1316221952 Q * fisted Ping timeout: 480 seconds 1316222127 J * fisted ~fisted@xdsl-87-78-216-172.netcologne.de 1316224168 Q * hparker Ping timeout: 480 seconds 1316224785 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1316229353 M * Bertl off to bed now ... have a good one everyone! 1316229357 N * Bertl Bertl_zZ 1316230578 J * c0d3rc0d3 ~root@c114-77-239-157.ipswc3.qld.optusnet.com.au 1316232268 Q * hparker Ping timeout: 480 seconds 1316232347 J * hparker ~hparker@2001:470:1f0f:32c:beae:c5ff:fe01:b647 1316232494 Q * c0d3rc0d3 Ping timeout: 480 seconds 1316239332 J * ghislain ~AQUEOS@adsl2.aqueos.com 1316239698 Q * ghislain Quit: Leaving. 1316243466 J * derjohn_mob ~aj@p4FFD3EE3.dip.t-dialin.net 1316243602 J * thierryp ~thierry@home.parmentelat.net 1316244244 Q * thierryp Remote host closed the connection 1316245080 J * thierryp ~thierry@home.parmentelat.net 1316245358 J * Aiken_ ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1316245358 Q * Aiken Read error: Connection reset by peer 1316245520 N * Aiken_ Aiken 1316245546 Q * Aiken 1316245557 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1316246245 Q * thierryp Remote host closed the connection 1316247193 J * sannes1 ~ace@cm-84.209.106.118.getinternet.no 1316248403 J * bonbons ~bonbons@2001:960:7ab:0:8cef:67cc:4d71:2538 1316254760 J * thierryp ~thierry@lns-bzn-47f-62-147-212-202.adsl.proxad.net 1316256764 Q * thierryp Remote host closed the connection 1316260513 M * Rockj hmf *continune to play with his annoying virtual network cards* 1316264712 Q * derjohn_mob Ping timeout: 480 seconds 1316264732 J * derjohn_mob ~aj@p4FFD3EE3.dip.t-dialin.net 1316265864 Q * Aiken Remote host closed the connection 1316266622 Q * fisted Ping timeout: 480 seconds 1316266795 J * fisted ~fisted@xdsl-87-78-209-216.netcologne.de 1316267626 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316272749 N * Bertl_zZ Bertl 1316272759 M * Bertl morning folks! 1316274996 M * Guy- Bertl: hi; so far, the 'no such device' routing error seems to have been eliminated by the patch you suggested 1316275037 M * Bertl excellent! will prepare a proper patch which keeps the hiding 1316276593 M * geb hi folks 1316276633 M * geb can i remove the memory limits set by cgroups without rebooting a vserver ? 1316276650 M * geb i want to switch back to rlimits 1316276715 M * daniel_hozac because rlimits don't work? 1316276757 M * geb i use cgroups for limiting memory, it don't works very well ... i am not sure i can use both cgroup + rlimits 1316276796 M * daniel_hozac rlimits for memory have been removed since.. 2.6.33? 1316276811 M * geb :( 1316276915 M * geb i really have a problem with cgroups: i use 2.6.32 and a guest can use more memory than i set in cgroup, it just go to the swap (causing lot of I/O) 1316276935 M * geb then it is killed by the oom, flushing the swap and causing another I/O flood 1316277080 M * geb my monitoring is kinda broken, but you can see the problem at http://supervision.toile-libre.org/munin/toile-libre.net/vservers1.toile-libre.net.html ( cpu spending lot of time on I/O) 1316277162 M * geb do you have an idea for that ? 1316277163 M * Bertl so, what behaviour do you invision instead? 1316277199 M * geb the same as rlimits: a vserver can't use more memory than the limit 1316277200 M * Bertl *envision 1316277224 M * Bertl just set a mem+swap limit as well, make it the same as the mem limit 1316277239 M * Bertl guest processes will get OOM kills once over the limit 1316277360 M * geb hum, interesting, i don't thought about that 1316277419 M * geb i remember that we already talked about it and you told me that it is really a problem you have with cgroups 1316277475 M * Bertl no, my problem with cgroups is that it swaps once over the memory limit instead of keeping that in memory till the host needs to swap 1316277501 M * Bertl but that was a design decision/feature of the cgroup system and nobody (so far) complained about it :) 1316277527 M * Bertl (at least not loud enough for the cgroup developers to hear) 1316277703 M * geb :) 1316277738 M * geb so the problematic question: if i use that (memory.memsw.limit = memory.limit), how will the host react when memory will be full ? 1316277785 M * Bertl what do you mean with 'host'? 1316277865 M * Bertl the host as in the host processes will not do much, the kernel will go and OOM kill inside the guest cgroup 1316277931 M * Bertl or are we talking host memory now? 1316277953 M * geb sorry, for not behing clear, kinda difficult to explain, sec :) 1316278064 M * geb if all memory is used by vservers, will the host swap, or will it just kill process inside guest cgroups ? 1316278116 M * Bertl as long as the guests are not over limit, no killing will happen I guess 1316278179 M * geb for example, if i have 4 guest, 1G of memory on the host, and put a 512M memory limit. What will append if the guests use all their there limits ? will the host just swap or ..? 1316278223 M * Bertl assuming the guests use slightly less than their assigned limit 1316278242 M * Bertl (or would use that amount if it were available) 1316278264 M * Bertl then no OOM killing should happen, as long as the swap space (on the host) suffices 1316278293 M * geb ok nice ! 1316278603 M * geb sorry for not behing clear but it was really not clear for me that the host will use swap if it needs memory for guests and the guest have memory.memsw.limit = memory.limit 1316278705 M * Bertl the memory limit will not be reached for any guest 1316278724 M * Bertl and the memsw limit doesn't care where the memory is, i.e. resident or swapped 1316278776 M * geb ok tks a lot then :) 1316278784 M * Bertl np 1316278997 Q * sweil Remote host closed the connection 1316279026 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316279074 Q * sweil Read error: Connection reset by peer 1316279078 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316279111 Q * sweil Remote host closed the connection 1316279124 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316279142 Q * sweil Read error: Connection reset by peer 1316282005 J * clopez ~clopez@238.10.117.91.dynamic.mundo-r.com 1316283138 M * Bertl Guy-: http://vserver.13thfloor.at/ExperimentalT/delta-dev-fix01.diff 1316283162 M * Bertl that one should do the trick, while keeping the devices hidden from guest processes 1316283171 M * Bertl (doesn't hide the routes yet though) 1316283342 Q * fback_ Quit: moving stuff 1316283360 M * Guy- Bertl: thanks; I think I can test it tomorrow 1316283452 M * Bertl great! let us know how it goes! 1316283467 M * Bertl ah, you have to revert your test changes 1316283961 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316284397 M * Guy- that's clear :) 1316284411 Q * sweil Remote host closed the connection 1316284461 J * sweil ~stefan@p5086E8FB.dip.t-dialin.net 1316284515 Q * sweil Read error: Connection reset by peer 1316286804 J * fback fback@red.fback.net 1316297720 Q * sannes1 Remote host closed the connection 1316298921 Q * bonbons Quit: Leaving 1316300397 J * Aiken ~Aiken@2001:44b8:2168:1000:21f:d0ff:fed6:d63f 1316303795 Q * clopez Ping timeout: 480 seconds