1320712768 Q * PowerKe Ping timeout: 480 seconds 1320713300 Q * fisted_ Remote host closed the connection 1320713313 J * fisted ~fisted@xdsl-87-78-212-166.netcologne.de 1320713658 Q * dowdle Remote host closed the connection 1320713682 J * PowerKe ~tom@94-226-105-186.access.telenet.be 1320713702 Q * clopez Ping timeout: 480 seconds 1320714868 J * _Radiance ~Radiance@193.16.154.187 1320714893 Q * _nono_ Ping timeout: 480 seconds 1320714944 Q * Radiance Read error: Connection reset by peer 1320714956 N * _Radiance Radiance 1320715707 Q * black Ping timeout: 480 seconds 1320716442 Q * hparker Quit: Quit 1320718715 M * Bertl off to bed now ... have a good one everyone! 1320718720 N * Bertl Bertl_zZ 1320723338 Q * nkukard Ping timeout: 480 seconds 1320723994 J * nkukard ~nkukard@41-133-198-167.dsl.mweb.co.za 1320725224 Q * mrjack_ Ping timeout: 480 seconds 1320734211 J * sannes1 ~ace@cm-84.209.106.118.getinternet.no 1320734269 J * _nono_ ~gomes@licencieux.ircam.fr 1320734869 Q * fisted Read error: Operation timed out 1320738527 J * derjohn_mob ~aj@213.238.45.2 1320739130 J * ghislain ~AQUEOS@adsl2.aqueos.com 1320739326 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1320743061 M * arekm /dev/hdv1 on / type ufs (defaults) 1320743095 M * arekm while in reality this is: /dev/md5 / xfs defaults 1 1 1320743183 M * arekm why it hides dev name, fs type for rootfs? 1320743185 M * arekm /dev/md3 on /boot type ufs (defaults) 1320743186 M * arekm /dev/md6 on /home type xfs (defaults) 1320743191 M * arekm for boot it only hides fs type 1320743197 M * arekm for home everything is true 1320744266 M * glen hello arekm :) 1320746766 J * clopez ~clopez@155.99.117.91.static.mundo-r.com 1320747236 M * arekm hello 1320747286 M * daniel_hozac arekm: the root mtab entry is copied from the config. 1320747408 M * arekm daniel_hozac: what about /boot then? 1320747580 M * daniel_hozac what fs type is it originally? 1320747684 J * BenG ~bengreen@cpc12-aztw24-2-0-cust146.aztw.cable.virginmedia.com 1320749881 M * arekm daniel_hozac: ext4 1320749917 N * Bertl_zZ Bertl 1320749921 M * Bertl morning folks! 1320749930 M * arekm hmm, I have such weird behaviour with cgroups that when processes in that cgroup ate all assigned memory then it's quite hard to kill these processes, load jumps the roof and usually I have to reboot whole machine 1320749948 M * arekm oom kills some processes in that cgroup but that doesn't help much 1320750022 M * Bertl maybe file an upstream bug report, I'm pretty sure there are still some bugs left in this area 1320750022 M * arekm has anyone seen this (it's vserver unrelated) 1320750051 M * arekm and I cannot try newer kernel because newer kernel kills ipmi for me :/ 1320750079 M * Bertl did you file a bug report for that? 1320750115 M * Bertl (sounds easier to fix to me :) 1320750130 M * arekm for cgroup? no. for ipmi? yes, I would have to bisect this to get help (and it's hard to do bisecting on production machine) 1320750164 M * Bertl and you are testing cgroup stuff on a production machine? 1320750185 M * arekm it isn't marked experimental afaik 8-) 1320750235 M * arekm the problem was that processes were eating all ram (without cgroup) and thus killing the machine, so result is the same unfortunately 1320750406 M * Bertl rising load sounds like something is respawning despite the fact that it gets killed over and over 1320750608 Q * clopez Ping timeout: 480 seconds 1320750947 Q * Aiken Remote host closed the connection 1320751207 M * arekm Bertl: these processes aren't easily killable.... while (true); do killall -9 process; done and have to like 20-30s for these (60 "process" processes) to get killed 1320751571 M * daniel_hozac or the load is because it starts swapping or becoming stuck in D state waiting for memory. 1320751572 N * sid4windr sid3windr 1320751620 M * arekm no swap here, so rather the other thing 1320751666 M * daniel_hozac re mtab, ext* will get translated to ufs for quota. 1320751692 M * daniel_hozac which kernel are you on with those cgroups? 1320751704 M * arekm 2.6.38.8 1320751909 M * daniel_hozac any chance you could try 3.1? 1320752007 M * arekm 3.0 kills ipmi on this machine... so not really (have to solve this one first) 1320752091 M * daniel_hozac how does it kill ipmi? 1320752201 J * clopez ~clopez@155.99.117.91.static.mundo-r.com 1320752347 M * arekm daniel_hozac: ipmi IP becomes unreachable 1320752421 M * arekm daniel_hozac: since BMC uses the same physical eth as OS (and it's quite common for OS driver to break BMC in such case) 1320752573 M * daniel_hozac ah. 1320752587 M * daniel_hozac yeah, those are pretty fragile. 1320752665 M * arekm all my new machine have separate eth for ipmi/kvm but still have to use few old ones 1320752679 M * daniel_hozac which driver do you have? has it had that many commits? 1320752751 M * arekm tg3, not much commits, just production machine 400km away from me 1320752877 M * arekm heh, raising mem limit to 6GB, killing processes and then lowering back to 5GB is some way to recover 1320754912 M * Bertl off for now ... bbl 1320754918 N * Bertl Bertl_oO 1320758103 J * thierryp ~thierry@sop033r.vpn.inria.fr 1320760025 Q * BenG Quit: I Leave 1320761123 Q * FireEgl Ping timeout: 480 seconds 1320762457 J * fisted ~fisted@xdsl-87-78-218-249.netcologne.de 1320762983 Q * thierryp Remote host closed the connection 1320763690 J * K ~Kyle@37.88.89.79.rev.sfr.net 1320763732 N * K Key 1320763738 N * Key K 1320763948 P * K 1320764519 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1320766730 J * dowdle ~dowdle@scott.coe.montana.edu 1320767913 Q * hparker Ping timeout: 480 seconds 1320768379 Q * ncopa Quit: Leaving 1320768683 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1320771407 Q * renihs Quit: narf 1320771583 Q * derjohn_mob Ping timeout: 480 seconds 1320771866 J * bonbons ~bonbons@2001:960:7ab:0:dd28:473e:ac86:43b1 1320771970 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1320774211 Q * FireEgl Read error: Connection reset by peer 1320774817 Q * hparker Quit: Quit 1320775086 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1320775177 J * FireEgl ~FireEgl@173-16-9-169.client.mchsi.com 1320777495 Q * FireEgl Read error: Connection reset by peer 1320777828 Q * clopez Ping timeout: 480 seconds 1320778037 Q * Radiance Read error: Connection reset by peer 1320778037 J * _Radiance ~Radiance@193.16.154.187 1320778037 N * _Radiance Radiance 1320778700 Q * fisted Ping timeout: 480 seconds 1320779296 J * fisted ~fisted@xdsl-87-78-218-249.netcologne.de 1320780299 Q * kshannon Read error: Connection reset by peer 1320782870 Q * PowerKe Quit: leaving 1320782887 J * PowerKe ~tom@94-226-105-186.access.telenet.be 1320783601 N * Bertl_oO Bertl 1320783605 M * Bertl back now ... 1320784799 J * clopez ~clopez@238.10.117.91.dynamic.mundo-r.com 1320787069 Q * sannes1 Remote host closed the connection 1320788615 Q * bonbons Quit: Leaving 1320791558 J * dkg ~dkg@lair.fifthhorseman.net 1320791612 M * dkg i have a vserver guest with two IP addresses; since they're both considered "secondary" IP addresses on the host, they both show up as "secondary" in the output of "ip addr" on the guest; how do outbound connections decide which IP address to bind to? 1320791638 M * dkg more to the point: how can i force one of those addresses to be the default for binding for outbound connections? 1320791684 M * Bertl if the routing tables give a clear decision which one to prefer, it will be taken, if the tables aren't clear and/or result in an IP not assigned to the guest, the first one will be used 1320791817 M * dkg ok, so its the routing tables on the host + the order of the addresses as returned from "ip addr" ? 1320791836 M * dkg (i'm just trying to clarify what you mean by "the first one") 1320791846 M * Bertl not necessarily the order in 'ip' more the order they are assigned to the guest 1320791858 M * Bertl (although they often coincide) 1320791946 M * dkg so /etc/vserver/foo/interfaces/0/ip comes before /etc/vserver/foo/interfaces/1/ip ? 1320791969 M * Bertl you can check that with /proc/virtnet//info 1320791977 M * dkg ah, excellent. 1320791981 M * dkg thank you Bertl! 1320791985 M * Bertl you're welcome! 1320795729 Q * ghislain Quit: Leaving.