1161216275 Q * DreamerC_ Quit: leaving
1161216296 J * DreamerC ~dreamerc@61-217-226-241.dynamic.hinet.net
1161216692 M * derjohn2 Bertl, n8 ... the rc41 debian packahge just building up ... till tomorrow !
1161216711 M * Bertl okay, cya!
1161217980 P * litage Leaving
1161219395 Q * Aiken Quit: Leaving
1161220545 J * Aiken ~james@tooax8-128.dialup.optusnet.com.au
1161222383 Q * Piet_ Quit: Piet_
1161223150 J * Aiken_ ~james@tooax6-175.dialup.optusnet.com.au
1161223476 Q * Aiken Ping timeout: 480 seconds
1161224886 J * marcfiu ~mef@c-68-39-177-97.hsd1.nj.comcast.net
1161224900 M * marcfiu Hello
1161224930 M * marcfiu Looks like the xen stuff has been merged into arch/i386 in recent kernels.
1161225017 M * marcfiu The 2.0.2.1 patch set does not contain changes to any Xen specific files. 
1161225048 M * marcfiu I am hoping that it simply wont be necessary to do anything special.
1161225078 M * Bertl will look into it, but probably it is already fine
1161225098 M * Bertl maybe ray6 knows more?
1161225118 M * marcfiu I'll have to dig up the old vserver-xen patches and double check that its probably fine as is.
1161226485 Q * ensc Killed (NickServ (GHOST command used by ensc_))
1161226495 J * ensc ~irc-ensc@p54B4F96A.dip.t-dialin.net
1161230811 M * Bertl okay, enough for me for today ... I'm off to bed now ... have a good one everyone! cya!
1161230819 N * Bertl Bertl_zZ
1161231156 P * marcfiu 
1161232421 Q * Greek0 Ping timeout: 480 seconds
1161234333 J * Greek0 ~greek0@85.255.145.201
1161238622 Q * sp Ping timeout: 480 seconds
1161238706 Q * _are_ Ping timeout: 480 seconds
1161238767 J * meandtheshell ~markus@85-124-233-97.work.xdsl-line.inode.at
1161240516 J * Hollow ~hollow@styx.xnull.de
1161240725 Q * Hollow 
1161240912 J * Hollow ~hollow@styx.xnull.de
1161241022 J * dna_ ~naucki@103-194-dsl.kielnet.net
1161242995 Q * dna_ Quit: Verlassend
1161243741 J * _mcp ~hightower@wolk-project.de
1161243745 Q * mcp Read error: Connection reset by peer
1161245344 J * eGnarF_ ~bartek@bk.crystone.se
1161245354 J * daniel_hozac|irssi ~daniel@c-2c1472d5.010-230-73746f22.cust.bredbandsbolaget.se
1161245360 Q * daniel_hozac Read error: Connection reset by peer
1161245502 Q * eGnarF Ping timeout: 480 seconds
1161245553 J * rgl Rui@217.129.151.190
1161245555 M * rgl good morning
1161245560 M * doener morning rgl 
1161246185 M * renihs morning 
1161246725 N * eGnarF_ eGnarF
1161246746 M * ensc|w mmh... kernel 2.6.18 is really bad :(  vserver patches do no work (slooow mouse with -t*, complete lockups with 2.02.2-rc2), some ominous ext3 corruption, .got table on my embedded devices are sometimes corrupted :(
1161246785 M * doener ensc|w: using the hard cpu scheduler?
1161246809 M * doener there was a bug that caused sllloooowwww mouses in there, not sure if it's fixed yet
1161246816 M * ensc|w yes
1161246850 M * ensc|w but this should be fixed in -rc2?
1161247116 M * doener I see no delta that would fix it, so I don't think that it's fixed in rc2
1161247406 M * ensc|w what's with the deltas for the do_IRQ() routines?
1161247426 M * ensc|w this looks like a candidate for a mouse fix
1161247583 M * doener hm, I considered that an extended "MySQL dies" fix, but yeah, Bertl said its intention was sth. else IIRC
1161247662 M * doener it's for 2.1.1 though
1161247754 M * ensc|w but as said, -rc2 locked completely (no mouse movement, no ethernet, no logentries). Hence I did not tried it on the machine with the slow mouse
1161247782 M * doener what kind of box is that?
1161247789 M * doener (the locking one)
1161247893 M * ensc|w HT, PREEMPT
1161247913 M * ensc|w slow mouse does NOT happen there
1161247924 M * ensc|w and -t9 seems to run without problems
1161248033 M * doener hm, that includes a switch from .18 to .18.1 (I guess you mean t8, right?)
1161248097 M * ensc|w yes
1161248105 M * ensc|w (switch to .18.1)
1161248211 J * azuli1 ~Carlos@user-10cm30m.cable.mindspring.com
1161248941 Q * azuli1 Ping timeout: 480 seconds
1161249119 J * debugger ~Rui@217.129.151.190
1161249273 M * fs I tested the slow mouse issue on 2.6.18.1 with 2.0.2.2-rc2 yesterday, with hardcpu and it was fixed
1161249450 J * azulin ~Carlos@user-10cm30m.cable.mindspring.com
1161249576 Q * rgl Ping timeout: 480 seconds
1161249981 N * daniel_hozac|irssi daniel_hozac
1161250002 M * doener daniel_hozac: hm, seems that the legacy interface broke
1161250015 M * daniel_hozac doener: hmm?
1161250054 M * daniel_hozac but yes, the do_IRQ is supposed to fix the mouse issues.
1161250076 Q * azulin Ping timeout: 480 seconds
1161250124 M * daniel_hozac ensc|w: what caused the lockup?
1161250413 M * ensc|w daniel_hozac: I think, a 'vserver ... restart'
1161250438 M * ensc|w but I saw it only once and rebooted into -t8
1161250439 M * daniel_hozac ensc|w: any panic or similar?
1161250460 M * ensc|w no; it was under X and there were no log entries
1161250472 M * ensc|w (completely dead, no network, mouse, disk)
1161250554 M * daniel_hozac magic sysrq?
1161250560 M * doener daniel_hozac: http://paste.linux-vserver.org/513
1161250577 M * doener legacy config vserver
1161250597 M * ensc|w daniel_hozac: no; did not worked anymore
1161250625 M * doener unless I'm reading the caps in the wrong way, CAP_NET_ADMIN is not set
1161250647 M * daniel_hozac should it be? :)
1161250659 M * doener no, but iptables should fail then :p
1161250675 J * lilalinux ~plasma@dslb-084-058-203-010.pools.arcor-ip.net
1161250704 M * doener (that's on 2.6..17.13-2.0.2.1
1161250731 M * daniel_hozac oh, so the iptables is inside the guest, ok.
1161250809 Q * lilalinux Remote host closed the connection
1161250857 M * daniel_hozac it doesn't work on 2.6.18.1-vs2.0.2.2-rc1 at least.
1161250883 M * doener with a plain chcontext-compat call it fails...
1161251013 M * daniel_hozac maybe CAP_NET_RAW is the culprit? would seem very strange though.
1161251070 N * debugger rgl
1161251082 J * lilalinux ~plasma@dslb-084-058-203-010.pools.arcor-ip.net
1161251097 M * doener hu? If I copy the chbind ... chcontext-compat ... bla call and do it myself, iptables fails?!
1161251099 M * rgl is this the way of boorstraping ubuntu http://paste.linux-vserver.org/514 ?
1161251197 Q * cdrx Quit: Leaving
1161251241 M * daniel_hozac rgl: why would you write to apps/debootstrap/uri twice?
1161251251 M * daniel_hozac but yes.
1161251270 M * rgl daniel_hozac, oh, the first write should not have been on that paste :)
1161251270 M * daniel_hozac doener: that seems very strange.
1161251282 M * rgl daniel_hozac, cool thx :)
1161251323 J * Piet hiddenserv@tor.noreply.org
1161251446 Q * marl_ Quit: Leaving
1161251569 M * wenchien I'm not sure if this is debian specific....  http://paste.linux-vserver.org/515
1161251644 M * daniel_hozac did you read NEWS?
1161251662 M * wenchien daniel_hozac: oh? ok, i'll read it.
1161251900 M * wenchien daniel_hozac: i see. thx. :)
1161251916 M * daniel_hozac np
1161251940 J * catweazle ~catweazle@e182046123.adsl.alicedsl.de
1161251992 Q * catweazle 
1161252377 M * doener daniel_hozac: hm, if I execute the whole thing in / it works as expected
1161252388 M * doener if I do it in /vservers/foo, I doesn't
1161252403 M * daniel_hozac i.e. it lets you run iptables?
1161252408 M * doener s/I/it/2
1161252410 M * doener yep
1161252440 M * daniel_hozac sounds very strange. tried stracing the the iptables?
1161252466 M * doener what would that tell me?
1161252484 M * daniel_hozac they have to be doing something different, no?
1161252497 M * daniel_hozac do you have a quick reproducer command?
1161252512 M * rgl you guys known how do I remove the klogd ubuntu package without also removing the syslogd package?  that is, apt-get --purge remove klogd whats to remove both :(
1161252735 M * Borg- w 5
1161252760 M * daniel_hozac shell-over-IRC doesn't really work that way :)
1161252832 M * SNy Hehe.
1161252839 M * doener daniel_hozac: it's CAP_NET_RAW
1161252920 M * SNy I'd rather assume it was meant to be a IRC-client-command, daniel_hozac.
1161252937 M * SNy As in "window numero 5".
1161252972 Q * rob-84x^ Read error: Operation timed out
1161253098 M * daniel_hozac doener: ok, but why does that change with different cwds?
1161253129 M * daniel_hozac SNy: hehe.
1161253130 M * doener maybe because the host is 64bit while the guest is 32bit
1161253141 M * doener could be a broken compat layer in the kernel...
1161253152 M * daniel_hozac doener: are you chrooting to cwd?
1161253169 M * doener yes, capchroot does that
1161253188 M * daniel_hozac ok.
1161254347 M * doener seems to be a 32/64bit problem
1161254353 M * doener investigating...
1161254550 M * doener hm, 32bit iptables compat layer was commited on April 1st... bad sign *g*
1161254566 M * daniel_hozac hahaha
1161255318 J * Piet_ hiddenserv@tor.noreply.org
1161255550 Q * Piet Ping timeout: 480 seconds
1161255992 Q * michal` Ping timeout: 480 seconds
1161256098 J * michal` ~michal@www.rsbac.org
1161256311 J * debugger ~Rui@217.129.151.190
1161256627 J * dna_ ~naucki@47-194-dsl.kielnet.net
1161256776 Q * rgl Ping timeout: 480 seconds
1161257614 Q * debugger Quit: Fui embora
1161257902 M * doener daniel_hozac: yep, bug, sending to security@ now...
1161257948 M * daniel_hozac ok, so a vanilla bug, cool.
1161258484 M * doener http://lkml.org/lkml/2006/10/19/81
1161258653 M * Borg- hehe nice..
1161258662 M * Borg- sounds a bit like ClearCase problem too ;)
1161258696 M * doener reminded me of the reiser4 requests that hit lkml right before the last flamewar about it
1161258760 M * daniel_hozac heh.
1161258795 M * daniel_hozac doener: have you seen the utrace patch by any chance?
1161258849 M * doener The one from June?
1161258980 M * daniel_hozac hmm, seems this one has gotten quite a few updates since then, so no.
1161259062 M * doener well, except for some references in the seccomp flamewar, I've no other traces of utrace in my lkml mbox
1161259576 M * daniel_hozac it seems the Fedora kernel is starting to deviate more and more from vanilla...
1161259617 M * doener wasn't Fedora also aiming for a nearly unpatched kernel? or was that Novell/SuSE?
1161259644 M * daniel_hozac i thought so.
1161259842 J * rob-84x^ rob@submarine.ath.cx
1161259965 Q * Piet_ Remote host closed the connection
1161260187 Q * MrX Quit: urk IRC v0.-1.4 - http://urk.sf.net/
1161260340 M * daniel_hozac is there any particular reason we're not vx_check'ing in may_attach?
1161260772 M * Hollow i think we need a better svn log browser..
1161260795 M * daniel_hozac how come?
1161260800 M * Hollow viewvc sucks :)
1161260819 M * Hollow it really misses a timeline view
1161260835 M * Hollow or a changeset view
1161260851 M * daniel_hozac yeah... changeset would be handy.
1161260899 M * Hollow unfortunately tracs svn browser cannot be used stand-alone :(
1161261001 M * daniel_hozac aww, i really like that one.
1161261012 M * Hollow yeah, me too
1161261094 M * Hollow well, do you think it would make sense to install trac? would imply a wiki for each project/svn repo, all other components of trac could be disabled, although i think a bug tracker could be handy as well
1161261122 M * Hollow of course public editing could be prevented
1161261187 M * daniel_hozac yet another wiki for each project seems counter-productive.
1161261217 M * daniel_hozac and util-vserver already has a bug tracker.c
1161261227 M * Hollow indeed... but if public editing is disabled we could just make a frontpage for each repo with some general info and link to the main wiki
1161261327 M * daniel_hozac i suppose.
1161261495 M * Hollow i'll take a look after fixing the sqldump for mysql 4.1 for the old wiki
1161261670 J * marcfiu ~mef@targe.CS.Princeton.EDU
1161262288 J * s0undt3ch_ ~s0undt3ch@81.193.63.42
1161262297 Q * s0undt3ch Read error: Connection reset by peer
1161262297 Q * dmax Read error: Connection reset by peer
1161262303 N * s0undt3ch_ s0undt3ch
1161262389 J * dmax ~semaj@81.193.63.42
1161262483 M * Hollow yay, semi-success
1161262727 Q * lilalinux Remote host closed the connection
1161262836 A * Hollow looks forward to PHP6
1161263004 Q * harry Killed (FloodServ ((FloodServ) Warning, you have triggered a network protection. Stop flooding!))
1161263138 M * Hollow ha!
1161263141 M * Hollow oldwiki works
1161263158 M * Hollow with umlauts and history and complete pages ;)
1161263160 J * ws larstr@colargol.tihlde.hist.no
1161263325 J * harry ~harry@d54C2508C.access.telenet.be
1161263676 M * daniel_hozac yay!
1161263821 Q * Aiken_ Quit: Leaving
1161264102 M * Hollow and the timestamps are ok too now... strange enough that tavi does not use epoch, but mysql TIMESTAMP data type.. *bah*
1161264181 M * Hollow but in the source it converts mysql TIMESTAMP to epoch, and reformats it to the format of TIMESTAMP... wtf
1161264209 M * daniel_hozac heh
1161265456 Q * virtuoso Remote host closed the connection
1161265591 J * virtuoso ~s0t0na@shisha.spb.ru
1161265708 Q * dna_ Quit: Verlassend
1161265828 M * Hollow FYI, the rewrite map for the wiki relocation has been updated...
1161265871 Q * ws Quit: Leaving
1161265873 M * daniel_hozac cool
1161265874 P * marcfiu 
1161266693 J * dna_ ~naucki@47-194-dsl.kielnet.net
1161267578 Q * Medivh Read error: Connection reset by peer
1161267647 J * Medivh ck@paradise.by.the.dashboardlight.de
1161268101 M * kevinp morning all :)
1161268125 M * kevinp Hollow, is there a way to get at the source of the oldwiki articles for easier migration?
1161268176 M * Hollow hm
1161268206 M * kevinp First it would make it easier to migrate manually, secondly, if there are certain compatibility changes that need to take place in all of them, it shouldn't be hard to migrate with a script.
1161268264 M * kevinp I assume someone has already checked for an existing conversion script from the oldwiki format to the new one?
1161268300 M * Hollow automatic migration is even more PITA than manual IMO
1161268320 M * Hollow since many pages should be consolidated
1161268425 M * daniel_hozac and restructured.
1161268505 M * Hollow e.g. VServer_Configuration is not a good title IMO
1161268517 M * Hollow this could be kernel config, util-vserver config, vcd config, etc pp
1161268535 M * daniel_hozac yeah...
1161268585 M * Hollow i also got an email from someone who migrated the HowTo+Read+ProcFS page about naming conventions
1161268714 A * Hollow sighs
1161268723 M * Hollow http://wiki.linux-vserver.org/ProblematicPrograms and http://wiki.linux-vserver.org/Problematic_Programs
1161268784 N * Bertl_zZ Bertl
1161268788 M * Bertl morning folks!
1161268820 M * Hollow morning Bertl
1161268829 M * renihs morning bertl :)
1161268847 M * Bertl Hollow: what I'm still missing is a mechanism to get an email when _any_ page is changed and a way to delete pages for special folks (like powerfox)
1161268853 M * Hollow Bertl: do we really need 3 relocation maps? if we do, please add the necessary subdomains to *.at and *.net
1161268857 M * kevinp Hollow, I understand the desired changes in structure, which would take quite sometime, but you also run the risk of information going stale if you're not able to update it
1161268888 M * Hollow Bertl: people in the sysops group can delete pages
1161268890 M * daniel_hozac kevinp: if you're updating it, why not take the time to do it properly?
1161268898 A * Hollow nods
1161268915 M * Bertl Hollow: okay, so we solved one out of two already :)
1161268933 M * Hollow Bertl: yep, i will also take a look at anti-spam plugins for mediawiki
1161268964 M * Bertl powerfox did an excellent job on removing/reverting spam on the old wiki
1161268973 M * kevinp I agree, but it would be *easier* and more people would be apt to do it if we could pull up the original source and start with that rather then convert from html
1161269006 M * Bertl (so given that the wiki notification can be made working, that should be available immediately for the new wiki too)
1161269007 M * Hollow Bertl: ok, i will give the necessary rights to powerfox
1161269058 M * Bertl btw, a 'delta' (diff) of the changes sent via email would be great, could go over some special wiki mailing list as we had
1161269108 M * Bertl I read you fixed the oldwiki copy? great!
1161269183 M * Bertl I thought about the request to get the wiki source for old pages
1161269219 M * Bertl why not adapt the save page (i.e. use the blacklist) to cancel saves compeltely (on the old wiki) but allow to enter the edit page?
1161269233 M * Bertl (should be no problem with the changes I did to tavi)
1161269261 M * Hollow well, i disabled the edit function when we switched...
1161269316 M * Bertl yes, which keeps folks from extracting the wiki source
1161269511 M * Hollow there is a watchlist feature in mediawiki, but you have to add pages manually to it..
1161269602 M * Bertl hmm, that's bad
1161269632 M * Bertl well, we probably have to hack that then .. mediawiki is php?
1161269668 M * Hollow yep.. i'm currently looking around for patches
1161269707 M * Bertl okay, great, if you find nothing suitable, I'll take a shot at it on the weekend
1161269863 J * foobear foobear@193.30.161.200
1161269866 A * foobear slaps Bertl around a bit with a large trout
1161269868 A * foobear slaps Bertl around a bit with a large trout
1161269870 N * foobear avi
1161269879 M * Bertl hmm, why would you do that?
1161269906 M * doener maybe he has just discovered that button in mIRC? *g*
1161269911 M * Hollow Bertl: i added powerfox to the sysops at least now
1161269960 M * Bertl doener: there is a button in mirc for that?
1161269960 M * avi Bertl, because i <3 u 
1161269974 M * avi remembah me Bretl o' mine ?!?!
1161269981 M * avi avi openbsd.org.il
1161269991 M * avi listen lord i have a quick and dumb question for u
1161270006 M * Bertl avi: my memory is not very good, let's hear your question :)
1161270010 M * avi loool
1161270023 M * avi Bertl: OK lord
1161270030 M * avi where do i set the caps for vserver2++
1161270046 M * avi sorry for the extreamly dumb questione...:D
1161270059 M * avi (for a specific vserver ofcourse)
1161270061 M * doener doener: I think I remember one, it's been ages since I used it... But what other reasons is there that you can see that line regularly from windows users? (especially those that laugh out out out loud ;)
1161270076 M * doener ehrm... s/doener/Bertl/
1161270103 M * daniel_hozac hehe
1161270123 M * daniel_hozac avi: /etc/vservers/<name>/bcapabilities
1161270129 M * Bertl daniel_hozac: what is the 'current' version of the flower page?
1161270142 M * Bertl the wiki one or the original one?
1161270147 M * daniel_hozac hmm? both should be accurate.
1161270161 M * Bertl ah, great, which one should I refer folks to then?
1161270176 M * Bertl (i.e. will they stay in sync?)
1161270186 M * daniel_hozac yes, they should.
1161270188 M * daniel_hozac i can't personally update the wiki one though.
1161270192 M * daniel_hozac so it might lag a bit.
1161270199 M * avi lool @ flowah page
1161270234 M * Bertl avi: http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1161270253 M * daniel_hozac and i think we've worked too hard to get the nongnu.org page up as the first hit on great flower page on google to move it :)
1161270289 M * Hollow hehe
1161270303 M * avi Bertl
1161270320 M * avi ur a good guy i always knew that... i'll call theo today and tell hin
1161270349 M * Bertl okay, I'm off now .. will be back later ...
1161270355 N * Bertl Bertl_oO
1161270405 M * avi <3
1161271168 M * daniel_hozac Bertl_oO, doener: shouldn't vx_barrier check for VX_ADMIN|VX_WATCH? is there any sense in denying access to the spectator context?
1161271191 M * daniel_hozac s/vx_barrier/[vd]x_barrier/
1161271633 M * WorkRoey hey all
1161271640 M * WorkRoey how do I make a vserver start automatically on system startup?
1161271748 M * daniel_hozac make the vservers-default (or whatever your distribution's script is) start on boot, and mark your guests with default.
1161271759 M * daniel_hozac i.e. echo default > /etc/vservers/<name>/apps/init/mark
1161271797 M * avi daniel_hozac
1161271808 M * avi who currently maintain the gentoo version of vserver?
1161271819 M * daniel_hozac Hollow and phreak``, AFAIK.
1161271871 M * Hollow pong
1161271874 M * Hollow :)
1161271927 M * avi Hollow lordy lord
1161271935 M * avi are you maintaining the gentahh version ?
1161271950 M * WorkRoey daniel_hozac:  thanks
1161271956 M * Hollow gentoo, yes
1161271963 M * avi ;pp;p;
1161272002 M * avi say why dont u make an exec that will allow one to edit/append configuration changes for a specific vserver? (such as caps, add/del nics etc) ?
1161272021 M * avi and maybe add it to the gentahh vserver port... 
1161272031 M * doener because every editor can do that?
1161272038 M * doener ;)
1161272045 M * Hollow beside that this isn't gentoo specific, an editor can be used quite good for this purpose
1161272051 M * avi well thats a lot of work once u need to maintain several machines
1161272066 M * Hollow but, there will be/there is a webfront end
1161272073 M * avi there is?
1161272081 M * Hollow for util-vserver there is openvcp
1161272088 M * avi lol
1161272089 M * WorkRoey also
1161272091 M * Hollow for vcd there will be a syscp plugin at least
1161272092 M * WorkRoey daniel_hozac:  hi again
1161272093 M * daniel_hozac and vserver ... config hopefully isn't too far off.
1161272098 M * WorkRoey I cant' seem to log in as root into the vservers
1161272103 M * avi didnt know and i used vservah for almost a year and half
1161272110 M * daniel_hozac WorkRoey: why not?
1161272131 M * WorkRoey I'mt rying to find out.  Nowhere in my vserver's /etc/login.defs or /etc/sshd_config does it say to deny root logins.
1161272148 M * WorkRoey daniel_hozac:  I can vserver enter but not ssh into them as root.
1161272185 M * daniel_hozac have you set a root password?
1161272193 M * daniel_hozac what do your logs say?
1161272195 J * stefani ~stefani@tsipoor.banerian.org
1161272200 M * avi So Guys, The Vserver project is undergoing changes?
1161272209 M * Hollow avi: every day :)
1161272226 M * avi LoL 
1161272226 M * WorkRoey daniel_hozac:  yeah I've set a root password.  Lemme check my logs, I didnt' think to do that ;)
1161272228 M * WorkRoey thanks
1161272234 M * avi I mean Like a new website etc
1161272275 M * Hollow avi: yes, we currently migrate to the new site
1161272290 M * WorkRoey Oct 19 11:19:31 sink sshd[7921]: error: PAM: Authentication failure for root from h-69-3-4-193.mclnva23.covad.net
1161272292 M * WorkRoey er
1161272293 M * WorkRoey yeah.
1161272298 M * WorkRoey daniel_hozac:  that's what it says.
1161272303 M * WorkRoey so I suspected pam actually
1161272306 M * WorkRoey mayeb I need a root_ok.so
1161272375 M * WorkRoey daniel_hozac:  nope, that's not it.
1161272384 M * WorkRoey (tried adding auth required root_ok.so to /etc/pam.d/ssh)
1161272386 M * doener WorkRoey: do you allow password logins for root?
1161272410 M * WorkRoey doener:  well like I said, /etc/login.defs and /etc/ssh/sshd_config don't say anything about denying it.
1161272417 M * WorkRoey doener:  well I shouldn't really allow for that, you're right.
1161272417 M * doener hm, well, guess pam wouldn't be involved otherwise...
1161272462 M * daniel_hozac WorkRoey: pam_unix is usually very verbose in its logging.
1161272470 M * WorkRoey alright then.
1161272472 M * WorkRoey well like I said
1161272479 M * WorkRoey I can log in as a user I've created in the vserver.
1161272485 M * WorkRoey and sudo or su from there.
1161272592 M * tokkee Can I make /proc/virtual visible in a guest?
1161272622 M * WorkRoey also
1161272624 M * Hollow off for a bit, have to get some groceries
1161272682 M * daniel_hozac tokkee: setattr --~hide?
1161272689 M * WorkRoey daniel_hozac:  I'm on debian/unstable for the host, with kernel 2.6.14 and vserver 2.1.0. Is it worth upgrading?  are there issues to note?
1161272705 M * daniel_hozac well, yeah.
1161272724 M * daniel_hozac 2.6.14 probably has more holes than your average swiss cheese :)
1161272792 M * tokkee daniel_hozac: Did it. Thx.
1161272815 M * doener WorkRoey: any other messages in your logs besides the one above?
1161273009 M * daniel_hozac WorkRoey: does su ask for the password? what are the differences between /etc/pam.d/su and ssh?
1161273043 M * WorkRoey doener:  well many, but what in particular?
1161273050 M * WorkRoey daniel_hozac:  so I should upgrade kernels and vserver.
1161273068 J * sp sp@ip-87-97.members.virt-ix.net
1161273194 M * WorkRoey daniel_hozac, doener: to what should I upgrade?
1161273199 M * WorkRoey I see many versions on linux-vserver.org
1161273260 M * WorkRoey Bertl_oO, Hollow:  hi! 
1161273277 M * WorkRoey avi:  and hi to you too
1161273278 M * WorkRoey em
1161273302 M * WorkRoey do I need to change anything before I update my kernel+vserver patch ?
1161273304 M * daniel_hozac i'd upgrade to 2.6.18.1-vs2.1.1-rc41
1161273322 M * WorkRoey daniel_hozac:  I didnt' know vs2.1.1-rc41 applied to 2.6.18.1 cleanly, tanks
1161273323 M * WorkRoey *thanks
1161273363 M * daniel_hozac umm, it's the only one it applies completely clean against.
1161273488 M * WorkRoey oh, heh
1161273489 M * WorkRoey ok.
1161273559 M * avi WorkRoey: :D:D:D:DD: <3
1161273665 M * WorkRoey which avi?
1161273671 M * WorkRoey avi from EFNet?
1161273674 M * WorkRoey avi from Freenode?
1161273679 M * WorkRoey avi:  �הלן
1161273744 M * avi loool
1161273746 M * avi avi 
1161273749 M * avi I KNOW U !!!?!!
1161273801 M * avi WorkRoey: ata ko hamud
1161274059 J * bonbons ~bonbons@83.222.36.111
1161274349 Q * phreak`` Ping timeout: 480 seconds
1161275386 M * Skram hey, Johnnie, around?
1161275402 M * Skram Just curious if you want to chat.. I am on a lunch break and can't get on AIM from this location :\
1161275418 M * Skram sort of silly how I can SSH out from here but not get to AIM.com
1161275441 M * node why do you need to get to AIM.com to use aim?
1161275499 M * Skram oh.. i could use some yucky text-based one on my shell server ;\
1161275568 J * notasnark ~sam@mailgate.glendale.org.uk
1161275980 J * lilalinux ~plasma@dslb-084-058-203-010.pools.arcor-ip.net
1161276060 M * notasnark Hello, I'm having a problem with a vserver guest under Gentoo, and I'm looking for some pointers on how to investigate/fix.
1161276103 M * notasnark Recently, one guest stopped starting up. Immediately after starting, a "/sbin/reboot -f" appears in the process list, and it shuts down.
1161276146 M * notasnark Anybody know how I can find out what is causing this?
1161276176 M * Hollow notasnark: which init style are you using? and which version of baselayout-vserver is running inside?
1161276245 M * notasnark I think baselayout-vserver-1.11.14-r4. Not sure about the init style. How do I check?
1161276278 M * Hollow in /etc/vservers/<name>/apps/init/style
1161276297 M * notasnark plain
1161276345 M * notasnark I recently upgraded the gcc version in this vserver. It was working fine before then, and fine until I restarted it some days later.
1161276357 J * _are_ ~are@stgt-d9be6728.pool.mediaWays.net
1161276403 M * notasnark However, I've seen comments saying that upgrading the gcc shouldn't be a problem (I'm running 3.3.6 on the host still).
1161276460 J * Rich_Estill ~restill@c-24-11-195-139.hsd1.mi.comcast.net
1161276485 M * Hollow the thing is... all /sbin/reboot have been removed for baselayout-vserver, except the one for runlevel 6
1161276516 M * Hollow are you sure it is baselayout-vserver and not baselayout?
1161276607 M * notasnark baselayout-vserver is what I've got installed on the others. And it was what I installed when I first built them many months ago.
1161276678 M * notasnark Hmm... it looks like both baselayout and baselayout-vserver are installed.
1161276712 M * notasnark (a quick emerge -p of both tells before the guest dies tells me I have sys-apps/baselayout-1.12.5-r2 and sys-apps/baselayout-vserver-1.11.14-r4)
1161276719 M * Hollow maybe an update of baselayout overwrote files
1161276749 M * notasnark Probably during one of the emerge -puD world I did whilst rebuilding gcc/glib
1161276755 M * Hollow you can simply chroot to your vservers root and fix it (remember to bind mount procfs and portage)
1161276849 M * notasnark What's the best way to mount /usr/portage? Do I do it on the guest before the chroot (in the right place in the /vserver/guest directory)?
1161276855 N * Bertl_oO Bertl
1161276860 M * Bertl back now ...
1161276881 M * Hollow notasnark: mount --bind /usr/portage /vservers/guest/usr/portage should suffice for the chroot
1161276944 M * Hollow notasnark: see the blue box on top of http://oldwiki.linux-vserver.org/GentooGuestHowto how to convert from baselayout to baselayout-vserver in a safe way
1161276986 M * notasnark Okay, thanks. I'll give that a go.
1161277044 M * Hollow feel free to ping me if you have problems, though i'm afk a bit now, dinner time...
1161277068 M * notasnark Thanks for your help (again!).
1161277073 M * Hollow yw!
1161277721 M * Bertl daniel_hozac: ping?
1161277778 Q * rob-84x^ Quit: That's it for today
1161277788 J * rob-84x^ rob@submarine.ath.cx
1161278127 J * crypt ice@p54A3B55B.dip0.t-ipconnect.de
1161278143 M * Bertl welcome crypt!
1161278148 M * crypt hi Bertl 
1161278326 M * crypt Bertl, i have a question: why does a node not using the swap? RAM is completly used but no swap
1161278349 M * Bertl mainly because swap is less efficient than ram (in your case)
1161278379 M * Bertl i.e. the host decides when to swap out stuff, as long as ram is available that is typically a no
1161278531 M * crypt hmm
1161278559 M * crypt but isn't it faster to use swap then accesing all from hdd?
1161278559 Q * derjohn2 Ping timeout: 480 seconds
1161278588 M * Bertl crypt: swap will a) never be used for stuff _on_ disk
1161278595 J * derjohn2 ~aj@dslb-084-058-196-127.pools.arcor-ip.net
1161278603 M * Bertl and b) swap is disk I/O too, so why should it be faster?
1161278742 M * crypt becaus i have 2 swap partiotions on 2 different hdd'S
1161278790 M * crypt the problem is i have ~20%wa and running a RAID1 configuration
1161278829 M * Bertl the following rules apply to linux swap (in general)
1161278848 M * Bertl - files/inodes are cached, and never swapped, just discarded
1161278878 M * Bertl - anonymous memory is swapped
1161278893 M * Bertl - shared memory can be swapped but usually will not
1161278915 M * Bertl when it comes to page eviction, the following order is (usually) used
1161278932 M * Bertl - discard unused 'swap caches'
1161278941 M * Bertl - discard file backed pages
1161278958 M * Bertl - swap out anonymous memory (lru)
1161278966 M * Bertl - try to swap out rest
1161278970 M * Bertl - OOM
1161279067 M * crypt ok then i have to upgrade my server or have to change my raid1 configuration into a raid0 conf to speed up my server
1161279108 M * doener for reads, raid1 should be as fast as raid0 AFAIK
1161279171 M * doener do you know which process is accessing the disk so heavily?
1161279233 M * crypt apache2
1161279286 M * doener log writing?
1161279294 M * Bertl most likely
1161279308 M * Bertl putting the logs on a fast disk (maybe even ram disk) would help
1161279317 M * doener yeah, pages should be in cache, just asking to be sure ;)
1161279328 M * Bertl adjusting the log to write a single log file every now and then helps too
1161279331 M * crypt how can i realize that in a vserver?
1161279344 M * crypt the ram disk
1161279358 M * doener make the host mount the ram disk on vserver startup
1161279359 M * Bertl /tmp for example is a ram disk
1161279364 M * doener using /etc/vservers/foo/fstab
1161279383 M * Bertl crypt: you should have an example in the fstab (in the config tree)
1161279429 M * crypt ahh yes indeed, so it could help to let apache log into /tmp and writing a script which transfere old not used logs onto hd
1161279441 M * crypt every night after log rotate
1161279455 M * Bertl for example, note that the current /tmp is limited to 16M by default
1161279476 M * crypt i know ;) i have already increased that
1161279477 M * Bertl i.e. you might want to raise that in this case (or have a separate one)
1161279498 M * doener hum, if apache is writing logs so heavily, that might use up all space on the tmpfs quite soon...
1161279499 M * Bertl this, btw _is_ space which can be swapped out :)
1161279515 M * crypt ah ok thx :)
1161279527 M * crypt i love the support you get here very much :)
1161279537 M * Bertl you're welcome!
1161279557 M * crypt i'll write a workaround for that :D
1161279561 M * doener Bertl: hm, the cache can be swapped out?
1161279575 M * crypt btw where can i see how much of tmp in the vserver is used?
1161279587 M * Bertl doener: tmpfs should we swapable
1161279593 M * crypt because df doesn't show that
1161279603 M * doener vnamespace -e $vserver df
1161279606 M * Bertl hmm? it should
1161279626 M * doener (assuming that you were starting df on the host)
1161279689 M * crypt thx doener :) i doesn't see this because this vserver is to old to have the tmp entry in fstab
1161279697 M * crypt running now for a year
1161279701 M * Bertl crypt: if, for whatever reason, df doesn't show it _inside_ the guest, you can 'just' add the tmp to mtab
1161279722 M * Bertl that will make df display that too ;)
1161279727 M * crypt ;)
1161279955 M * crypt is this entry correct?
1161279955 M * crypt none /dev tmpfs size=256m,mode=1777 0 0
1161280125 M * crypt in /etc/vservers/$vserver/fstab
1161280171 M * Bertl /dev ? 
1161280178 M * crypt äh /tmp
1161280212 M * Bertl I would have suspected /var/log/httpd but yeah, it seems fine
1161280258 M * crypt ok thx :) that what you said is even a better idea, i think ;)
1161280258 M * Bertl note that you probably want a script on shutdown to copy the log files to permanent store too (and populate the dir on startup with the last version)
1161280359 M * crypt i have to put these into rc6 and rc1 right?
1161280388 M * Bertl depends on the config you have, check what is executed and adjust accordingly
1161280400 M * crypt ok thx
1161280573 M * WorkRoey hi
1161280577 M * WorkRoey bertl!!
1161280585 M * WorkRoey I am untarring the latest kernel.
1161280586 M * WorkRoey an I get:
1161280587 M * Bertl hey WorkRoey!
1161280588 M * WorkRoey tar: linux-2.6.18.1/scripts/kconfig/expr.c: Cannot open: No such file or directory
1161280595 M * WorkRoey hey hey Herbert :)
1161280601 M * WorkRoey and I get a BUNCH of them.
1161280603 M * WorkRoey what's up with that??
1161280624 M * Bertl you are talking about vanilla kernel tars?
1161280627 M * WorkRoey yeah
1161280630 M * WorkRoey 2.6.18.1
1161280638 M * Bertl did you check the md5sum?
1161280649 M * WorkRoey nope
1161280651 M * WorkRoey lemme do that.
1161280694 M * WorkRoey http://kernel.org/ doesn't seem to give an md5sum anywhere
1161280750 M * WorkRoey I see a .sign
1161280759 M * WorkRoey how can I verify authenticity with a .sign ???
1161280790 M * Bertl Comment: See http://www.kernel.org/signature.html for info
1161280832 M * WorkRoey thanks
1161280836 M * Bertl np
1161281023 M * WorkRoey Bertl:  
1161281025 M * WorkRoey root@magneto:/home/katzr/linux# gpg --verify linux-2.6.18.1.tar.bz2.sign  linux-2.6.18.1.tar.bz2
1161281025 M * WorkRoey gpg: Signature made Sat 14 Oct 2006 02:03:29 AM EDT using DSA key ID 517D0F0E
1161281025 M * WorkRoey gpg: Can't check signature: public key not found
1161281046 M * doener try gpg --recv-keys $ID
1161281046 M * WorkRoey I will try importing the keyring
1161281056 M * WorkRoey what is $ID?
1161281064 M * doener 517D0F0E
1161281067 M * WorkRoey ah
1161281069 M * doener see your paste ;)
1161281581 J * pagano ~pagano@131.154.5.21
1161281585 M * pagano hi all :)
1161281608 M * Bertl welocme pagano!
1161281699 M * pagano hi Bert, I have a request for u :)
1161281737 M * pagano http://www.cs.princeton.edu/~mef/research/vserver/paper.pdf
1161281760 M * pagano here, can u tell me where I can found the image inside the paper>
1161281775 M * pagano the problem is that all the pdf is in black and white
1161281819 M * pagano ...also the images, so it is impossible to understand the graphics
1161282312 M * WorkRoey gzip: stdin: unexpected end of file
1161282312 M * WorkRoey tar: Unexpected EOF in archive
1161282312 M * WorkRoey tar: Error is not recoverable: exiting now
1161282317 M * WorkRoey Bertl:  I tried the .tar file, too.
1161282319 M * WorkRoey same stuff.
1161282564 M * Bertl WorkRoey: and the pgp signature matches?
1161282592 M * Bertl pagano: best contact mef
1161282636 M * pagano sorry bert ?
1161282661 M * pagano do u prefer a mail?
1161282668 M * Bertl well, the author of the paper (marc) he shows up every now and then
1161282684 M * Bertl and/or you can contact him via email too I guess
1161282703 M * pagano oh well, thanks.
1161282870 M * ensc Bertl: l
1161282887 M * ensc ... are the do_IRQ() patches really ok?
1161282906 M * ensc what is with locking of 'current'?
1161282932 M * ensc mmh... have to restart the irssi mem hog ...
1161282949 M * WorkRoey Bertl:  didn't try the pgp signature.  I can't get the stupi thing to work at all
1161283002 M * Bertl ensc: what issues do you see with 'locking' of current?
1161283015 M * ensc complete hangups...
1161283026 M * ensc but I do not know whether they are related
1161283056 M * ensc I just saw it when trying 2.6.18.1 + patch-2.6.18.1-vs2.0.2.2-rc2.diff
1161283058 M * Bertl I don't think the irq fixes can cause that
1161283076 M * Bertl WorkRoey: well, I presume your 'downloaded' archive is broken
1161283086 M * Bertl WorkRoey: I can do an md5 sum for you though
1161283118 M * WorkRoey I did the md5 sum
1161283125 M * WorkRoey I just wanted the source md5sum
1161283127 M * WorkRoey the 'good' sum.
1161283142 M * Bertl that's what I meant :)
1161283153 M * WorkRoey ah
1161283158 M * WorkRoey can ya get me a good md5sum?
1161283161 M * WorkRoey thanks :) :)
1161283267 M * crypt Bertl, i just created the ramdisk, but now %wa is ~70% :/
1161283289 M * Bertl would suggest that the issue was somewhere else
1161283295 M * Bertl check with iostat
1161283302 M * Bertl WorkRoey: 38f00633b02f07819d17bcd87d03eb3a  linux-2.6.18.1.tar.bz2
1161283313 M * Bertl WorkRoey: and that one passed the sign check too
1161283481 M * WorkRoey thanks
1161283526 M * WorkRoey ok
1161283528 M * WorkRoey that checks out
1161283531 M * WorkRoey my thing is pristine.
1161283536 M * WorkRoey now what else could be the problem?

1161283536 M * crypt Bertl, http://paste.linux-vserver.org/516 maybe you can comment that?
1161283565 M * Bertl WorkRoey: check disk space and use tar tjf on the file
1161283580 M * Bertl WorkRoey: also remove any existing files, btw, what filesystem?
1161283593 M * WorkRoey er
1161283594 M * WorkRoey lemme check
1161283606 M * WorkRoey I'm just hoping hoping hoping it's not something with the lvm
1161283632 M * Bertl crypt: try the interactive version (-i 1 or so IIRC)
1161283641 M * WorkRoey I had other weird stuff, too, like a file that shows up with 'ls' but that I cannot 'cat' ('no such file or directory') but that I /can/ see in the vserver guest
1161283725 M * ensc Bertl: lockups with -rc2 seem to be 100% reproducible
1161283734 M * ensc (that was the good message)
1161283749 M * ensc bad is, that there are no message/oopses :(
1161283804 M * Bertl ensc: let's activate the NMI oopser and Mutex debugging (as well as the vserver debugging) and magic sysctl
1161283823 M * ensc sysctl does not work
1161283829 M * Bertl why so?
1161283854 M * Bertl I mean, it doesn't work in general or when it hangs?
1161283855 M * ensc I mean it is activated and works usually. But not after the lockup
1161283868 M * Bertl okay, that is where the NMI oopser comes into play
1161283872 Q * lilalinux Remote host closed the connection
1161283883 M * crypt Bertl, http://paste.linux-vserver.org/517 is this correct?
1161283885 M * Bertl make sure it is activated (kernel config and boot options)
1161283890 M * Bertl crypt: sec
1161283904 M * crypt i have time ;)
1161283946 M * Bertl and does vmstat show heavy swap in/out?
1161283953 M * Bertl or is it just the disk I/O
1161283964 M * Bertl and more improtant, what I/O scheduler do you use?
1161283990 M * ensc anticipatory IO scheduler
1161283995 M * crypt this are things i have never thougt about ;) i'm using debian etch
1161284005 M * ensc lockups happen during 'vserver ... restart'
1161284029 M * Bertl ensc: that's a good test case, if reproduceable
1161284029 M * ensc (for initfaked vservers)
1161284051 M * Bertl let's see if you can get the NMI working
1161284059 M * Bertl that should at least give us an oops
1161284104 M * Bertl and do you know, does 2.1.x work or show the same behaviour?
1161284110 M * ensc how do I activate the NMI oopser?
1161284116 M * ensc never tried 2.1
1161284125 M * Bertl there is a short howto in Doumentation (kernel tree)
1161284358 M * ensc Bertl: __group_send_sig_info+0x33
1161284363 M * ensc do_notify_parent
1161284379 M * Bertl could you upload the entire trace?
1161284394 M * Bertl and do you have a 'working' addr2line?
1161284399 M * ensc I just see it at screen
1161284419 M * Bertl hmm, serial console?
1161284433 M * Bertl or photo?
1161284573 M * crypt Bertl, where do i see which I/O sheduler i'm using?
1161284747 M * Bertl /sys/block/<devicename>/queue/scheduler
1161284778 M * crypt there is no such file
1161284804 M * Bertl # cat /sys/block/hda/queue/scheduler 
1161284804 M * Bertl noop [anticipatory] deadline cfq 
1161284832 M * crypt same
1161284838 M * crypt sry missed the queu
1161284853 M * Bertl well, you want to check the devices you actually have in use :)
1161284861 M * Bertl but probably they are on default
1161284875 M * Bertl and does the vmstat show heavy swap activity?
1161284910 M * crypt the 1st hd; sda has the same sheduler u posted
1161285020 M * crypt unfortunatly i can't say whether this is high swap usage: http://paste.linux-vserver.org/518
1161285062 M * Bertl well, see the 'swap' column?
1161285069 M * Rich_Estill photo?  Thats an idea.  Do we need to record the morse code beeps too?
1161285094 M * crypt then i would say i haven't a high swap usage
1161285097 J * cunha ~adray@109.102.broadband6.iol.cz
1161285103 M * Bertl well, a camera snapshot is better than nothing :)
1161285110 M * Bertl welcome cunha!
1161285141 M * Bertl crypt: so I'd say there is still some heavy filesystem I/O going on
1161285143 M * cunha Hi, 
1161285159 M * crypt yes so it seems :(
1161285193 M * crypt is there a tool what could give me app which writes a lot to hdd?
1161285217 M * Bertl not really, unless you have I/O tracking enabled
1161285226 M * Bertl (which is a very new feature, btw)
1161285242 M * Bertl crypt: do you have several guests on that machine?
1161285251 M * crypt yes 7
1161285253 M * Bertl crypt: is the hard cpu scheduler activated?
1161285269 M * cunha How to make 'free' show me information about one guest and not about the whole machine?
1161285271 M * crypt no
1161285327 M * Bertl cunha: http://oldwiki.linux-vserver.org/Caps+and+Flags
1161285340 M * Bertl you want the VIRT_MEM
1161285382 M * WorkRoey Bertl:  I tried tar tzf, and it gives me:
1161285383 M * crypt cunha, try using openvcp ;)
1161285399 M * WorkRoey gzip: stdin: unexpected end of file
1161285399 M * WorkRoey tar: Unexpected EOF in archive
1161285399 M * WorkRoey tar: Error is not recoverable: exiting now
1161285399 M * WorkRoey Bertl: k
1161285415 M * WorkRoey Bertl:  about halfway into the test
1161285443 M * Bertl never said anything about tzf :)
1161285460 M * Bertl bz2 -> tjf :)
1161285484 M * Bertl crypt: how would that help with 'free'?
1161285495 M * crypt we set the flag automatically
1161285508 M * crypt if you want it ;)
1161285558 M * Bertl i.c. ...
1161285599 M * crypt Bertl, i thought hard cpu shedule doens't work yet?
1161285617 M * Bertl why do you think so?
1161285638 M * cunha Bertl: Yes, I've found it, but I don't know how to use it. I have to make 'flags' file in /etc/vservers/guest and then? What must I write into it?
1161285677 M * WorkRoey Bertl:  OK, the tar bzf finished fine this time.  I think it was a space issue.
1161285679 M * WorkRoey thanks!
1161285682 M * Bertl cunha: http://www.nongnu.org/util-vserver/doc/conf/configuration.html
1161285692 N * _mcp mcp
1161285694 M * crypt Bertl, i thought reading such a thing in the wiki (a long time ago)
1161285702 M * Bertl cunha: (for the pathes) and see the 'config' column, that's what you write there
1161285714 M * crypt Bertl, would it be good to activate this feature?
1161285730 M * Bertl well, it would help you now to figure the responsible guest
1161285743 M * Bertl (which could ease isolating the issue)
1161285802 M * crypt how could i see this by setting this flag?
1161285863 M * Bertl you could 'pause' a guest, which should either show a difference (in the I/O) or not
1161285940 M * crypt hum ok i'll check
1161286015 Q * phedny_ Ping timeout: 480 seconds
1161286072 M * Bertl crypt: probably best is to shut down one guest at a time
1161286094 M * Bertl crypt: for a short period, and check if the I/O is reduced significantly
1161286149 M * crypt ok i'll check that thanks for your help!!!!
1161286187 M * crypt another question: does util-vserver performe each morning at 6 o'clock some tasks?
1161286217 N * mcp _mcp
1161286217 M * WorkRoey hey Bertl
1161286229 M * WorkRoey the rc40 patch doesn't apply cleanly to 2.6.18.1
1161286232 M * WorkRoey Bertl:  I get:
1161286235 M * WorkRoey Bertl:  Hunk #1 FAILED at 1.
1161286235 M * WorkRoey 1 out of 1 hunk FAILED -- saving rejects to file Makefile.rej
1161286257 M * WorkRoey patching file include/linux/Kbuild
1161286257 M * WorkRoey Hunk #1 FAILED at 61
1161286282 M * WorkRoey patching file Makefile
1161286282 M * WorkRoey Hunk #1 FAILED at 1.
1161286285 J * phedny_ ~mark@volcano.p-bierman.nl
1161286293 M * WorkRoey Bertl:  is that known?
1161286323 M * node WorkRoey: what arguments are you passing to patch?
1161286330 M * WorkRoey -P1
1161286349 M * WorkRoey cat ../../linux/patch-2.6.18-vs2.1.1-rc40.diff  | patch -p1
1161286407 M * Bertl are you in the kernel dir?
1161286505 M * WorkRoey yes
1161286505 M * WorkRoey yes I am
1161286522 M * WorkRoey a bunch of stuff--most stuff--passes, but two files fail
1161286535 M * Bertl hmm, better get the correct patch then
1161286552 M * WorkRoey for 2.6.18.1?
1161286556 M * WorkRoey you said that it worked
1161286568 M * Bertl there is one for 2.6.18.1 too
1161286586 M * WorkRoey where at??
1161286597 M * WorkRoey it's not patch-2.6.18-vs2.1.1-rc40.diff   ?
1161286614 M * WorkRoey vs2.0.2.2-rc2
1161286615 M * WorkRoey ?
1161286624 M * WorkRoey isn't that old???
1161286676 M * cunha Bertl: Thanx
1161286704 M * WorkRoey Bertl:  should I be using 2.0 or 2.1???
1161286806 M * Bertl cunha: everything working now as expected?
1161286818 M * Bertl WorkRoey: depends on what features your want/need
1161286847 M * WorkRoey nm
1161286854 M * WorkRoey I was using the wrong patch file for 2.6.18.1
1161286859 M * WorkRoey I was using rc40 instead of rc41
1161286862 M * WorkRoey my bad.
1161286923 M * Bertl ensc: any more details?
1161286948 M * Bertl s/more/further/
1161287023 M * daniel_hozac Bertl: pong
1161287036 M * ensc Bertl: me is fighting with the serial console
1161287047 M * daniel_hozac crypt: not unless you told it to.
1161287055 M * ensc but I see kernel warnings at kernel/pid.c:270/pid_task()
1161287065 M * Wonka mmmh. how's IPv6 support in devel?
1161287078 M * crypt daniel_hozac, i have each morning a cpu and load peak at 6:20
1161287082 M * ensc Bertl: called by kill_proc_info()
1161287083 M * daniel_hozac Wonka: non-existant, you'll need bonbons' patch ;)
1161287083 M * Bertl ensc: warnings are fine
1161287085 M * crypt but only on vserver nodes
1161287091 M * Wonka daniel_hozac: :(
1161287103 M * daniel_hozac crypt: do your guests run something at that time?
1161287110 M * Bertl ensc: what issues with the serial console do you have?
1161287119 M * Wonka daniel_hozac: i heard it was going to be included in devel "soon"
1161287126 M * daniel_hozac Wonka: yep.
1161287127 M * Bertl daniel_hozac: I'm currently cleaning up the networking code
1161287129 M * crypt daniel_hozac, i checked alll
1161287143 M * crypt no guest perfoming tasks at this time of da
1161287145 M * crypt y
1161287148 M * daniel_hozac Bertl: yeah, i figured.
1161287149 M * Bertl daniel_hozac: how much efford would it be to move the HIDE_NETIF from VXF to NXF?
1161287159 M * ensc stupid things like typing /dev/tty0 instead of S0
1161287163 M * daniel_hozac crypt: even with timezone configuration factored in?
1161287169 M * crypt UTC
1161287179 M * Bertl ensc: ah, if you need help somewhere, just let me know
1161287180 M * daniel_hozac even inside the guests?
1161287193 M * daniel_hozac Bertl: quite a bit of effort, i think. i don't think we support network flags at all yet.
1161287204 M * crypt no when i have another timezone then it's for example 8:20
1161287220 M * Bertl daniel_hozac: well, I can work around the legacy case in the kernel
1161287248 M * Bertl daniel_hozac: what I would like to see is that the new NXF_HIDE_NETIF is passed on network context creation by default (first step)
1161287258 M * cunha Bertl: It  seems so
1161287276 M * Bertl daniel_hozac: and that some --flag attribute (which actually should have be there for some while now :) is added (second step)?
1161287302 M * Bertl daniel_hozac: third step would be support in the config I guess with nflags or so=
1161287302 M * daniel_hozac Bertl: the API doesn't even support passing flags right now :)
1161287314 M * Bertl daniel_hozac: the kernel api?
1161287315 M * daniel_hozac yeah, i've thought about this before.
1161287319 M * daniel_hozac no, the util-vserver API.
1161287323 M * Bertl ah, k
1161287361 M * Bertl so step 1 should be straight forward, and even backwards compatible, as the flag is not used (yet)
1161287382 M * Bertl i.e. just adding the new flag there, unconditionally
1161287430 M * daniel_hozac VXF_HIDE_NETIF will be ignored?
1161287474 M * Bertl yep, it becomes a 'blank' for now
1161287493 M * Bertl (i.e. we won't reuse it too soon ...
1161287523 M * Bertl legacy, i.e. ipv4root stuff will be handled kernel side
1161287528 M * daniel_hozac yeah, shouldn't be too hard then.
1161287547 M * Bertl we had that on for backwards compatibility anyways
1161287562 Q * cunha Quit: Leaving
1161287584 M * Bertl daniel_hozac: okay great! would be cool to have that in the next release (if possible)
1161287601 M * daniel_hozac ok, will do.
1161287602 M * Bertl (bit/value is the same as for VXF)
1161287633 M * Bertl patches for further testing will be available today I guess
1161287651 Q * _are_ Ping timeout: 480 seconds
1161287688 Q * meandtheshell Quit: exit (0);
1161287770 M * Bertl ensc: I added a 'warning' to pid_task() this will cause 'oops' like dumps, but they are no oopses
1161287820 M * Bertl ensc: I'm still looking for the details of those warnings (i.e. what exactly causes them and why they show up)
1161287860 J * cunha ~cunha@109.102.broadband6.iol.cz
1161287869 M * Bertl wb cunha!
1161287881 Q * cunha 
1161288226 J * cunha ~adray@109.102.broadband6.iol.cz
1161288281 M * Bertl hmm ..
1161288350 M * daniel_hozac i get crashes with 2.0.2.2-rc2 as well.
1161288375 M * Bertl well, probably we have a bug somewhere, do we have traces yet?
1161288446 M * daniel_hozac not yet, will try to get it.
1161288455 M * daniel_hozac only seems to happen with guests that run their own init though.
1161288470 M * ensc Bertl: it's not much; no stack trace
1161288471 M * Bertl hmm, ensc reported fake init restart, IIRC
1161288480 M * ensc process is in minit (the faked init)
1161288487 M * daniel_hozac i got a really long trace.
1161288487 P * cunha 
1161288492 M * ensc and in __wake_up_snc
1161288498 M * daniel_hozac (i.e. the beginning had already scrolled off the screen)
1161288509 M * daniel_hozac i'll enable serial/net console.
1161288515 M * Bertl okay
1161288568 M * Bertl ensc: do you think you could test with devel? would be interesting if that shows the same issues, or if we just messed it up for stable rc
1161288634 M * ensc Bertl: bad... it happens on my workstation which is not prepared for experimental kernel recompilations
1161288658 M * Bertl ensc: and rc1 was fine? 
1161288667 M * ensc dunno
1161288675 M * ensc I tested the last -tX only
1161288677 M * Bertl okay, which version was fine?
1161288801 M * ensc -t8
1161288809 M * Bertl so that was 2.6.18 then
1161288813 M * ensc and kernel 2.6.18
1161288830 M * Bertl okay, I'll check what changed between them ..
1161288880 M * daniel_hozac WARN_ON won't kill the process that caused it, will it?
1161288887 M * Bertl nope
1161288893 M * daniel_hozac that's what i thought.
1161288901 M * daniel_hozac kernel is ready now, i'll reboot and reproduce again.
1161288909 M * Bertl excellent, tx!
1161289009 M * WorkRoey Bertl:  (it's compiling now)
1161289035 M * Bertl ensc: which arch x86 or x86_64?
1161289044 M * ensc x86
1161289064 M * ensc Bertl: http://ensc.de/kernel-kosh-config
1161289065 M * daniel_hozac x86 here too.
1161289074 M * ensc HT, PREEMPT
1161289091 M * Bertl ok, tx!
1161289149 M * daniel_hozac hmm, netconsole didn't work, will have to enable serial console.
1161289186 M * Bertl ensc: is it a HT cpu? daniel_hozac SMP/SMT?
1161289196 M * ensc Bertl: yes
1161289205 M * daniel_hozac kernel is SMP, but it's a single CPU with no HT.
1161289212 M * Bertl okay
1161289438 Q * bonbons Quit: Leaving
1161289616 M * Bertl hmm, hmm .. wasn't there something with misalignment on xchg?
1161289730 M * ensc misalignment on x86?
1161289758 M * Bertl shouldn't apply here, because the task struct is the same in stable vs devel
1161289770 M * daniel_hozac soft lockup.
1161289829 M * daniel_hozac http://daniel.hozac.com/tmp/minicom.cap
1161289873 M * Bertl thats interesting ...
1161289916 M * Bertl could you addr2line the following addresses for me?
1161289967 M * Bertl c011ebb8, c01053c7, c01281c0
1161289995 M * daniel_hozac include/linux/list.h:32
1161289996 M * daniel_hozac arch/i386/kernel/irq.c:130
1161289996 M * daniel_hozac kernel/signal.c:1911
1161290026 M * daniel_hozac looks like something is trying to kill init while it still has children.
1161290041 M * Bertl yep, something like that
1161290056 M * Bertl I suspect we are too permissive regarding signals
1161290072 M * Bertl probably the guest shutdown manages to kill off init early or so
1161290090 M * daniel_hozac i'm not even sure what's causing the halt.
1161290111 M * daniel_hozac it is my more-broken-than-not Gentoo guest though, so it could be some guest misconfiguration.
1161290121 M * daniel_hozac (i get this when i run vserver gentoo start)
1161290126 M * Bertl keep a backup :)
1161290157 J * _are_ ~are@62.112.159.81
1161290161 A * ensc gets it at 'vserver ... stop'
1161290174 M * daniel_hozac yeah, something in the start process is calling halt.
1161290174 M * ensc (but not such a nice backtrace)
1161290182 M * Bertl ensc: yeah, would be interesting to see a trace from you too
1161290191 M * daniel_hozac i see vshelper 8002 halt in the process list before the box crashes.
1161290234 M * ensc Bertl: I did not got a stacktrace overall; just 'Process: minit', some hex machine code and then 'console shutdown'
1161290295 M * Bertl but you have minit running inside, yeah?
1161290313 M * ensc yes, that's the init process (pid 1)
1161290318 M * Bertl (so with an init process, not just fake init)
1161290326 M * ensc that's fake init
1161290335 M * Bertl nope, that's a real init :)
1161290377 M * Bertl but the question is, is it a process _inside_ the guest or not?
1161290410 M * ensc Bertl: http://paste.linux-vserver.org/519
1161290423 M * ensc yes, it's inside the guest
1161290438 M * Bertl so not what the kernel calls fake init then :)
1161290730 M * Bertl okay, I have some potential candidates for this
1161290750 M * Bertl but no real explanation yet
1161290907 M * ensc .... git bisect ... ;)
1161290918 M * Bertl bertl bisect :)
1161290923 M * Bertl *aaah*
1161290945 M * Bertl http://vserver.13thfloor.at/Stuff/delta-2022-optA.diff
1161290946 N * _mcp mcp
1161290950 M * Bertl http://vserver.13thfloor.at/Stuff/delta-2022-optB.diff
1161290960 M * Bertl try to revert one of those
1161290965 M * derjohn does this stuff affect dvl-rc42, too ?
1161290974 M * Bertl IMHO no
1161290980 M * derjohn H ?
1161290982 M * derjohn :)
1161290993 M * Bertl it looks like we 'missed' something on the backport
1161291010 M * Bertl interesting thing is, that it seems to work under certain conditions
1161291025 M * Bertl i.e. works quite fine (including start/stop) on my powerbook
1161291036 M * daniel_hozac i can't reproduce it with a non-messed up guest.
1161291057 M * Bertl what does the messed up guest start/do?
1161291073 M * Bertl could we strace -fF into the startup?
1161291075 M * daniel_hozac i think it calls halt -f early during boot.
1161291086 M * Bertl halt hmm ...
1161291099 M * Bertl so we have a real init and a halt -f then?
1161291110 M * Bertl that could be the same for ensc, right?
1161291111 M * daniel_hozac right.
1161291182 M * Bertl with the reboot kill enabled?
1161291200 M * daniel_hozac not here.
1161291246 M * Bertl but something like this works, or?
1161291249 M * Bertl chcontext --xid 100 --flag fakeinit -- bash -c "bash -c 'halt -f'"
1161291286 M * daniel_hozac indeed.
1161291303 M * Bertl what about:
1161291305 M * Bertl chcontext --xid 100 --flag fakeinit -- bash -c "sleep 5 & bash -c 'halt -f'"
1161291326 M * daniel_hozac that causes the crash.
1161291346 M * Bertl excellent, let me see if I can get that with qemu too
1161291421 M * Bertl yep, even this is sufficient
1161291422 M * Bertl chcontext --xid 100 --flag fakeinit -- bash -c "sleep 5 & halt -f"
1161291452 M * Bertl same for devel
1161291470 M * Bertl I'll upload the traces
1161291482 M * derjohn ???
1161291484 P * Rich_Estill Leaving
1161291490 M * derjohn Bertl, devel is affected?
1161291517 M * Bertl yep
1161291519 M * derjohn i was just about type reboot for testing a rc41 ...
1161291538 M * derjohn well *whine* ... I'll compile again ! :/
1161291556 M * derjohn will you provide rc42 (42 , nice :) ) ?
1161291799 M * Bertl http://paste.linux-vserver.org/520
1161291857 J * ntrs_ ~ntrs@68-188-51-87.dhcp.stls.mo.charter.com
1161291874 M * Bertl note: I'm recompiling my kernel now, as the qemu test run shows strange kernel build versions 
1161292067 Q * dna_ Quit: Verlassend
1161292220 M * daniel_hozac seems there are problems with the history tracing as well?
1161292252 M * Bertl  yup
1161292266 Q * ntrs Ping timeout: 480 seconds
1161292710 J * Aiken ~james@tooax6-204.dialup.optusnet.com.au
1161292723 M * Bertl welcome Aiken!
1161292747 M * Aiken hi Bertl
1161293337 J * Osgiliath ~abigor@kurzweil.no-ip.org
1161293346 M * Bertl welcome Osgiliath!
1161293356 M * Osgiliath hi Bertl
1161293383 M * Bertl daniel_hozac: http://paste.linux-vserver.org/521
1161293399 M * Osgiliath 2 day ago, I posted a mail in the mailing list about Oracle 10g, i don't know if you remember ...
1161293407 M * Bertl yup
1161293423 M * Bertl I suggested to strace, IIRC
1161293432 M * Osgiliath yep
1161293452 M * Osgiliath i successed in the install, but after 2 reboots, it doesn't start !
1161293459 M * Osgiliath so i just did a strace
1161293476 M * Osgiliath and i got all this stuff :
1161293477 M * Osgiliath 3759  open("/oracle/app/oracle/product/lib/tls/i686/mmx/cmov/libclntsh.so.10.1",,
1161293477 M * Osgiliath  O_RDONLY) = -1 ENOENT (No such file or directory)
1161293477 M * Osgiliath 3759  stat64("/oracle/app/oracle/product/lib/tls/i686/mmx/cmov", 0xbf97d418) = --
1161293477 M * Osgiliath 1 ENOENT (No such file or directory)
1161293478 M * Osgiliath 3759  open("/oracle/app/oracle/product/lib/tls/i686/mmx/libclntsh.so.10.1", O_RDD
1161293478 M * Osgiliath ONLY) = -1 ENOENT (No such file or directory)
1161293506 J * vil ~vladimir@catv-tisnov-86.selfnet.cz
1161293511 M * Bertl Osgiliath: okay, and they are supposed to be there?
1161293514 M * Bertl welcome vil!
1161293528 M * Osgiliath the libs are supposed to be in /oracle/app/oracle/product/lib/ 
1161293540 M * vil hello Bertl
1161293563 M * Bertl Osgiliath: maybe, but what I wonder is, did they get lost or did they miss since ever?
1161293630 M * Osgiliath hum it's not possible because Oracle started after the install !
1161293653 M * Osgiliath and i stopped and restarted oracle few times
1161293659 M * Bertl ah, now you are jumping to conclusions :)
1161293705 M * Bertl Osgiliath: try for example the following: 'strace -fF true 2>&1 | grep NOENT'
1161293796 M * Osgiliath access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
1161293796 M * Osgiliath open("/etc/ld.so.preload", O_RDONLY)    = -1 ENOENT (No such file or directory)
1161293796 M * Osgiliath open("/oracle/app/oracle/product/lib/tls/i686/mmx/cmov/libc.so.6", O_RDONLY) = -1 ENOENT (No such file or directory)
1161293796 M * Osgiliath stat64("/oracle/app/oracle/product/lib/tls/i686/mmx/cmov", 0xbfe94138) = -1 ENOENT (No such file or directory)
1161293807 M * Osgiliath and 20 others lines like these ...
1161293815 M * Bertl so does true fail?
1161293908 M * Osgiliath hu... i don't understand, sorry 
1161293941 M * Bertl if you execute 'true' does it fail somehow?
1161294034 M * Osgiliath ah ok sorry, yes it does work !
1161294050 M * Bertl okay, so the ENOENT on various libraries are 'normal'
1161294072 M * Osgiliath i don't know what 'true' does, but it does nothing ... (no error)
1161294078 M * Bertl i.e. no indication of a fault per se, unless you _know_ they they were there before
1161294092 M * Bertl true just returns with exit(0)
1161294105 M * Bertl (in the beatiful version of glibc here :)
1161294150 M * Osgiliath ok, so the 'No such file or directory' is not the problem there
1161294161 M * Bertl as I said, not necessarily
1161294180 M * Bertl if you _know_ the library xy was there before, it _might_ be an issue
1161294236 M * Osgiliath they were not
1161294311 P * vil 
1161294411 M * Osgiliath so, are there any way to find the problem with strace ?
1161294428 M * Bertl well, if you upload the strace somewhere, we could have a lookat it :)
1161294473 M * Osgiliath great, i'll do that
1161294530 M * Osgiliath http://kurzweil.no-ip.org/oracle.trace2
1161294620 M * Osgiliath it seams there is a link with DNS no ?
1161294660 M * Bertl yes, but it sees to me like the process gets a signal and exits
1161294674 M * Bertl so nothing spectacular actually
1161294679 M * Osgiliath yes, that's me
1161294695 M * Osgiliath because the prog is freezed
1161294714 M * Osgiliath maybe i shouldn't interrupt it :/
1161294789 M * Bertl hmm :)
1161294812 M * Osgiliath i upload a new one, oracle.trace3
1161294833 M * Osgiliath and i didn't interrupt the program this time
1161294935 M * Osgiliath the log doesn't change anymore, so you can look oracle.trace3
1161294954 M * Bertl well, it is waiting for another thread/action
1161294975 M * Osgiliath hum, and what is it ?
1161294986 M * Bertl it is the process you are stracing
1161295015 M * Bertl so no bad behaviour so far ... maybe some server/client should answer but doesn't?
1161295047 M * Osgiliath i don't think so 
1161295047 M * Bertl no idea what the intuitively named 'lsnrctl' is supposed to do?!
1161295057 M * Osgiliath because it is a listener
1161295063 M * Osgiliath which listen port 1521
1161295074 M * Bertl fine, so it is listening now?
1161295078 M * Osgiliath nop
1161295089 M * Bertl hehe, how do you know?
1161295109 M * Bertl what I see here is, that it is waiting on a futex
1161295116 M * Osgiliath 23:58 root@oracle /oracle/app/oracle/product# netstat -at | grep 1521
1161295116 M * Osgiliath zsh: done       netstat -at |
1161295116 M * Osgiliath zsh: exit 1     grep 1521
1161295132 M * Bertl so once that futex will get released, it will continue
1161295145 M * Osgiliath ok
1161295154 J * samueltc ~samuel@adsl-68-122-125-205.dsl.pltn13.pacbell.net
1161295158 M * samueltc hi!
1161295166 M * Bertl Osgiliath: now it might be that we are dealing with some futex issue here, but could as well be an oracle issue
1161295170 M * Bertl welcome samueltc!
1161295195 M * Osgiliath yes ...
1161295209 M * Osgiliath it's not a good news !
1161295213 M * Bertl do you have the source code at hand?
1161295238 M * Osgiliath just of the listener ?
1161295263 M * Bertl well, for a start ...
1161295279 M * samueltc anyone here is working on snmpd agent to monitor vservers?
1161295288 M * samueltc also what about the progress of the monitoring api?
1161295289 M * Osgiliath hum, i should find this in the oracle setup files ...
1161295293 M * Bertl Osgiliath: you could try to recreate the futex stuff in a simple test case
1161295349 M * Osgiliath hu, how can i do this ?
1161295381 M * Bertl well, there is some code using that futex
1161295394 M * Bertl (one code part in the client and another one in the server)
1161295452 M * samueltc http://collectd.org/plugins/vserver.shtml is nice
1161295528 M * hardwire` any thoughts on thawing a vserver?
1161295537 M * hardwire` yes.. flame me.. I care not
1161295555 M * Bertl hmm, you mean after freezing it? how?
1161295564 M * hardwire` :)
1161295583 M * hardwire` just freezing the state of all the memory its mapped to.. storing to file.. and assuming storage won't change
1161295596 M * Bertl hehe, nah, not a good idea
1161295603 M * hardwire` whazzu
1161295613 M * Bertl requires too much virtualization to realize
1161295646 A * hardwire` hangs head
1161295653 M * hardwire` thats sad
1161295664 M * hardwire` and.. why do I have a little booger after my nick?
1161295667 N * hardwire` hardwire
1161295684 M * Bertl you can use Xen for this, should work fine
1161295851 M * hardwire yup
1161295856 M * hardwire Xen is fine for it
1161295859 M * Osgiliath i'm sorry Bertl, i did'nt find the source code ...
1161295862 M * hardwire just a wee bit pokier
1161296354 M * Bertl daniel_hozac: here is the history dump isolated: http://paste.linux-vserver.org/528
1161296430 M * Bertl it basically means that the per cpu history buffer is invalid :(
1161296632 M * Bertl hmm, maybe per_cpu is not defined for all configured cpus anymore
1161296807 M * Osgiliath Bertl, can you explain why the listener starts in a freshly build vserver ?
1161296831 M * Osgiliath because i find this very strange !
1161296840 M * Bertl maybe because the futex doesn't exist yet?
1161296846 M * Bertl i.e. it is not locked
1161296930 M * Hollow Bertl: what are the vc_{g,s}et_pflags syscalls supposed to do?
1161296953 M * Bertl they will (in the near future) allow to set process flags
1161296956 M * Osgiliath ok, and this futex is not unlocked when the vserver restart ?
1161296971 M * Hollow Bertl: e.g.?
1161296978 M * Bertl Osgiliath: nope
1161296981 M * Hollow or are process flags a vanilla feature?
1161297003 M * Bertl Hollow: again, in the future, e.g. 'spectator' flags
1161297027 M * Hollow ok :)
1161297064 M * Osgiliath therefore i'm fucked ;)
1161297079 M * hardwire spectator?
1161297105 M * Bertl Osgiliath: well, no, actually you could write a little tool (similar to what the server does) to unlock the futex
1161297141 M * Bertl or, if you do not care about the wasted memory
1161297152 M * Osgiliath yeah but i really don't know about futex ...
1161297152 M * Bertl simply change the context (xid) after shutdown
1161297158 M * Osgiliath hum
1161297170 M * Osgiliath i'll try this ...
1161297174 M * Bertl they should be properly isolated, so that might help
1161297182 M * notasnark Hollow: Your suggestion earlier to re-install baselayout-vserver worked a charm, and my vserver is now back up and running. Thanks a lot.
1161297240 M * Hollow notasnark: good to hear :) if your guest uses the vserver profile, it should not happen to install baselayout on updates if it was not installed previously
1161297284 M * Bertl daniel_hozac: yep, with 1 cpu, access to cpu=1 kills it, with 3 cpus, access to cpu=3 (4th cpu)
1161297329 M * Bertl (this probably changed with 2.6.18)
1161297349 M * notasnark I think the problem was that I followed the instructions on how to upgrade to the new gcc, and that probably said to set things to use the standard 2006.1 profile - which I did without thinking. My system upgrade then broke everything!
1161297376 M * Osgiliath Bertl : I stopped the vserver, changed the context number, restarted it, and it does the same thing ... :(
1161297412 M * Bertl what kernel version/patches do you use?
1161297424 M * Osgiliath 2.6.17.13-vs2.0.2.1
1161297489 M * Osgiliath and i write the context id in the /etc/vservers/oracle/context
1161297689 M * Bertl okay, would be nice if you could dig out some futex tests
1161297703 M * Bertl I'm 100% sure google will find some test code for that
1161297734 M * Bertl and try to reproduce the issue, that would speed up the debugging process a lot
1161297763 M * Osgiliath ok, so i need to find a test code which unlock the futex ?
1161297784 M * Bertl well, for your purpose, the unlock would help, I guess
1161297799 M * Bertl but I think it might be an isolation issue in linux-vserver
1161297812 M * Bertl what I wonder is, what happens on a real linux system?
1161297824 M * Bertl are you screwed after the futex is locked too?
1161297845 M * Bertl (i.e. do you have to reboot then, or is it mended somehow)
1161297866 M * matti :)
1161297880 M * matti Greetings from Poland  :)
1161297893 M * Bertl huh? back?
1161297898 M * matti Yep.
1161297901 M * Osgiliath hmm, reboot doesn't change anything
1161297904 M * Hollow Bertl: in virt_stat vc_data.offset refers to the virtualized guest time feature, right?
1161297905 M * matti I take one day off.
1161297913 M * Bertl Osgiliath: hmm?
1161297913 M * Hollow hi matti :)
1161297923 M * matti I need to sort out some of my problems in Poland ;]
1161297928 M * matti Hollow: :)
1161297930 M * Bertl Osgiliath: please elaborate ...
1161297930 M * Osgiliath no , really :(
1161297935 M * Osgiliath hm
1161297945 M * matti Bertl: Easyjet++ ;]
1161297951 M * Bertl Hollow: yep
1161297951 M * gdm  matti you went back for one day?
1161297954 M * matti Yes.
1161297964 M * gdm wow
1161297971 M * matti gdm: My company is paying for my trip, so I don't care ;p
1161297972 M * Osgiliath if i reboot the host, the futex is still locked 
1161297988 M * Bertl Osgiliath: ah, no, that's not possible :)
1161297994 M * gdm hehe, kool. i am going to bed now. good night all
1161298003 M * Bertl Osgiliath: in this case _something_ must have already 'locked' it
1161298018 M * matti gdm: Sweet dreams ;)
1161298018 M * Osgiliath humpf !
1161298041 M * Bertl Osgiliath: the futexes are kernel structures, they vanish on reboot
1161298047 M * Bertl (at least on host reboot)
1161298049 M * Osgiliath ok
1161298053 M * matti k, I am tired also - good night ;]
1161298057 P * stefani I'm Parting (the water)
1161298068 M * Bertl have a good one gdm! matti!
1161298080 M * Osgiliath so, maybe this is not the problem, is it ?
1161298107 M * matti Bertl: Thanks :)
1161298127 J * comfrey ~comfrey@h-64-105-215-75.sttnwaho.covad.net
1161298147 M * Bertl wb comfrey!
1161298150 M * Hollow Bertl: any reason why you moved sock_acc_field from cacct_def.h to cacct.h? (in contrary to vhi_name_field which is in *_cmd.h)
1161298203 M * Osgiliath but i cannot understand how the listener could start twice (just after install), and after host or guest reboot, freezed !
1161298227 M * Bertl Hollow: yes, the vhi_name_field is not used outside of cmd code
1161298243 M * Bertl Hollow: but the sock_acc enum is used elsewhere too
1161298254 M * Hollow ok..
1161298261 M * Bertl Osgiliath: let me give a simple example:
1161298301 M * Bertl - check for file xy, if it is there, read it, and increment the value
1161298310 M * Bertl - if it is not there, set the value to 0
1161298317 M * Bertl - write the value back to file xy
1161298329 M * Bertl - if value is >2, lock the futex
1161298335 M * Osgiliath yep
1161298348 M * Bertl so, granted that's constructed and would be stupid
1161298365 M * Bertl but I think the truth is somewhat similar
1161298391 M * Bertl i.e. the oracle _changes_ something, maybe some changes are bad or need 'special' undoing on reboot
1161298413 M * Bertl I would suggest to do the following, if you want to narrow it down
1161298419 M * Osgiliath yep
1161298427 M * Bertl - create a new guest with a brand new oracle install
1161298444 M * Osgiliath yep, i did it twice ! :p
1161298447 M * Bertl - make a snapshot (e.g. via lvm) of the filesystem before you start anything
1161298458 M * Osgiliath i dit it as well
1161298464 M * Bertl - kick the oracle (I assume it works now)
1161298481 M * Bertl - go back to the snapshot part
1161298493 M * Bertl iterate as many times as it 'works'
1161298508 M * Bertl then check what changed in the guest layout
1161298534 M * Osgiliath ok, i see
1161298555 M * Bertl alternatively, the following could be helpful:
1161298573 M * Bertl - strace -fF -o _every_ oracle process involved (server, client, whatever)
1161298592 M * Bertl - compare two traces, one which 'works' and one which 'fails'
1161298607 M * Osgiliath yes
1161298860 M * Osgiliath i think i'll do the second one ;p
1161298881 M * Bertl okay, just make sure that you do not miss any of those processes
1161298892 M * Bertl i.e. check/look for runlevel scripts and such
1161298916 M * Bertl I can help you somewhat with analyzing the differences
1161298916 M * Osgiliath yes, no problem because i did it ;p
1161298923 M * Osgiliath ok
1161298973 M * Osgiliath for now, thanks a lot !
1161298980 M * Bertl you're welcome!
1161298990 M * Osgiliath you're doing great work with VS !
1161298998 M * Bertl thanks!
1161299117 J * s0undt3ch_ ~s0undt3ch@81.193.56.46
1161299132 Q * s0undt3ch Killed (NickServ (GHOST command used by s0undt3ch_))
1161299135 N * s0undt3ch_ s0undt3ch
1161299147 M * Osgiliath maybe I should install PostgreSQL :)))
1161299158 J * _dmax ~semaj@81.193.56.46
1161299232 M * Bertl wb _dmax!
1161299243 M * Bertl Osgiliath: could help a lot :)
1161299502 Q * dmax Ping timeout: 480 seconds
1161299506 N * _dmax dmax
1161299735 M * Hollow Bertl: ok, the new stat syscalls are in libvserver svn and tested, seems all well
1161299754 M * Bertl great!
1161299786 M * Bertl Hollow: so you have now everything you need (regarding data)
1161299845 M * Hollow yep, all data is now available via syscalls
1161299865 M * Bertl excellent!
1161299880 M * Hollow indeed :)
1161299882 M * Hollow thanks a lot
1161299902 M * Bertl you're welcome! thanks for your time and work on it too!
1161299935 M * Hollow i guess vstatd will be adapted over the weekend, i will report back then and show some nice graphs etc :)
1161299960 M * Bertl good, maybe even have a posting on the ML?
1161299987 M * Hollow yep, will do... vstatd is independent from vcd now anyway since it got its own package
1161299994 M * Hollow i.e. it can be used with util-vserver too
1161300014 M * Bertl great, probably folks will create pakages too
1161300034 M * FaUl is there any schedule for 2.1.1 yet? i'd realy like to test some of the next releases with ipv6-patches applied
1161300053 M * Bertl 2.1.1 should be already out, but it isn't :)
1161300060 M * Hollow heh
1161300077 A * Hollow is already looking forward to the legacy cleanup in 2.2
1161300096 M * FaUl so it is expected 'really soon'?
1161300106 M * Bertl yep
1161300117 M * FaUl excellent ;-)
1161300132 M * FaUl </burns>
1161300147 M * Bertl hehe
1161300167 M * Hollow off to bed now.. cu tomorrow
1161300178 M * derjohn2 Hollow, bye 
1161300214 M * FaUl so, i'm going to bed now, first lecture is schedeuled for 8:15 so I have to get up at about 6:30 which is roughly 5 hours ;-)
1161300230 M * Bertl k, have a good one!
1161300234 M * derjohn2 FaUl, I have a debian kernel with v6 support ready. also a dpatch-tree.
1161300249 M * derjohn2 i just wait for rc42 before putting it public
1161300270 M * FaUl derjohn2: i personally never used distribution-kernels...
1161300286 A * FaUl is in bed now, anyway
1161300296 M * derjohn2 Bertl, FaUl : but integrating the V6 stuff in devel, i a good idea. but after 2.1.1 
1161300329 M * Bertl for those still awak/with us: here is the updated oops after fixing the history trace: http://paste.linux-vserver.org/529
1161300336 M * Bertl *awake
1161301069 M * hardwire can I pick on the vserver logo?
1161301087 M * hardwire is the V like a big green checkmark?
1161301101 M * Bertl yep
1161301120 M * hardwire weird
1161301184 M * Bertl de gustibus non est disputandum
1161301452 Q * crypt Quit: Leaving
1161301785 M * hardwire why is that?
1161301978 M * derjohn2 hardwire, latin
1161302007 M * hardwire not what is that.
1161302013 M * hardwire why is it non disputable
1161302021 M * derjohn2 ah
1161302048 M * derjohn2 yes .... because it's a personal preference I guess ;)
1161302057 M * hardwire http://www.getac.com/pages/products/prod_M220.html
1161302060 M * hardwire mmm
1161302061 M * hardwire wow
1161302069 M * hardwire I had a mil spec toughbook at my last job
1161302075 M * hardwire I don't like having a dell now
1161302078 M * hardwire it looks so breakable