1165623042 P * stefani I'm Parting (the water) 1165623286 N * Piet_ Piet 1165623871 Q * Piet Remote host closed the connection 1165623930 J * Piet hiddenserv@tor.noreply.org 1165623945 J * dreamind apwdsl@p548AB2E1.dip0.t-ipconnect.de 1165624021 M * dreamind Hi folks :) 1165624382 M * daniel_hozac hello 1165624610 M * dreamind hi daniel_hozac :) 1165624618 A * dreamind finally got his macbook pro :) 1165624650 M * daniel_hozac sounds nice 1165624881 Q * dna Quit: Verlassend 1165624991 M * dreamind yes, 17" notebooks are nice ;) 1165625030 M * daniel_hozac i'm of the opposite opinion... ;) 1165625036 M * dreamind hehe ;) 1165625037 A * daniel_hozac prefers a 12" notebook. 1165625044 M * dreamind I'm ok with that too ;) 1165625053 M * dreamind but for me a 17" desktop replacement is nice 1165625148 M * dreamind and I'm happy this one now has 2gb ram, will get nice when running parallels :D 1165625326 M * daniel_hozac i bet. 1165625797 M * Bertl dreamind: that is one of those intel books, yes? 1165625805 M * dreamind Bertl: yup, core 2 duo 1165625820 M * Bertl yeah, did you install linux on that yet? 1165625821 M * dreamind hi Bertl btw (: 1165625828 M * dreamind Bertl: nope, will run mac os x. 1165625839 M * Bertl what? no dual boot? 1165625842 M * dreamind nope. 1165625846 M * dreamind only virtualized 1165625870 M * dreamind remember, mac os x is a unix too, and its quite perfect for me on that laptop. 1165625888 M * Bertl well, it's ancient mach/bsd :) 1165625997 M * dreamind wow this thing is fast... just installed macports... 1165626693 Q * dreamind Read error: Connection reset by peer 1165627167 J * Johnnie ~jdlewis@jdlewis.org 1165627174 Q * Johnnie 1165627822 J * Johnnie ~jdlewis@jdlewis.org 1165629610 Q * Piet Quit: Piet 1165630062 M * Bertl okay, compiling a ppc kernel now 1165630097 M * Bertl daniel_hozac: any specific requirements? should it work with 2.2.0-rc2? 1165630112 M * daniel_hozac should work fine with everything. 1165630117 M * daniel_hozac :) 1165630141 M * Bertl will 2.2.0 exercise the 'new' code? (on 2.6.19, of course)? 1165630150 J * Aiken_ ~james@tooax6-182.dialup.optusnet.com.au 1165630155 M * daniel_hozac if it has a high enough version and the VCI-bit, yes. 1165630161 M * Bertl morning Aiken_! 1165630200 M * Bertl well, I'm compiling the version I uploaded :) 1165630229 M * daniel_hozac so it should be fine :) 1165630374 Q * mcp Read error: Connection reset by peer 1165630375 J * _mcp ~hightower@wolk-project.de 1165630476 Q * Aiken Ping timeout: 480 seconds 1165630651 M * daniel_hozac does 2.2 have the space fix already? 1165630682 M * daniel_hozac won't work too well without that. 1165630709 M * Bertl well, we'll see and adjust ... but IIRC, it should have all but the proc changes 1165630740 M * daniel_hozac 2.2.0-rc2's timestamp is older than delta-space-fix0[12], at least. 1165630756 M * Bertl indeed 1165630782 M * Bertl okay, will update/check 1165631359 Q * Johnnie Read error: Connection reset by peer 1165631517 M * Bertl hmm, why is your patch named cacct-fix01, while I seem to have it here as cacct-clean2? 1165631564 M * daniel_hozac cacct-clean02 does more than cacct-fix01, no? 1165631594 M * Bertl ah, so that is the 'fix only' part then? 1165631633 J * Johnnie ~jdlewis@jdlewis.org 1165631643 M * daniel_hozac well, it's more of a quick-fix thing. i guess the headers should be moved around similarly to how they are in devel, with a cacct_def.h containing _vx_cacct and the enum. 1165631690 M * daniel_hozac in 2.0, _vx_cacct is still in context.h, while the enum is in cacct.h. 1165631713 M * daniel_hozac (thus the include cacct.h from context.h) 1165631732 M * Bertl hmm, the branches seem to get out of sync ... 1165631748 M * Bertl I'm currently looking at 2.1.x for 2.6.19 1165631777 M * Bertl that has the enum in cact.h 1165631783 M * Bertl *cacct.h 1165631798 M * Bertl and the proper defines in cacct_def.h 1165631803 M * daniel_hozac ah, i misremembered that then. 1165631821 M * Bertl okay, let's check 2.2.x then 1165631856 M * daniel_hozac looks the same as devel. 1165631881 M * Bertl okay, so the only 'affected' is 2.0.x on 2.6.18.x 1165631886 M * daniel_hozac yep. 1165631892 M * Bertl okay, good 1165631901 M * Bertl that needs to be refurbished anyway 1165631935 M * Bertl the question for 2.0.x is, who will keep it alive in the future ... 1165631950 M * daniel_hozac is there any interest in keeping it alive? 1165631951 M * Bertl probably some distros are locked to that branch 1165631971 M * daniel_hozac i.e. Debian? 1165631991 M * Bertl the question is, will somebody care to backport stuff, or will they migrate to 2.2.x? 1165632073 M * Bertl obviously marc cared enough to backport the scheduler stuff for planetlab 1165632073 M * daniel_hozac i'm guessing most people will be migrating to 2.2. 1165632096 M * Bertl although I do not know if that one got tested ... 1165632162 M * daniel_hozac has 2.2 removed the quota hashes? 1165632169 M * Bertl yes 1165632184 M * daniel_hozac okay. 1165632205 M * Bertl I still kept them in 2.3.x for now 1165632320 M * daniel_hozac is the fs/xfs/quota/xfs_qm_syscalls.c hunk (vx_capable in 2.1 vs. capable in 2.2) intentional? 1165632358 M * Bertl is it a quota check? 1165632390 M * Bertl I think it is unintentional, but will look into it shortly 1165632418 M * daniel_hozac yeah, it's a quota check. 1165632452 M * daniel_hozac 2.2.0-rc2 also sets the SPACES VCI bit twice (once after COW, once after history). 1165632464 M * Bertl I didn't plan to remove the quota checks, so I assume it was lost 1165632571 M * Bertl do you have a patch fragment at hand for the quota? 1165632579 M * Bertl I'll fix up the VCI manually 1165632701 J * mcp ~hightower@wolk-project.de 1165632713 Q * _mcp Read error: Connection reset by peer 1165632804 M * Bertl btw, I think we should add a prefix entry for ipv4 too 1165633124 M * daniel_hozac hmm, we already have the mask, what would the prefix do? 1165633266 Q * Aiken_ helium.oftc.net nobelium.oftc.net 1165633266 Q * Simbalin helium.oftc.net nobelium.oftc.net 1165633266 Q * Smutje helium.oftc.net nobelium.oftc.net 1165633266 Q * Curus helium.oftc.net nobelium.oftc.net 1165633266 Q * ensc helium.oftc.net nobelium.oftc.net 1165633266 Q * mountie helium.oftc.net nobelium.oftc.net 1165633266 Q * Medivh helium.oftc.net nobelium.oftc.net 1165633266 Q * bogus helium.oftc.net nobelium.oftc.net 1165633266 Q * trippeh helium.oftc.net nobelium.oftc.net 1165633277 J * Aiken_ ~james@tooax6-182.dialup.optusnet.com.au 1165633277 J * Simbalin ~Simba@deb30.mgts.by 1165633277 J * Smutje ~Smutje@xdsl-87-78-98-134.netcologne.de 1165633277 J * Curus ~Curus@kbhn-vbrg-sr0-vl209-213-185-8-10.perspektivbredband.net 1165633277 J * ensc ~irc-ensc@p54B4E0AE.dip.t-dialin.net 1165633277 J * mountie ~mountie@CPE0080c6fe323f-CM000a739acaa4.cpe.net.cable.rogers.com 1165633277 J * Medivh ck@paradise.by.the.dashboardlight.de 1165633277 J * bogus ~bogusano@fengor.net 1165633277 J * trippeh atomt@x.vx.no 1165633338 Q * shedi Quit: Leaving 1165633363 M * daniel_hozac interdiff -p1 patch-2.6.19-vs2.1.x-t7.diff patch-2.6.19-vs2.2.0-rc2.diff | filterdiff -i '*/fs/xfs/quota/xfs_qm_syscalls.c' | patch -Rp1 -d linux... 1165633684 M * daniel_hozac well, i'm tired so i'm gonna get some sleep now. let me know how it goes with 0.30.212-rc5 and i'll release it in the morning (if there are no problems). good night all! 1165634187 M * Bertl okay, np, have a good one! 1165634188 J * dmax ~semaj@81.193.57.225 1165634264 J * s0undt3ch ~s0undt3ch@81.193.57.225 1165634291 Q * s0undt3ch 1165634298 M * Bertl wb dmax! 1165634311 J * s0undt3ch ~s0undt3ch@81.193.57.225 1165636901 J * marcfiu ~mef@lupin.hongo.wide.ad.jp 1165636906 M * marcfiu hello 1165637007 M * marcfiu Bertl: is anyone working on demonstrating soundly that vserver subsumes lsm functionality? 1165637299 Q * Johnnie Remote host closed the connection 1165638344 M * Bertl marcfiu: hey! and hmm? 1165638529 J * Johnnie ~jdlewis@jdlewis.org 1165638579 M * marcfiu Bertl: still thinking? 1165638612 M * marcfiu Bertl: I'm in Tokyo right now and completely time confused. What time is it for you now? 1165638625 M * Bertl 5:30am, but that's fine 1165638645 M * Bertl I don't understand your question though ... 1165638666 M * Bertl why should somebody do that? 1165639537 M * Bertl marcfiu: still around? 1165640714 Q * marcfiu Ping timeout: 480 seconds 1165642205 J * marcfiu ~mef@lupin.hongo.wide.ad.jp 1165642210 M * marcfiu oops 1165642220 M * marcfiu bertl: I'm still around.... 1165642368 M * marcfiu bertl: wrt demonstrating soundly that vservers (or in the long run a containerized linux) subsumes lsm would open the door to eliminating integrated LSM support. 1165642405 M * marcfiu Bertl: so I was just curious whether there was anyone in the vserver community either already doing this or interested in doing this. 1165642459 M * Bertl well, I remember somebody trying to implement Linux-VServer with LSM 1165642474 M * Bertl and as far as I know, he failed at some point ... 1165642510 M * Bertl haven't seen anybody actually use LSM, but I have heard of folks combining certain LSM features with Linux-VServer 1165642540 M * Bertl but AFAIK, there is no project showing that Linux-VServer would eliminate LSM .. if that is what you mean 1165643229 Q * Zaki Ping timeout: 480 seconds 1165643829 J * hardwire ~hardwire@rdbck-7535.palmer.mtaonline.net 1165643844 M * hardwire heh 1165643847 M * hardwire I called MTA about my DSL 1165643847 M * Bertl wb hardwire! 1165643958 M * hardwire err 1165643960 M * hardwire wrong window 1165643963 M * hardwire hey ther Bertl 1165644034 M * hardwire hows teh life 1165644037 M * hardwire hows teh kids 1165644058 M * Bertl life is fine, no kids yet :) 1165644063 M * hardwire cats/dog? 1165644075 M * Bertl two cats, one dog, all of them are fine 1165644080 M * hardwire been walked? 1165644083 M * hardwire dogs love walks 1165644087 M * hardwire they are all woowoo! 1165644109 M * Bertl yeah, that's true ... luckily my parents care for the dog ... 1165644122 M * hardwire aww so yer just a little guy 1165644135 A * Bertl is more a cat person ... 1165644182 M * hardwire ok.. lemme ask this and this is conditional.. if you are over 25 and living with your parents you have no obligation to answer.. but we will know you are a weirdo.. 1165644185 M * hardwire how old are you? 1165644199 M * Bertl lol 1165644223 M * Bertl well, I'm 36, and I do not live with my parents .. but I visit them often :) 1165644237 M * hardwire aww.. so you aren't a little guy 1165644249 M * Bertl compared to the universe :) 1165644257 M * hardwire "my parents take care of the dog" made it sound like you live in the basement and they are closer to the doors to the outside. 1165644289 M * Bertl not at the moment, as I'm not visiting them :) 1165644335 M * Bertl so strictly speaking, it is not my dog ... 1165644343 M * hardwire gotcha 1165644363 M * Bertl but ... that wasn't your question, and I didn't imply that either 1165644374 M * Bertl although you might have thought I did :) 1165644415 M * marcfiu Bertl: yeah... what I am after is knowing whether Linux-Vserver (or, in general, a containerized Linux) would/should eliminate LSM. 1165644446 M * Bertl marcfiu: I think, LSM originally hat a slightly different purpose 1165644464 M * marcfiu Bertl: right 1165644464 M * hardwire I imply a lot of things 1165644470 M * hardwire say you are pretty bright for a little guy 1165644474 M * Bertl marcfiu: and while it was never really used, AFAIK, it can still be used for other things .. 1165644503 M * marcfiu where the "it" in "it can still be used for other things" refers to LSM? 1165644511 M * Bertl marcfiu: so, yes, most of the security aspects are interchangeable 1165644520 M * hardwire http://www.woot.com/ 1165644524 M * hardwire somebody buy me that 1165644547 M * hardwire Bertl: so did you pack up the cats when your parents kicked you out? 1165644589 M * Bertl hardwire: nope, nobody kicked me, and the cats didn't require to be packed up :) 1165644605 M * hardwire ok 1165644608 M * hardwire you know what sucks 1165644610 M * hardwire my DSL provider 1165644617 M * hardwire I have half a fucking t1 sync. 1165644628 M * hardwire you would think I could stream a 96kbps mp3 station in 1165644634 M * Bertl marcfiu: but I think the LSM was kind of ignored by most folks who wanted some security in the kernel 1165644639 M * hardwire but I have a man on the inside. 1165644646 M * hardwire who is gonna adjust and tweak 1165644649 M * marcfiu Bertl: agreed. 1165644660 M * Bertl marcfiu: most security enhancements do not use it because it is too complicated or broken or not detailed enough 1165644688 M * Bertl marcfiu: unfortunately mainline folks keep still arguing that it can be used for all kind of stuff and folks should better use it 1165644737 M * marcfiu Bertl: so LSM will hang around until something else comes along that generates a significant uptake 1165644754 M * Bertl hardwire: was on 64kbit dialup for the last three years .. now I enjoy reasonable transfer rates on aDSL 1165644772 M * hardwire Bertl: I bet the ISDN went straight to their IP provider 1165644787 M * hardwire instead of going through billions of internal "ooh look this makes things happy" hardware. 1165644805 M * Bertl marcfiu: I guess so, nevertheless, it would be quite interesting to evaluate actual users of LSM 1165644821 M * Bertl marcfiu: i.e. figure who _does_ use it atm 1165644858 M * marcfiu Bertl: NSA 1165644888 M * Bertl ah, good point 1165644894 M * marcfiu Bertl: and it is unlikely they will tell you how they use it. 1165644962 M * Bertl right 1165648779 J * Aiken__ ~james@tooax8-247.dialup.optusnet.com.au 1165649106 Q * Aiken_ Ping timeout: 480 seconds 1165649695 Q * mrrm Remote host closed the connection 1165649708 J * mrrm ~urkel@tor-irc.dnsbl.oftc.net 1165649716 M * Bertl wb mrrm! 1165649939 Q * marcfiu Ping timeout: 480 seconds 1165650124 Q * Adrinael Quit: Bleh 1165651406 J * Adrinael adrinael@st12-127.tky.hut.fi 1165653333 Q * nebuchadnezzar Quit: ERC Version 5.1.4 (IRC client for Emacs) 1165653437 J * nebuchadnezzar ~nebu@zion.asgardr.info 1165653524 M * Bertl daniel_hozac: ppc just finished to build the kernel, but I'm now off to bed, so the tool testing will have to wait till I get up again .. do not hesitate to release the 212 though, as it is probably the only way to get more testers :) 1165653541 M * Bertl have a good one everyone! cya later ... 1165653546 N * Bertl Bertl_zZ 1165653548 J * pmenier ~pmenier@ACaen-152-1-90-193.w86-208.abo.wanadoo.fr 1165653566 M * pmenier hello all 1165653598 M * pmenier last patch vs-2.1.1.4 doesn't work for me 1165653610 M * pmenier srvweb:/usr/local/sbin# vserver vweb1 enter 1165653610 M * pmenier vnamespace: vc_enter_namespace(): Invalid argument 1165653700 N * Bertl_zZ Bertl 1165653715 M * Bertl pmenier: what tools do you use? 1165653738 M * pmenier i use util-vservers-0.30.212-rc2 1165653738 M * Bertl pmenier: and what does testme.sh output? 1165653752 M * pmenier i did'nt try testme.sh 1165653768 M * Bertl okay, give it a quick spin, and let me know if anything fails 1165653788 M * pmenier ok i try it 1165653791 M * Bertl btw, 0.30.212-rc5 is there 1165653809 M * pmenier ok i install it 1165653970 M * pmenier here is the output of testme 1165653977 M * pmenier srvweb:/usr/local# ./testme.sh 1165653977 M * pmenier Linux-VServer Test [V0.16] Copyright (C) 2003-2006 H.Poetzl 1165653977 M * pmenier chcontext is working. 1165653977 M * pmenier chbind is working. 1165653977 M * pmenier Linux 2.6.19-vs2.1.1.4 #1 SMP Sat Dec 9 09:34:06 CET 2006 i686 1165653978 M * pmenier Ea 0.30.212 273/glibc (DSa) 1165653978 M * pmenier VCI: 0002:0102 273 03000407 (TbLgn) 1165653980 M * pmenier --- 1165653980 M * pmenier [000]# succeeded. 1165653982 M * pmenier [001]# succeeded. 1165653982 M * pmenier [011]# succeeded. 1165653984 M * pmenier [031]# succeeded. 1165653984 M * pmenier [101]# succeeded. 1165653986 M * pmenier [102]# succeeded. 1165653986 M * pmenier [201]# succeeded. 1165653988 M * pmenier [202]# succeeded. 1165654013 M * Bertl looks fine to me .. please check with rc5 if your issue remains 1165654041 M * pmenier ok i finish to compil kernel and install it 1165654092 M * Bertl I'm off to bed now .. if you encounter issues with that version, please contact daniel_hozac (who will be up in a few hours at most) 1165654104 M * Bertl have a good one! 1165654109 N * Bertl Bertl_zZ 1165654110 M * pmenier ok ok :-) Good night and thanks 1165654440 J * kir_home ~kir@213.152.157.70 1165654646 M * pmenier it works fine with util-vservers-0.30.212-rc5 1165654658 M * pmenier bye 1165654663 Q * pmenier Quit: KVIrc 3.2.0 'Realia' 1165655189 Q * yang Quit: KVIrc 3.2.4 Anomalies http://www.kvirc.net/ 1165655977 J * meandtheshell ~markus@85-125-231-228.dynamic.xdsl-line.inode.at 1165656083 Q * nebuchadnezzar Quit: ERC Version 5.1.4 (IRC client for Emacs) 1165656164 J * nebuchadnezzar ~nebu@zion.asgardr.info 1165656509 J * yang ~yang@yang.sponsor.oftc.net 1165657899 Q * fs Remote host closed the connection 1165657905 J * dna ~naucki@21-242-dsl.kielnet.net 1165657946 Q * Curus Ping timeout: 480 seconds 1165657960 J * Curus ~Curus@kbhn-vbrg-sr0-vl209-213-185-8-10.perspektivbredband.net 1165658430 J * fs fs@213.178.77.98 1165659394 Q * kir_home Quit: Ухожу я от вас 1165659404 J * DavidS ~david@85.125.165.34 1165659493 J * trippeh_ atomt@x.vx.no 1165659493 Q * trippeh Read error: Connection reset by peer 1165660378 J * pmenier ~pmenier@ACaen-152-1-90-193.w86-208.abo.wanadoo.fr 1165661535 J * iorikong ~PalPV3@218.81.239.134 1165661590 M * iorikong f 1165661724 Q * iorikong Remote host closed the connection 1165662457 M * eyck is it normal for debian's 2.6.18-3-vserver to OOPS everytime a vserver is shut down (looks like right after reboot request from guest) ? 1165662503 M * waldi nope 1165662524 M * eyck so, this 100% repeatable, 1165662660 M * matti Hi eyck :) 1165662673 M * eyck hi matti 1165663196 M * eyck anyhoo, I backed off, and it's now working nice. 1165664415 Q * Aiken__ Quit: Leaving 1165665430 Q * rob-84x^ Quit: That's it for today 1165665440 J * rob-84x^ rob@submarine.ath.cx 1165665658 Q * ensc Killed (NickServ (GHOST command used by ensc_)) 1165665668 J * ensc ~irc-ensc@p54B4D1C5.dip.t-dialin.net 1165666950 J * yarihm ~yarihm@84-74-17-6.dclient.hispeed.ch 1165667476 M * daniel_hozac eyck: yes it is. 1165668390 Q * nebuchadnezzar Remote host closed the connection 1165668529 J * nebuchadnezzar ~nebu@zion.asgardr.info 1165668885 M * daniel_hozac ah, you're here pmenier. i just replied to your email. 1165668903 M * daniel_hozac pmenier: do you have a way to reproduce this every time? 1165670304 J * Piet hiddenserv@tor.noreply.org 1165670764 J * Zaki ~Zaki@88.213.59.61 1165670772 J * sebastian ~info@pD957E17B.dip.t-dialin.net 1165671175 M * pmenier sorry i'm back again 1165671208 M * pmenier when i say loose their context i type : vserver-sta and the nmae of vserver disappear 1165671235 M * pmenier when it's ok i got : 1165671235 M * pmenier srvweb:/var/log# vserver-stat 1165671235 M * pmenier CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME 1165671235 M * pmenier 0 82 986.7M 220.4M 4m15s84 1m41s00 2h49m11 root server 1165671235 M * pmenier 20 18 424.1M 117.1M 0m10s24 0m01s12 2h48m00 vweb1 1165671236 M * pmenier 21 15 729.1M 107M 0m02s70 0m01s10 2h05m32 vweb2 1165671236 M * pmenier 22 10 63.6M 22.4M 0m00s50 0m00s50 2h05m14 vweb3 1165671259 M * pmenier when it's out : srvweb:/var/log# vserver-stat 1165671259 M * pmenier CTX PROC VSZ RSS userTIME sysTIME UPTIME NAME 1165671259 M * pmenier 0 82 986.7M 220.4M 4m15s84 1m41s00 2h49m11 root server 1165671259 M * pmenier 20 18 424.1M 117.1M 0m10s24 0m01s12 2h48m00 1165671259 M * pmenier 21 15 729.1M 107M 0m02s70 0m01s10 2h05m32 1165671260 M * pmenier 22 10 63.6M 22.4M 0m00s50 0m00s50 2h05m14 1165671429 M * daniel_hozac hmm, that's strange. 1165671519 M * daniel_hozac but vserver-stat uses a lot of old APIs, so it's possible it's relying on /proc. 1165671742 M * pmenier i don't know.... i was replaying to your mail .I have Recompil util-vserver with enable-apis=NOLEGACY 1165671762 M * pmenier it works... for the moment 1165671897 M * pmenier i've no way to reproduce this. It appears suddenly this morning. The machine was up since 1 day without any problems 1165672573 M * pmenier I must go away a few hours. I would be back ~ 19pm 1165672589 Q * pmenier Quit: KVIrc 3.2.0 'Realia' 1165672643 J * pmenier ~pmenier@ACaen-152-1-90-193.w86-208.abo.wanadoo.fr 1165675177 J * bonbons ~bonbons@ip-83-134-145-130.dsl.scarlet.be 1165675272 J * Val ~val@v41.org 1165675331 M * Val Hi 1165675388 M * Val did someone experience automount problem inside a guest (using vnamespace from main host) since 2.6.18-vs2.0.2.2-rc8 ? 1165675703 M * Val for exemple : vnamespace -e `cat /etc/vservers/guest/context` automount -v /vservers/guest/home file /etc/auto.guest-home 1165675745 M * Val in /etc/auto.guest-home : test -fstype=auto,rbind :///etc/test 1165675760 M * Val and : val -fstype=auto,rbind :///home/val 1165675782 M * Val /etc/test is in / filesystem 1165675792 M * Val /home is a mount point 1165675837 M * Val inside the guest 'cd /home/test' work, 'cd /home/val' doesn't 1165675845 M * Val +s 1165675882 M * Val automount says it does not find /home/val 1165675907 M * Val util-vserver 0.30.211-6 (debian) 1165675931 M * Val kernel 2.6.18-vs2.0.2.2-rc8 1165676535 M * daniel_hozac /home is most likely unmounted in the namespace. 1165676561 M * daniel_hozac any particular reason you're using automount instead of just putting the mounts in the fstab? 1165676637 M * Val yes, to provide generic configuration using autofs for multiple guests 1165676646 M * Val it worked well 1165676694 M * Val but now it seems automount is unable to mount mounted filesystem using rbind 1165676754 M * daniel_hozac yes, because of namespace cleanup. 1165676759 M * Val arg 1165676778 M * Val is it a way to get this working ? 1165676778 M * daniel_hozac you can either disable it, or add the required filesystems to /etc/vservers/.defaults/namespace-cleanup-skip 1165676793 M * Val oh 1165676863 M * Val ok let's try 1165676865 M * Val thnaks 1165676938 M * Val /etc/vservers/.defaults/namespace-cleanup-skip is not on the great flower page ? 1165677044 M * Val what is the syntax of the file /etc/vservers/.defaults/namespace-cleanup-skip ? 1165677053 M * Val one path per line ? 1165677112 Q * yarihm Ping timeout: 480 seconds 1165677126 M * daniel_hozac yes. 1165677146 M * daniel_hozac it's a new feature in 0.30.212, and since i haven't released it yet, the flower page hasn't been updated. 1165677154 M * Val arg 1165677172 M * Val ok i'll try nonamespace-cleanup flag in guest config 1165677183 M * daniel_hozac well, namespace-cleanup-skip should work fine. 1165677197 M * Val in the guest config ? 1165677208 M * daniel_hozac in .defaults if you want it to apply for all guests. 1165677234 M * daniel_hozac but you can have it in the guest config as well. 1165677252 M * Val /etc/vservers/.defaults# cat namespace-cleanup-skip 1165677257 M * Val /all 1165677276 M * daniel_hozac you have a filesystem mounted at /all? 1165677281 M * Val yup 1165677296 M * Val a luks encrypted fs 1165677300 M * daniel_hozac should work fine then. 1165677326 M * daniel_hozac (if that's the mount you're expecting to be there) 1165677357 M * Val the mount autofs will try is /all/test -> /home/test 1165677365 M * Val rbind 1165677387 M * Val but it doesn't :'( 1165677410 M * daniel_hozac and you restarted the guest? 1165677415 M * Val yes 1165677437 M * daniel_hozac so does vnamespace -e cat /proc/mounts list /all? 1165677438 M * Val i try now with the same file in the guest config dir 1165677487 M * Val yes, automount is listed on /all 1165677502 M * Val automount(pid25263) /var/lib/vservers/vdumb/home autofs rw,fd=5,pgrp=25263,timeout=30,minproto=2,maxproto=4,indirect 0 0 1165677512 M * daniel_hozac hmm? you're automounting automounts? 1165677520 M * Val no 1165677536 M * Val /all is mounted with main host /etc/fstab 1165677552 M * daniel_hozac and it is listed in the guest namespace's /proc/mounts? 1165677563 M * Val no 1165677573 M * Val only the automount proc 1165677585 M * Val like before... nothing more 1165677630 M * daniel_hozac well, vnamespace -e cat /proc/mounts should list quite a few things that aren't listed inside the guest. 1165677640 M * Val in one other host, with the same conf. but not an updated kernel things like autofs /home user dir works well 1165677654 M * daniel_hozac same utils? 1165677658 M * Val same 1165677662 M * Val only -rc6 1165677671 M * Val (for the kernel) 1165677680 M * daniel_hozac so you have 0.30.211-6 on both hosts? 1165677689 M * Val yes 1165677695 M * Val kernel differ 1165677705 M * daniel_hozac and you have restarted the guests since upgrading to that on both? 1165677711 M * Val no ! 1165677740 M * Val with the problem i get now, i will not reboot the other until i get a solution 1165677745 M * daniel_hozac namespace cleanup was enabled by default in 0.30.211-4 or so. 1165677754 M * Val hum 1165677765 M * daniel_hozac (when it was upgraded to 0.30.212-rcX) 1165677778 M * Val i'll try with an older package of util-vserver 1165677791 M * daniel_hozac or you could just disable namespace cleanup altogether. 1165677807 M * Val mmm 1165677824 M * Val namespace-cleanup-skip does not work 1165677832 M * daniel_hozac it should. 1165677847 M * Val so i put a file : nonamspace-cleanup ? 1165677855 M * daniel_hozac yep. 1165677867 M * Val an emtpy one 1165677881 M * daniel_hozac doesn't matter, as long as the file exists. 1165677906 M * Val okay 1165677942 M * Val guest restarted, autofs doesn't work 1165677951 M * Val same problem 1165677955 M * Val ... 1165678127 M * Piet is there something like traceroute which works without raw socket access capability? 1165678138 J * yarihm ~yarihm@84-74-17-6.dclient.hispeed.ch 1165678150 M * Piet 'traceroute: raw socket: Operation not permitted' 1165678223 M * daniel_hozac tracert? 1165678229 M * daniel_hozac uh, tracepath, i mean. 1165678365 M * daniel_hozac Val: could you paste cat /proc/mounts; vnamespace -e cat /proc/mounts to paste.linux-vserver.org? 1165678377 M * Val ok 1165678483 J * dreamind apwdsl@p548ACB46.dip0.t-ipconnect.de 1165678497 M * dreamind Hi folks :) 1165678543 M * Val daniel_hozac : http://paste.linux-vserver.org/746 1165678582 M * dreamind hi daniel_hozac :) 1165678587 M * daniel_hozac Val: ok, and now a vserver --debug vdumb start 1165678594 M * daniel_hozac hello dreamind. 1165678608 M * Snow-Man daniel_hozac: Trying that patch from yesterday, btw. 1165678689 M * Piet thanks daniel 1165678745 M * daniel_hozac Snow-Man: ok, thanks. 1165678778 M * Snow-Man Looks like that fixed it. 1165678781 M * Snow-Man Linux aratan 2.6.18-3cacctfix-vserver-amd64 #1 SMP Fri Dec 8 17:25:57 EST 2006 x86_64 GNU/Linux 1165678784 M * Snow-Man No oopsen. :) 1165678808 M * daniel_hozac great! 1165678918 M * Snow-Man yah, definitely looks good. 1165678950 M * Snow-Man Gonna have a -rc9 soon? :) 1165678974 M * daniel_hozac i think so, there are quite a few other fixes pending for 2.0. 1165678986 M * daniel_hozac (like fixing the CPU scheduler) 1165679001 M * Snow-Man waldi: Would be really nice to have the oopsen fix in the next Debian kernel... :) 1165679172 M * Val daniel_hozac : http://paste.linux-vserver.org/747 1165679426 M * Val daniel_hozac : /etc/vservers/vdumb/scripts/*/vautomount is a script that do 'vnamespace -e ... automount .../home file /etc/auto.-home' 1165679481 M * daniel_hozac could you do a vserver vdumb stop; vserver --debug vdumb start? your trace got cut off just as it was getting interesting... 1165679502 M * Val arg 1165679503 M * Val ok 1165679606 Q * sebastian 1165679660 M * Val daniel_hozac : http://paste.linux-vserver.org/748 1165679770 Q * dreamind Quit: dreamind 1165679818 M * daniel_hozac Val: well, you don't have /etc/vservers/.defaults/nonamespace-cleanup or /etc/vservers/vdumb/nonamespace-cleanup 1165679867 M * Val oh, yes i removed it ... sorry 1165679907 M * Val i restart then... 1165679914 M * daniel_hozac but you're right .defaults/namespace-cleanup-skip doesn't work. 1165679921 M * daniel_hozac bad assumption on my part. 1165679926 M * Val :] 1165679946 M * Val well 1165679949 M * Val restarted 1165679953 M * Val autods works 1165679957 M * Val autoFs 1165679981 M * Val many thanks 1165680006 M * Val i will wait for new util-vserver tools to remove nonamespace-cleanup and test namespace-cleanup-skip 1165680386 M * daniel_hozac http://svn.linux-vserver.org/projects/util-vserver/changeset/2420 should fix it. 1165680483 M * Val ok 1165680487 M * Val thx 1165680703 J * shedi ~siggi@inferno.lhi.is 1165680787 Q * yarihm Quit: Leaving 1165680934 J * Blissex ~Blissex@82-69-39-138.dsl.in-addr.zen.co.uk 1165681463 T * daniel_hozac http://linux-vserver.org/ | latest stable 2.02.1, exp 2.02.2-rc8, devel 2.1.1.3, (2.6.19) 2.2.0-rc3, 2.1.1.4, stable+grsec 2.0.2.1, devel+grsec 2.1.1 | util-vserver-0.30.212 | libvserver-1.0.2 & vserver-utils-1.0.3 | He who asks a question is a fool for a minute; he who doesn't ask is a fool for a lifetime -- share the gained knowledge on the Wiki, and we'll forget about the minute ;) 1165681476 M * daniel_hozac util-vserver 0.30.212 released. 1165681498 M * derjohn daniel_hozac, thx ! ! 1165681520 M * derjohn micah, time for a new upstream release ... 1165681536 M * daniel_hozac micah, ensc, Hollow, phreak``, etc: if you could update your packages, i'd appreciate it :) 1165681814 M * Hollow daniel_hozac: yeah, will do so shortly :) 1165687482 M * daniel_hozac ... and the channel died :) 1165687507 M * pmenier daniel_hozac: i just try util-vservers 0.30.212. It seems vservers are stable 1165687672 M * daniel_hozac good, thanks. 1165687996 M * micah daniel_hozac: will do 1165688027 J * lilalinux ~plasma@80.69.41.2 1165689641 N * Bertl_zZ Bertl 1165689648 M * Bertl morning folks! 1165689657 M * daniel_hozac morning Bertl! 1165689691 M * Bertl pmenier: everything working as expected now? 1165689733 M * daniel_hozac except for the proc bug, i guess. 1165689762 M * Bertl hmm? fix should be in, no? 1165689777 M * daniel_hozac well, doesn't seem to be fixed. 1165689780 M * daniel_hozac (see the ML) 1165689812 M * daniel_hozac though it's not an oops now, just a warning. 1165689830 M * Bertl precisely, it happens, as expected, but it just warns 1165689847 M * daniel_hozac hmm, that's expected? 1165689861 M * Bertl I would have been alarmed if that warning hadn't shown up 1165689882 M * Bertl because then, the last issue would have been unsolved 1165689886 M * daniel_hozac could only happen if the context vanished between proc_virtual_lookup and proc_vx_read_info, right? 1165689912 M * Bertl yep, so I think we have a race with revalidate and/or we have a missing check 1165689955 M * daniel_hozac maybe we should get the [vn]x_info somewhere? 1165689978 M * daniel_hozac or at least start returning -ENOENT on !vxi. 1165690066 M * Bertl nah, getting the vx_info is not an option 1165690077 M * Bertl it would keep the proc entry around forever 1165690089 M * daniel_hozac hmm, true. 1165690106 M * Bertl the ENOENT is the check I meant 1165690124 M * Bertl but I still have to analyze the traces 1165690229 M * Bertl what is the vserver-stat failing issue? 1165690239 M * Bertl is that really related? 1165690243 M * daniel_hozac i'm not sure. 1165690271 M * Bertl it uses the proc information, no? 1165690310 M * daniel_hozac not for the xids, i think. 1165690328 M * Bertl so how does it get the name? 1165690395 M * daniel_hozac it should do the usual xid to name mapping, by getting the vx_info->name and comparing the run file. 1165690425 M * Bertl okay, and the vx_info-> name is done via proc or syscall? 1165690431 M * daniel_hozac syscall. 1165690440 M * daniel_hozac vc_get_vhi_name(VHIN_CONTEXT) 1165690459 M * Bertl hmm, maybe that is wrong then? 1165690475 M * daniel_hozac what would be messing it up? 1165690485 M * Bertl UTS namespaces 1165690493 M * daniel_hozac ah, hm... 1165690539 M * Bertl does retrieving uts information from outside work as expected 1165690555 M * daniel_hozac yeah. 1165690834 Q * Val Quit: thx for the help... bye bye 1165691079 M * Bertl pmenier: okay, can you trigger the issue? 1165691262 M * Bertl daniel_hozac: do you have vserver debugging compiled in? 1165691330 M * daniel_hozac i do, but i haven't been able to trigger the issue for a while 1165691349 M * daniel_hozac not since i added my debugging printk. 1165691360 M * Bertl could you enable the switch tracing and paste me the lines a vserver-stat generates? 1165691423 M * daniel_hozac http://paste.linux-vserver.org/749 1165691428 M * Bertl tx 1165691444 M * daniel_hozac i have to go afk for a bit, i'll be back soon. 1165691454 M * Bertl ok, np, have fun! 1165693342 M * pmenier Bertl I'm back again. can i help you ? You want a trace of something , 1165693374 M * Bertl not atm, tx 1165693532 M * pmenier bye everybody 1165693542 Q * pmenier Quit: KVIrc 3.2.0 'Realia' 1165694593 Q * DavidS Quit: Leaving. 1165696212 M * Bertl daniel_hozac: sorry, but I'm unable to find the 0.30.212 release ... 1165696270 M * Bertl linux-vserver.org doesn't mention it, there is no ML announcement, people.* doesn't have it ... where is it? :) 1165696355 M * daniel_hozac hmm? 1165696372 M * daniel_hozac linux-vserver.org/Downloads has it, i sent an ML announcement, it's on ftp.linux-vserver.org... 1165696389 M * Bertl funny, ftp is empty for me 1165696390 M * daniel_hozac i suppose i should add an item to the wiki news section. 1165696397 M * daniel_hozac ? 1165696409 M * daniel_hozac http://ftp.linux-vserver.org/pub/utils/util-vserver/ 1165696435 M * Bertl hmm, now I get something ... strange 1165696472 M * Bertl okay, I guess I need more caffeine :) 1165696482 M * daniel_hozac ah, that reminds me. should i add 2.1.1.4 as the recommended 2.1.1 release on the wiki? 1165696500 M * Bertl yes, please, same goes for 2.2.0 and 2.3.0 1165696500 M * daniel_hozac i.e. as the entry for "The latest version of the Linux-VServer development patch is" 1165696615 M * daniel_hozac should i put the 2.3.0* patches in the testing subdirectory? 1165696678 M * Bertl yes, I think that can't hurt 1165696850 M * Bertl nice, so the gentoo build scripts have been fed back? 1165696881 M * Bertl micah, daniel_hozac: what about the debian newvserver stuff? any progress there? 1165696898 M * daniel_hozac yeah, Hollow has been integrating pretty much all of the Gentoo scripts. 1165696922 M * daniel_hozac ah, i forgot about that! 1165696925 M * Bertl does that mean that I can build gentoo guests on non gentoo distros too? or is that just wishful thinking? 1165696939 M * daniel_hozac well, that's been possible for a while. 1165696945 M * daniel_hozac Gentoo is just a tarball, after all. 1165696977 M * Bertl cool, does it work via network? i.e. automatically? 1165696989 M * daniel_hozac no, you have to get the tarball yourself. 1165697004 M * daniel_hozac then build -m template ... -- -d gentoo -t stage3... 1165697034 M * Bertl i.c. any plans to add 'debian' like network bootstrapping? 1165697045 M * daniel_hozac i don't know if that's at all possible with Gentoo 1165697072 M * daniel_hozac Hollow, phreak``? 1165697083 M * Bertl hmm, sounds like all it would require is a 'curl/wget stage3' then the template install? 1165697112 M * daniel_hozac well, sure, but then you'd have to figure out the stage3 URL somehow. 1165697121 M * Bertl btw, what about caching network install stuff somehow? 1165697149 M * Bertl i.e. allow to specify a cache dir for network based installs 1165697176 M * daniel_hozac well, that would require support in the underlying tools. 1165697196 M * daniel_hozac but i know at least yum and apt have that, so i guess it would be worth looking in to. 1165697219 M * Bertl I think debootstrap should have it too, or if not, it would be worth adding that 1165697226 M * Bertl (for the debian folks, of course) 1165697264 M * daniel_hozac i suppose the "easier" solution is an apt-proxy... 1165697286 M * Bertl yes, but that requires setting up stuff and such 1165697309 M * Bertl and I'm not sure somebody using suse or redhat will manage to do that properly 1165697350 M * Bertl hmm, did I miss something, or do the tool now require the docu to be built? 1165697379 M * Bertl (i.e. I would love to avoid that 200MB additional packages just for building the rpm :) 1165697501 M * daniel_hozac hmm, shouldn't be required, i don't think. 1165697508 M * daniel_hozac at least, i haven't changed anything. 1165697515 M * Bertl well, it says: 1165697523 M * Bertl doxygen is needed by util-vserver-0.30.212-0.src 1165697523 M * Bertl tetex-latex is needed by util-vserver-0.30.212-0.src 1165697543 M * Bertl and if I build with nodeps (those are explicit dependancies) 1165697549 M * Bertl then it fails with: 1165697558 M * Bertl cp: cannot stat `lib/apidoc/html': No such file or directory 1165697607 M * Bertl well, I probably can touch that (or create a dir) to satisfy the build process :) 1165697705 M * Bertl just commented it out, should work ... 1165697727 M * Bertl but some --without-doc would be nice 1165697747 M * Bertl well, it would be even nicer to add the documentation to the package 1165697791 M * Bertl (at least I don't think there are dependancies on the actual build configuration at build time, no?) 1165697813 M * daniel_hozac nope. 1165697827 M * daniel_hozac where is it coming from? 1165697882 M * Bertl %files devel -f %name-devel.list 1165697882 M * Bertl %defattr(-,root,root,-) 1165697882 M * Bertl %doc lib/apidoc/latex/refman.pdf 1165697882 M * Bertl %doc lib/apidoc/html 1165697896 M * daniel_hozac ah, the spec file. 1165697920 M * Bertl BuildRequires: doxygen tetex-latex 1165697943 M * daniel_hozac right. 1165697970 M * daniel_hozac i haven't touched the specfile though. 1165698019 M * Bertl yeah, I remember 0.30.211 bithing about that too 1165698027 M * Bertl *bitching 1165698083 M * Bertl after update to 0.30.212 on x86_64 I get: 1165698089 M * Bertl * The configured vshelper '/sbin/vshelper' does not match the 'vshelper' script of the util-vserver package 1165698111 M * daniel_hozac update by RPM? 1165698121 M * Bertl yep, maybe a missing post install script? 1165698148 M * daniel_hozac hmm, it should be contained in the package. 1165698171 M * daniel_hozac rpm -qf /sbin/vshelper should list as belonging to util-vserver. 1165698186 M * daniel_hozac where does /sbin/vshelper point to? 1165698191 M * daniel_hozac s/where/what/ 1165698237 M * Bertl sorry, I already did: 1165698238 M * Bertl echo "/lib/util-vserver/vshelper" >/proc/sys/kernel/vshelper 1165698244 M * Bertl as suggested by the tools 1165698251 M * Bertl and that fixed it 1165698261 M * daniel_hozac humm, /lib/util-vserver? 1165698266 M * daniel_hozac with the RPM? 1165698280 M * Bertl yeah, but rpm -qf /sbin/vshelper gives the RPM 1165698296 M * daniel_hozac the RPM should install to /usr, i'm quite sure. 1165698333 M * Bertl here is something even more interesting: 1165698334 M * Bertl /sbin/vshelper -> ../usr/lib64/util-vserver/vshelper 1165698352 M * daniel_hozac hmm, i think you have two installs. 1165698356 M * Bertl so I probably have a mix of rpm and normal install 1165698365 M * Bertl could we add something to catch that RSN? 1165698383 M * daniel_hozac how? 1165698384 M * Bertl some kind of 'consistency check' or whatever 1165698396 M * daniel_hozac well, we already have the sanity checking, so it's just a matter of adding it. 1165698407 M * daniel_hozac i'm just not sure how to detect it. 1165698407 M * Bertl maybe add a version/release number to each function file/etc 1165698420 M * Bertl could be generated at compile time, some hash or so 1165698424 M * Bertl like for the kernel modules 1165698456 M * Bertl if that fails for any file, bail out with a version mismatch 1165698476 M * Bertl I think that would help with many issues reported here 1165698498 M * daniel_hozac well, once you're in the vserver script, everything is using absolute paths. 1165698531 M * Bertl so what actually happened here? 1165698547 M * daniel_hozac the RPM install overwrote your /sbin/vshelper link 1165698563 M * daniel_hozac AFAICT that's the only inconsistency in your install. 1165698569 M * daniel_hozac and it was caught. 1165698596 M * Bertl okay 1165698623 M * daniel_hozac if it hadn't been overwritten, you would've happily gone along using your older utils :) 1165698635 M * Bertl maybe add a short notice at the top like this: 1165698678 M * Bertl * The configured vshelper '/sbin/vshelper' does not match the 'vshelper' script of the util-vserver package. maybe you have two competing versions installed? 1165698708 Q * transacid Ping timeout: 480 seconds 1165698710 M * Bertl because I remember now, I already encountered this last time 1165698728 M * Bertl and I wasn't able to get to that conclusion this time ... 1165698809 M * Bertl ah, uninstallation of util-vserver-sysv-0.30.212-1mdk and util-vserver-legacy-0.30.212-1mdk fails 1165698818 M * Bertl will force now and try again 1165698821 M * daniel_hozac hmm? 1165698832 M * daniel_hozac why? 1165698878 M * Bertl error: %preun(util-vserver-legacy-0.30.212-1mdk.x86_64) scriptlet failed, exit status 1 1165698888 M * Bertl error reading information on service vservers-legacy: No such file or directory 1165698900 M * Bertl and similar ... trying now again from scratch 1165699012 M * Bertl works now, so I guess it was confused by the mix 1165699043 M * daniel_hozac okay. 1165699044 M * Bertl yeah, probably the manual uninstall removed those files 1165699065 M * Bertl and the rpm couldn't uninstall because they were not there anylonger 1165699075 M * Bertl (strange behaviour IMHO, but probably okay) 1165699079 M * daniel_hozac yeah, sounds likely. 1165699166 M * Bertl btw, you use bash as shell too, right? 1165699259 J * transacid ~transacid@transacid.de 1165699284 M * Bertl wb transacid! 1165699339 M * transacid thx 1165699342 Q * bonbons Quit: Leaving 1165699349 M * transacid fucking screen-key-bindings 1165699362 M * transacid always this C-a s 1165699479 M * Hollow Bertl, daniel_hozac: well, building own gentoo guests from source directly is a pita IMO, but probably the vserver-build script could download the stage on demand from one of the mirrors... 1165699506 M * daniel_hozac Bertl: yeah, it's all bash here :) 1165699524 M * Bertl daniel_hozac: okay, do you use reverse history search? 1165699531 M * daniel_hozac yeah, quite often. 1165699571 M * Bertl so do I, but everywhere I go, I have to re-bind the forward search, because C-s (the default) doesn't work ... 1165699608 M * Bertl was that just a stupid idea of the readline folks? or ignorance on the terminal sequences? 1165699615 M * Bertl or am I just missing some magic? 1165699645 M * daniel_hozac that has never worked for me either. 1165699661 M * Bertl yeah, well, C-q and C-s are flow control 1165699685 M * Bertl so that is not quite unexpected, I typically rebind it to C-f as that is not really used in bash 1165699700 M * daniel_hozac makes sense. 1165699745 M * Bertl okay, so probably just ignorance, no magic involved :) 1165699755 M * daniel_hozac probably :) 1165700629 Q * Piet Remote host closed the connection 1165700629 J * Piet_ hiddenserv@tor.noreply.org 1165700927 M * Bertl daniel_hozac: hmm, why do I get a debug rpm with x86_64 but not with ppc? 1165700956 M * Bertl (not that I would have used it yet) 1165701057 M * Bertl ppc guest startup and shutdown works fine 1165701139 M * Bertl daniel_hozac: ah, another thing which seems to be still 'unsolved', when I forget to shutdown the guest, I will definitely get an unclean shutdown, followed by a complete disk check on reboot 1165701173 M * Bertl you remember we talked about the vkillall just before shutdown? 1165701756 Q * duckx Remote host closed the connection 1165701822 J * duckx ~Duck@tox.dyndns.org 1165702361 M * Bertl daniel_hozac: hmm, seems I'm getting closer to the proc issue now 1165702379 Q * matti Remote host closed the connection 1165702381 J * matti matti@linux.gentoo.pl 1165702781 M * Bertl yep, this triggers it (almost reliably) 1165702790 M * Bertl chcontext --xid 42 -- sleep 10 & sleep 2 && cat /proc/virtual/*/status 1165702814 Q * lilalinux Remote host closed the connection 1165702836 M * Bertl I suspect the lookup avoids the actual dir lookup, so the xid is not propagated ... have to check that with the fd code, as I copied the code from there 1165705133 Q * dna Quit: Verlassend 1165705170 J * mrrm_ ~urkel@tor-irc.dnsbl.oftc.net 1165705305 Q * mrrm Remote host closed the connection 1165705940 M * Bertl okay, http://vserver.13thfloor.at/Experimental/delta-proc-fix02.diff fixes the proc issues for me 1165706078 M * Snow-Man Bertl: -rc9 soon? :) 1165706135 M * Bertl something like that :) will upload updated versions soon 1165706172 M * Snow-Man Well, being the selfish person I am, my main concern is getting the oopsen fix released which was affecting me. :) 1165706197 M * Bertl i.c. :) 1165706355 M * daniel_hozac 2.0 needs some love though. 1165706387 M * Bertl yeah, will focus on that now ... 1165706411 M * Bertl daniel_hozac: do you have any idea how vnamespace could become a zombie? 1165706440 M * daniel_hozac Bertl: so the idea is that [xn]id_array is overwritten by another thread? 1165706453 M * Bertl nah, not at all, much simpler 1165706464 M * Bertl the xid is consumed by the string conversion :) 1165706490 M * daniel_hozac oh! lol 1165706493 M * Bertl (one of the most stupid bugs I ever saw :) 1165706508 M * daniel_hozac and yet so hard to spot :) 1165706520 M * daniel_hozac re: vnamespace, no idea, that seems really strange. 1165706556 M * Bertl okay, we have some more details on the vnamespace 1165706567 M * Bertl a manual guest restart gets rid of the zombie 1165706591 M * Bertl it could be related to older tools, ntrs, what version is that? 1165706620 M * Bertl 210 and 211 if I got that right 1165706653 M * daniel_hozac what owns the zombie? 1165706655 M * Bertl ntrs: you execute a restart from a script, yes? 1165706659 M * daniel_hozac i mean, what is creating it? 1165706671 M * ntrs yes, from a script 1165706707 M * Bertl and it is restart, not stop and start, yes? 1165706727 M * daniel_hozac though restart is basically stop and start anyway. 1165706756 M * ntrs yes, it is restart 1165706766 M * Bertl it might be related to 'stop' being killed from inside? 1165706791 M * daniel_hozac hmm, it shouldn't be. 1165706795 M * ntrs I tried the script again and it produced a zombie right away 1165706815 M * Bertl daniel_hozac: well, you know the /usr/lib64/util-vserver/vserver.stop: line 85: 24659 Killed 1165706827 M * ntrs the guest is restarted though and seems to be working fine. 1165706834 M * daniel_hozac Bertl: right. 1165706836 M * Bertl daniel_hozac: sysv init, kill scripts inside the guest 1165706850 M * daniel_hozac AFAIK that error just means that the script the utils run was killed. 1165706862 M * ntrs I do get this when even manually restarting the guest 1165706864 M * ntrs /usr/lib64/util-vserver/vserver.stop: line 85: 24659 Killed "${NICE_CMD[@]}" ${USE_VNAMESPACE:+$_VNAMESPACE --enter "$S_CONTEXT" -- } $_VCONTEXT $SILENT_OPT --migrate --chroot --xid "$S_CONTEXT" -- "${INITCMD_STOP[@]}" 1165706880 M * Bertl daniel_hozac: yeah, but let me ask the critical question, who reaps the script? 1165706892 J * derjohn2 ~aj@dslb-084-058-200-047.pools.arcor-ip.net 1165706896 M * ntrs the script is run as root 1165706898 M * daniel_hozac child_reaper? 1165706925 J * Aiken ~james@tooax8-230.dialup.optusnet.com.au 1165706940 M * Bertl wb derjohn2! morning Aiken! 1165707103 M * Aiken good morning 1165707185 M * derjohn2 Bertl, nice to see you! 1165707210 M * Bertl derjohn2: the pleasure is mine ... 1165707220 M * Bertl Aiken: any progress with the alpha issue? 1165707245 Q * Blissex Remote host closed the connection 1165707309 M * daniel_hozac Bertl: so if i'm reading this patch correctly... shouldn't this be a pretty reproducible crash? 1165707319 M * daniel_hozac s/crash/WARN_ON/ 1165707391 M * Aiken I got sick of compiling kernels and did something else for a couple of days 1165707401 M * Bertl daniel_hozac: yeah, see my example :) 1165707404 M * Aiken I booted it a few min ago to get the oops 1165707420 M * Bertl Aiken: i.c. 1165707468 M * daniel_hozac Bertl: ah, got it... so the * is what does it, right? 1165707494 M * daniel_hozac (my reproducers used the xid) 1165707531 M * Aiken anyone else getting anything like this when trying to update from the linus tree? http://paste.linux-vserver.org/755 1165707584 M * Bertl daniel_hozac: the whole thing got a little obfuscated by the fact the the reporter uses a legacy enabled kernel 1165707603 M * ntrs Bertl, are we talking about the zombie vnamespace still? 1165707605 M * ntrs I got lost 1165707611 M * Bertl daniel_hozac: here, with recent tools, the guest name does not disappear, but it cannot be shut down 1165707630 M * Bertl ntrs: nah, we are talking three issues here right now 1165707640 M * ntrs ok I thought so. 1165707656 M * daniel_hozac Bertl: hmm, it vanishes as expected for me. 1165707659 M * ntrs I would really appreciate a fix to this vnamespace zombie issue. 1165707663 M * daniel_hozac (if we're still talking the proc issue) 1165707702 M * Bertl ntrs: let's create another one (zombie) and check the contents of /proc//vinfo and ninfo 1165707716 M * ntrs Bertl, I have a zombie ready 1165707733 M * Bertl also try to modify the script so that it does the shutdown (stop) part, but not the startup 1165707741 M * ntrs [root@edwards ~]# cat /proc/18344/vinfo 1165707741 M * ntrs XID: 0 1165707741 M * ntrs [root@edwards ~]# cat /proc/18344/ninfo 1165707741 M * ntrs NID: 0 1165707749 M * Bertl would be interesting to see if that keeps a context around too 1165707771 M * Bertl ah, also check /proc//status then, but please upload to 1165707780 M * Bertl paste.linux-vserver.org or pm me 1165707869 M * daniel_hozac well, if it kept the context around, starting it should fail. 1165707874 M * daniel_hozac ntrs: what kernel is that? 1165707889 M * Bertl 2.1.1.2.3+ 1165707894 M * ntrs 2.6.17.13 with 2021 1165707903 M * ntrs same thing happens with 2123 1165707906 M * ntrs 2113 1165707920 M * Bertl (the 2.1.1.3 is actually a 2.1.1.2.3+) 1165707933 M * ntrs ok 1165707950 M * Bertl but 2.6.17.13 is interesting 1165707965 M * ntrs I can move to the other machine 1165707969 M * ntrs with the latest kernel and patch 1165708013 M * Bertl I don't think that will change anything 1165708072 M * Bertl I think we first need to find a way to reproduce that with a simple script, do you think that works from bash or so? 1165708108 M * ntrs Bertl, the script is a one liner 1165708129 M * Bertl in php? or bash? 1165708142 M * ntrs bash or sh 1165708179 M * Bertl okay, and when you run that script manually on a login shell, you get a zombie? 1165708308 M * ntrs Bertl, I will check now 1165708379 M * Bertl okay, if that works, try to reduce it a little further, by removing the condition, and replacing the name with a fixed name 1165708741 Q * FireEgl Remote host closed the connection