he is the only one reporting that :) 1269375006 M * derjohn_mob A friend of mine told me he uses it since 2.6.29 (?) or so successfully. But a quick look at the sparse docs, it looks like the networking stuff is done with virtualized nics. 1269375015 Q * ghislain Quit: Leaving. 1269375085 M * Bertl it is indeed .. but I'm still waiting for somebody to show a production install with LXC 1269375097 M * Bertl don't get me wrong, I'm all for it 1269375195 M * derjohn_mob I never doubt that. You once told me, that you would be willing to concentracte on the details, if some new container suff comes up in the kernel ... 1269375785 Q * dna Ping timeout: 480 seconds 1269376310 Q * SlackLnx Quit: I'll Be Back 1269376738 J * niki ~niki@188.228.10.177 1269377158 J * Bushmills ~l@scarydevilmonastery.net 1269377256 J * ntrs_ ~ntrs@77.28.108.12 1269377257 Q * ntrs Read error: Connection reset by peer 1269377513 J * ghislain ~AQUEOS@adsl2.aqueos.com 1269377955 Q * ghislain Quit: Leaving. 1269378200 Q * Bushmills Read error: Connection reset by peer 1269378843 J * Bushmills ~l@scarydevilmonastery.net 1269379430 J * dna ~dna@31-216-103-86.dynamic.dsl.tng.de 1269379700 Q * Bushmills Remote host closed the connection 1269379847 Q * petzsch Quit: Leaving. 1269380341 Q * dna Quit: Verlassend 1269380381 J * Bushmills ~l@scarydevilmonastery.net 1269380502 Q * Bushmills Remote host closed the connection 1269380980 J * Bushmills ~l@scarydevilmonastery.net 1269381130 Q * Bushmills Remote host closed the connection 1269381258 Q * niki Quit: Leaving 1269381327 J * imcsk8 ~ichavero@148.229.1.11 1269383543 J * Bushmills ~l@scarydevilmonastery.net 1269383551 Q * ntrs_ Ping timeout: 480 seconds 1269383590 Q * Bushmills Remote host closed the connection