1315529318 Q * nox Quit: If the world is without walls and fences -- who needs Windows and Gates ?
1315529318 N * n01101111x nox
1315530546 J * jeroen__ ~jeroen@imap.powerinternet.eu
1315535053 M * Bertl off to bed now ... have a good one everyone!
1315535057 N * Bertl Bertl_zZ
1315542117 M * hparker Problem with memory limits 2.6.39.4-vs2.3.1-pre9.2 that doesn't make sense http://dpaste.com/611622/
1315544050 J * sannes ~ace@cm-84.209.106.118.getinternet.no
1315545560 J * ghislain ~AQUEOS@adsl2.aqueos.com
1315545813 Q * FireEgl Ping timeout: 480 seconds
1315546383 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com
1315546850 J * thierryp ~thierry@home.parmentelat.net
1315547369 Q * thierryp Remote host closed the connection
1315547835 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no
1315552226 J * thierryp ~thierry@zankai.inria.fr
1315553875 J * Aiken_ ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1315553876 Q * Aiken Read error: Connection reset by peer
1315554814 J * ghislain1 ~AQUEOS@adsl2.aqueos.com
1315555042 Q * ghislain Ping timeout: 480 seconds
1315555734 J * derjohn_mob ~aj@ip-81-210-228-102.unitymediagroup.de
1315562887 M * karasz moin
1315562924 M * karasz with:
1315562935 M * karasz util-vserver 0.30.216-pre2883 master
1315562937 M * karasz and
1315562948 M * karasz 2.6.32.41-vs2.3.0.36.29.7-dist
1315562959 M * karasz cgroups for cpu works fine
1315562994 M * karasz do I need http://vserver.13thfloor.at/ExperimentalT/delta-memcg-fix04.diff forapplying cgroups for  memory ?
1315564035 Q * derjohn_mob Ping timeout: 480 seconds
1315566330 N * Bertl_zZ Bertl
1315566334 M * Bertl morning folks!
1315567002 M * karasz hello Bertl 
1315567115 M * Bertl the memcg fix is for 2.6.34 and later, so shouldn't be needed for 2.6.32
1315571340 Q * ghislain1 Quit: Leaving.
1315572520 J * derjohn_mob ~aj@ip-81-210-228-102.unitymediagroup.de
1315573941 J * ghislain ~AQUEOS@adsl2.aqueos.com
1315576737 M * Mr_Smoke What does that do again ?
1315576750 M * Mr_Smoke (i'm running 2.6.38.*)
1315576805 M * Mr_Smoke In other words ,how necessary is that ?
1315576960 Q * fisted Ping timeout: 480 seconds
1315577201 J * fisted ~fisted@xdsl-87-78-215-162.netcologne.de
1315577476 Q * Aiken_ Remote host closed the connection
1315577898 M * Mr_Smoke Bertl: ? what does that memcg patch do exactly ?
1315577915 M * Mr_Smoke and has it been integrated in one of the later patches ?
1315578535 M * Bertl it adapts the patch to a mainline change
1315578551 M * Bertl (which went unnoticed for one release)
1315578573 M * Bertl so yes, it is included in later patches
1315578996 M * Mr_Smoke ok
1315579125 J * davidquist 2e0ec06f@ircip3.mibbit.com
1315580016 J * dowdle ~dowdle@scott.coe.montana.edu
1315580167 J * fback_ fback@red.fback.net
1315580269 Q * fback Read error: Connection reset by peer
1315580930 M * Rockj Bertl: http://pastie.org/2508892
1315580943 M * Rockj vc_set_sched, something missing in normal kernel config?
1315580956 Q * ncopa Quit: Leaving
1315581307 M * Rockj using vserver 0.30.215 - and  absint 3.0.4-vs2.3.1-pre10.1
1315582082 M * Rockj :-/
1315582986 M * ard you know what is coming now?
1315583008 M * ard to use 3.0.4-vs2.3.1-pre10.1, you need a recent vserver-utils package
1315583037 M * davidquist Hi guys, i have a small issue with an init script (/etc/init.d/oracle) of a guest. I am not able to stop oracle database services using vserver <guest> stop method. The init script which executes "su - oracle -c 'cd $RUN; ctl stop all'" runs however successfully within the guest or with vserver exec method.
1315583049 M * davidquist With vserver stop/restart method I get "could not open session" displayed after "su - oracle -c 'cd $RUN; ctl stop all'" is executed. Can you assist me in troubleshooting the issue. A detailed log including strace output can be found here: http://dev.quist.ch/vserver/start_oracle_issue.txt
1315583253 M * Guy- davidquist: a workaround would be to modify that line so it's like cd $RUN; sudo -H -u oracle ctl stop all
1315583286 M * Guy- davidquist: I don't know oracle -- maybe it also needs some envvars to be set, but you can then sudo ... env var1=val1 ... ctl stop all
1315583294 M * ard it's a problem with su
1315583299 M * Guy- yes,
1315583308 M * Guy- that's why I'm suggesting using sudo instead
1315583316 M * ard ah, ok ;-)
1315583352 M * ard maybe it's ok to put into the pam config that su doesn't need a session?
1315583379 M * davidquist how could the be configured?
1315583385 M * Guy- that may have complicated implications
1315583394 M * ard there is /etc/pam.d/su
1315583401 M * ard it includes various other files
1315583408 M * Guy- davidquist: /etc/pam.d/su, remove the line with session required pam_unix.so oslt
1315583409 M * ard you might want to take a look at the session part
1315583421 M * Guy- or @include common-session or whatever
1315583447 M * Guy- but this may have adverse effects if you use su for any other purpose too
1315583463 M * Guy- (like, resource limits wouldn't be set up when you su to a user)
1315583478 M * ard yes :-(. But on the bright side: if you use vserver, you probably only use oracle on that vserver
1315583487 M * Guy- yes
1315583492 M * Guy- still, I think sudo is cleaner
1315583495 Q * derjohn_mob Ping timeout: 480 seconds
1315583499 M * ard Me too
1315583503 M * ard sudo rocks
1315583517 M * ard better control of what can be done
1315583522 M * davidquist exactly just one service per vserver instance
1315583532 M * davidquist let me check /etc/pam.d/su and sudo
1315583733 M * davidquist one strange thing is that with starting the oracle services there is no issue with su
1315583775 M * davidquist what is different is that there are open files while trying to stop the services
1315583781 M * davidquist lsof | wc -l 
1315583794 M * davidquist returns 3800 open files
1315584212 Q * thierryp Remote host closed the connection
1315584717 J * bonbons ~bonbons@2001:960:7ab:0:d0db:a6f7:9fe2:f95d
1315585217 M * ser guys, if i set ionice 3 for few vservers and i do not set anything for the rest, what is the ionice for them? i can read "none" - but this magick word "none" means 1, 2 or 3? :) :)
1315585265 M * ser i am aksing because i have an impression it means 4
1315585287 M * ser as it looks they slowed down a lot
1315585382 M * Rockj ard: but 30.215 is newest vserver-utils as far as I see
1315585977 M * daniel_hozac ser: what?
1315585988 M * daniel_hozac ionice requires two values. what did you set for which one?
1315586001 M * geb ser,               Note  that  before kernel 2.6.26 a process that has not asked for an io priority formally uses "none" as scheduling class, but the io scheduler will treat such pro-
1315586002 M * geb               cesses as if it were in the best effort class. The priority within the best effort class will be dynamically derived from the cpu nice level of the process: io_pri-
1315586002 M * geb               ority = (cpu_nice + 20) / 5.
1315586008 M * geb (man ionice)
1315586038 M * ser geb: it is 2.6.26 
1315586067 M * ser and what precisely mean none? i still do not get it :)
1315586102 M * ser daniel_hozac: i have not set any for few vservers
1315586105 M * daniel_hozac 2.6.26? that kernel never really worked well.
1315586122 M * ser daniel_hozac: sorry!
1315586136 M * ser 2.6.36 of course
1315586196 M * ser and i have a feeling they are slower then these vservers i set ionice 3
1315586309 M * daniel_hozac ionice 3 doesn't really mean anything.
1315586319 M * daniel_hozac ionice requires a class and a priority.
1315586321 M * daniel_hozac what did you set?
1315586333 M * ser ionice 3 does not require class
1315586408 M * ser for a process on a vserver i have not set anything i can read:
1315586410 M * ser # ionice 26146
1315586410 M * ser none: prio 4
1315586410 M * ser none: prio 4
1315586446 M * ser for a process inside a vserver with 3 i can read:
1315586449 M * ser # ionice 15730
1315586449 M * ser idle
1315586449 M * ser idle
1315586473 M * ser and i have a feeling A is slower in drive access then B
1315586629 M * ser and i am pretty confused :)
1315587166 M * Bertl Rockj: sorry, my irc timed out ... but yes, 0.30.215 is many years old, so it won't work with a recent kernel (it won't work properly with most kernels we currently have patches for)
1315587290 M * Bertl ser: it's not just I/O class/niceness which controlls the actual I/O speed/behaviour, the I/O path itself is quite important as well
1315587340 M * Bertl for example, it won't help if A is in a realtime class but shares I/O with 100 other processes, while B is on its own I/O path with a -c3
1315587404 M * hparker I'm having a problem with memory limits in 2.6.39.4-vs2.3.1-pre9.2 with util-vserver 0.30.216_pre2987 that doesn't make sense http://dpaste.com/611622/
1315587452 M * Rockj Bertl: so which util packaage should I compiile?:)
1315587462 M * Rockj the one linking from the frontpage is the same version
1315587762 M * daniel_hozac hparker: and the problem is?
1315587778 M * daniel_hozac Rockj: 0.30.216-pre2987 is what you want.
1315587916 M * hparker daniel_hozac: no memory.* In the cgroup directory
1315587942 M * daniel_hozac did you mount the cgroups with memory enabled?
1315588011 M * hparker none                    /dev/cgroup cgroup              cpu,cpuset,memory       0 2
1315588033 M * daniel_hozac and cat /proc/mounts shows that?
1315588063 M * hparker hrrmm... No :P
1315588086 M * hparker But I'm pretty sure it does with .35, let me reboot it
1315588924 M * hparker Ok, so there's something strange going on with my mount... Thanks for kicking me in the right direction... Now to figure out why on boot it doesn't mount properly
1315588992 M * Bertl Rockj: the front page links the most recent stable version as well as the development versions#
1315589147 M * Bertl in general, stable util-vserver version works with the stable kernel patch
1315590160 M * davidquist Hey guys, i found the root cause of my su issue. I had to remove on the guest the oracle recommended entries from /etc/security/limits.conf. Strace showed before messages like:
1315590173 M * davidquist setrlimit(RLIMIT_NOFILE, {rlim_cur=128*1024, rlim_max=128*1024}) = -1 EPERM (Operation not permitted)
1315590183 M * davidquist setrlimit(RLIMIT_MEMLOCK, {rlim_cur=3500000*1024, rlim_max=3500000*1024}) = -1 EPERM (Operation not permitted)
1315590221 M * davidquist without those entries everything works how it should.
1315593805 M * hparker Editing /etc/vserver/.defaults/cgroup/subsys seems to have fixed everything up, thanks!
1315593883 M * ser Bertl: yes but the problem is i feel that setting ionice made things worse
1315593922 M * ser Bertl: i hoped setting ionice 3 for few most annoying servers help to put them in the background
1315593937 M * ser but it looks it does it in an opposite way
1315593970 M * ser and i wonder why - should i try to explictly set other, more important servers to class 2?
1315594048 M * ser because i am concerned about that "none" value, to be honest i do not know what precisely "none" means
1315594180 M * Rockj Bertl: http://pastie.org/2509926
1315594260 M * daniel_hozac you're missing -m for mirror.
1315594309 M * Rockj indeed
1315594310 M * Rockj thanks
1315594463 Q * nicholi Quit: leaving
1315594747 J * nicholi ~nicholi@rrcs-76-79-196-34.west.biz.rr.com
1315594970 M * Rockj think I got it up and running
1315594977 M * Rockj just some annoying locale errors etc while using apt
1315594979 M * Rockj in the guest
1315596347 M * Guy- I'm trying to get autofs working inside a guest, to mount nfs shares, and it's failing in different ways
1315596352 M * Guy- if I don't use nolock, I get in dmesg:
1315596356 M * Guy- svc: failed to register lockdv1 RPC service (errno 111).
1315596365 M * Guy- with nolock, the first mount works but the second one hangs
1315596379 M * Guy- (regardless of which of the two I do first)
1315596398 M * Guy- what am I missing? I'd rather not mount with nolock anyway
1315596411 M * Guy- also there is this: vxW: ['rpc.statd',31039:#583|583|583] denied [0x22] access to inode proc:ffff88011bcce718[#0,1733365]
1315596421 M * Guy- that file would be proc/sys/fs/nfs/nsm_local_state
1315596502 M * Guy- (the second mount hangs in autofs4_wait, fwiw)
1315596531 M * Guy- oh, at one point I also had "lockd_up: makesock failed, error=-111"
1315596557 M * Guy- this points toward some missing capability, but which one? I have the SECURE*MOUNT capabilities and SYS_ADMIN
1315597130 M * Guy- oh, 111 is ECONNREFUSED
1315597168 M * Guy- so part of my problem is probably that I don't have a portmapper on the host - I can put one there, but will that help? the guest has one...
1315597459 M * Guy- rpc.statd tries to open /proc/sys/fs/nfs/nsm_local_state for writing
1315597481 M * Guy- I suppose that fails due to /proc security
1315597769 Q * ghislain Quit: Leaving.
1315598308 M * Guy- OK, I'll try with mount --make-shared
1315598317 M * Guy- and mount the nfs stuff on the host instead
1315599108 M * Guy- well, that works at least
1315600937 Q * tomfmason Read error: Connection reset by peer
1315601405 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1315603532 Q * cuba33ci Read error: Connection reset by peer
1315603620 J * cuba33ci ~cuba33ci@111-240-170-123.dynamic.hinet.net
1315604961 Q * bonbons Quit: Leaving
1315605774 Q * Aiken Remote host closed the connection
1315606070 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f
1315608050 Q * Aiken Quit: Leaving
1315608867 Q * davidquist Quit: http://www.mibbit.com ajax IRC Client
1315611407 Q * dowdle Remote host closed the connection