1291337530 Q * bsingh Ping timeout: 480 seconds 1291338068 J * bsingh ~balbir@122.172.4.106 1291343975 Q * bsingh Read error: Connection reset by peer 1291344724 J * bsingh ~balbir@122.172.17.118 1291346126 Q * puck Ping timeout: 480 seconds 1291346370 J * puck ~puck@leibniz.catalyst.net.nz 1291346952 Q * manana Remote host closed the connection 1291350068 Q * DoberMann Ping timeout: 480 seconds 1291350301 J * DoberMann ~james@2a01:e35:8b44:84c0:230:18ff:fea3:7188 1291356427 Q * bsingh Ping timeout: 480 seconds 1291357123 J * bsingh ~balbir@122.167.248.67 1291358106 Q * Snow-Man_ Ping timeout: 480 seconds 1291358197 Q * bsingh Ping timeout: 480 seconds 1291358690 M * Bertl off to bed now ... have a good one everyone! 1291358694 N * Bertl Bertl_zZ 1291359679 J * Snow-Man ~sfrost@tamriel.snowman.net 1291359836 J * ncopa ~ncopa@3.203.202.84.customer.cdi.no 1291360938 Q * derjohn_mob Ping timeout: 480 seconds 1291361472 J * bsingh ~balbir@122.172.40.172 1291361713 J * ghislain ~AQUEOS@adsl2.aqueos.com 1291361852 J * petzsch ~markus@dslb-088-075-170-221.pools.arcor-ip.net 1291362460 J * derjohn_mob ~aj@213.238.45.2 1291363447 Q * bsingh Read error: Connection reset by peer 1291363463 J * harobed ~harobed@arl57-1-82-231-110-14.fbx.proxad.net 1291364897 M * pmjdebruijn https://lkml.org/lkml/2010/12/2/388 1291365953 J * thierryp ~thierry@zankai.inria.fr 1291366389 J * BenG ~bengreen@cpc6-aztw22-2-0-cust100.aztw.cable.virginmedia.com 1291366791 J * bsingh ~balbir@122.172.164.178 1291368355 J * barismetin ~barismeti@zanzibar.inria.fr 1291368761 J * kir ~kir@swsoft-msk-nat.sw.ru 1291369327 Q * FireEgl Remote host closed the connection 1291369645 Q * BenG Quit: I Leave 1291369816 Q * ntrs_ Ping timeout: 480 seconds 1291370236 J * FireEgl FireEgl@2001:470:e056:9:b0f7:dd3:93a9:a9cf 1291370282 J * ntrs ~ntrs@vault08.rosehosting.com 1291372507 Q * bsingh Ping timeout: 480 seconds 1291373861 Q * ktwilight_ Read error: Connection reset by peer 1291373879 J * ktwilight ~keliew@91.176.59.107 1291373925 Q * thierryp Read error: Connection reset by peer 1291373947 J * thierryp ~thierry@zankai.inria.fr 1291374530 Q * petzsch Read error: Connection reset by peer 1291374710 Q * thierryp Read error: Connection reset by peer 1291374720 J * thierryp ~thierry@zankai.inria.fr 1291375968 Q * niki Quit: Ex-Chat 1291377044 J * niki ~niki@94.145.207.11 1291377464 Q * barismetin Remote host closed the connection 1291378420 J * nkukard_ ~nkukard@dsl-247-39-113.telkomadsl.co.za 1291378740 Q * _WildPIkachu_ Ping timeout: 480 seconds 1291378980 Q * nkukard_ Ping timeout: 480 seconds 1291379157 J * nkukard_ ~nkukard@196-215-87-245.dynamic.isadsl.co.za 1291381180 P * kir Leaving. 1291381268 N * Bertl_zZ Bertl 1291381272 M * Bertl morning folks! 1291381670 Q * thierryp Remote host closed the connection 1291381692 J * thierryp ~thierry@zankai.inria.fr 1291381896 J * bsingh ~balbir@122.167.168.74 1291382543 Q * derjohn_mob Ping timeout: 480 seconds 1291385644 J * barismetin ~barismeti@zanzibar.inria.fr 1291386193 J * derjohn_mob aj@88.128.148.60 1291386676 Q * barismetin Remote host closed the connection 1291386816 J * barismetin ~barismeti@zanzibar.inria.fr 1291387559 M * Bertl off for now ... bbl 1291387563 N * Bertl Bertl_oO 1291387627 J * Walex ~Walex@o1.phyip3.dur.ac.uk 1291388287 Q * harobed Ping timeout: 480 seconds 1291389681 J * hparker ~hparker@2001:470:1f0f:32c:215:f2ff:fee0:9872 1291389697 Q * hparker Remote host closed the connection 1291390294 J * kazuya kazuya@freebsdsecured.org 1291390536 Q * barismetin Remote host closed the connection 1291391146 Q * thierryp Ping timeout: 480 seconds 1291391368 J * barismetin ~barismeti@zanzibar.inria.fr 1291393204 Q * barismetin Remote host closed the connection 1291393656 Q * Walex Remote host closed the connection 1291394690 J * dna ~dna@dslb-094-222-116-198.pools.arcor-ip.net 1291395075 Q * derjohn_mob Ping timeout: 480 seconds 1291395697 J * derjohn_mob aj@88.128.17.176 1291396985 Q * FireEgl Remote host closed the connection 1291397235 J * petzsch ~markus@p4FF43E2B.dip.t-dialin.net 1291398796 J * FireEgl FireEgl@2001:470:e056:8:6d61:4412:ef85:a624 1291400593 Q * derjohn_mob Ping timeout: 480 seconds 1291401312 Q * FireEgl Ping timeout: 480 seconds 1291404155 J * hparker ~hparker@2001:470:1f0f:32c:215:f2ff:fee0:9872 1291405439 J * mkiez mike@user.custhost.se 1291405766 M * mkiez Bertl_oO: there? 1291405833 N * Bertl_oO Bertl 1291405838 M * Bertl mkiez: what's up? 1291405878 M * mkiez i am not sure how to explain but i will try 1291406007 M * mkiez i had a bit of fun with cgroups but i did not notice anything wrong with it till now 1291406042 Q * petzsch Quit: Leaving. 1291406043 M * Bertl cgroups as in vanilla kernel or cgroups with Linux-VServer? 1291406047 M * mkiez ks26083 mysqld invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0 ks26083 mysqld cpuset=xtc mems_allowed=0 1291406068 M * mkiez and it is just dead 1291406087 M * Bertl sounds like 'expected' after OOM killing 1291406119 M * mkiez mhm 1291406123 M * mkiez but why? 1291406153 M * mkiez Bertl: xtc is vserver guest 1291406156 M * Bertl most likely because you reached 'a' memory limit 1291406186 M * mkiez on a guest? 1291406200 M * Bertl not necessarily, but possibly 1291406235 M * mkiez how i can fix that? 1291406263 M * Bertl depends on what caused the OOM kill, if you ran out of memory host wide, you need to add more memory to 'fix' it 1291406296 M * mkiez Bertl: but everything worked fine till now 1291406297 M * mkiez :/ 1291406310 M * Bertl if you indeed did hit a guest limit, you want to raise that limit or reduce the amount of memory used by the guest apps 1291406463 M * Bertl what does /proc/meminfo show on the host? 1291406466 M * Bertl (please use paste.linux-vserver.org for everything longer than 3 lines) 1291406555 M * mkiez http://paste.linux-vserver.org/18474 1291406652 M * Bertl well, besides the unfortunate highmem (you should avoid that on systems with less than 3GB) it looks like you still have some resources left 1291406674 M * Bertl so I'd assume that you did hit a guest/cgroup limit for that guest 1291406709 M * mkiez well in my history i only see cpuset.cpus 1291406710 M * Bertl (or for whatever reason, mysql allocated more than 1GB at once) 1291406720 M * mkiez nothing with mem at all 1291406743 M * Bertl was the host rebooted since? 1291406774 J * dna_ ~dna@dslb-088-074-211-076.pools.arcor-ip.net 1291406783 M * mkiez yes 1291406816 M * Bertl well, then we won't get much information from that (but I suspected so, as the swap space is prestine) 1291406829 M * mkiez well i got some log 1291406842 M * mkiez http://mike.custhost.se/bld.txt 1291406878 M * Bertl tftornado.py? 1291406907 M * mkiez its happend first time with tftornado.py 1291406913 M * mkiez second time with mysql 1291406934 M * mkiez tftornad its kind of torrent client script 1291406974 M * Bertl well, seems you actually used up all the memory 1291406989 M * Bertl Dec 3 09:45:33 ks26083 Free swap = 0kB 1291407014 M * Bertl i.e. you apps (all together) used more than 2.5GB 1291407050 M * mkiez are you sure its because of that? 1291407067 M * Bertl lighttpd alone used half a gigabyte of ram 1291407095 M * mkiez 33% of 2gb now 1291407161 M * Bertl apache2 can get quite hungry as well, and php/cgi ... well, you probably know :) 1291407191 M * Bertl for torrent it mostly depends on the chunks and how much the client caches 1291407200 M * mkiez but why it did not happend before? 1291407202 Q * dna Ping timeout: 480 seconds 1291407209 M * mkiez *happen 1291407226 M * Bertl all those processes have flexible memory requirements 1291407250 J * cuba33ci_ ~cuba33ci@111-240-214-134.dynamic.hinet.net 1291407255 M * Bertl i.e. httpd without requests uses like 10M memory, with request, caches grow up to 2GB 1291407260 M * mkiez how to avoid that apart of adding more ram ;p? 1291407303 M * Bertl you can limit the applications (where possible) to a number where you are sure that the sum over all apps will fit into memory + swap 1291407324 M * Bertl you can also increase the swap space to handle sudden bursts or memory leaks 1291407353 M * mkiez its quiet fast filling up 1291407354 M * mkiez total used free shared buffers cached 1291407355 M * mkiez Mem: 2052920 1497400 555520 0 4576 706004 1291407400 M * Bertl the buffers/cached space is instantly available when memory is required 1291407455 M * Bertl so those are not a problem, but RSS allocations (ps auxwww) are 1291407554 M * mkiez ps auxwww shows me that i using less than 35% 1291407561 M * mkiez mem 1291407595 M * Bertl you need to check with vps actually to see guest processes 1291407602 Q * cuba33ci Ping timeout: 480 seconds 1291407607 M * Bertl i.e. vps auxwww 1291407610 N * cuba33ci_ cuba33ci 1291407654 M * Bertl but yes, I do not expect it to use up all the memory right now, but over time, especially if there are http requests and php scripts running, you will use up all the memory 1291407666 M * Bertl maybe there is even a memory leak in one of your applications 1291407748 M * Bertl monitoring seems the best approach here, i.e. get some monitoring/graphing solution and record memory usage for certain apps or guests, once you identify an unusual increas, you know who's to blame 1291407762 M * Bertl anyway, I need a nap right now .. bbl 1291407769 N * Bertl Bertl_zZ 1291407770 M * mkiez su -c "echo 3 > /proc/sys/vm/drop_caches" 1291407782 M * mkiez is this gonna help me for now? 1291407793 M * mkiez after adding it to cron ;p? 1291408197 Q * hparker Remote host closed the connection 1291408377 J * bonbons ~bonbons@2001:960:7ab:0:2c0:9fff:fe2d:39d 1291408442 J * _WildPIkachu_ ~nkukard@dsl-146-40-05.telkomadsl.co.za 1291408868 Q * nkukard_ Ping timeout: 480 seconds 1291410919 J * BenG ~bengreen@cpc6-aztw22-2-0-cust100.aztw.cable.virginmedia.com 1291411037 Q * BenG 1291412695 Q * dna_ Quit: Verlassend 1291413515 Q * imcsk8 Quit: Leaving 1291415757 J * daniel_hozac ~daniel@c-aa3771d5.08-230-73746f22.cust.bredbandsbolaget.se 1291415777 Q * bonbons Quit: Leaving 1291415825 Q * daniel_hozac 1291415843 J * daniel_hozac ~daniel@c-aa3771d5.08-230-73746f22.cust.bredbandsbolaget.se 1291416463 N * Bertl_zZ Bertl 1291416467 M * Bertl back now ... 1291416498 M * Bertl mkiez: nope, dropping caches will not help you a bit, just decrease overall performance 1291416524 M * mkiez Bertl: it did not happend before 1291416525 M * Bertl (no idea why folks believe that dropping caches would help :) 1291416570 M * mkiez i unset cgroup and some grsec options now im putting it back 1291416577 M * mkiez it might be because of this? 1291416585 M * Bertl mkiez: I guess you have some problem understanding memory, allocation and memory consumption 1291416599 J * thierryp ~thierry@home.parmentelat.net 1291416600 M * mkiez no it just not happend before 1291416619 M * Bertl okay, let me make an easily understandable example for you: 1291416621 M * mkiez Bertl: 4 times in 3houers? 1291416671 M * Bertl did you record the memory usage during that period? 1291416741 M * Bertl your total ram is 2GB, unfortunately broken up into low and high mem (something you should change when you build the next kernel) 1291416760 M * Bertl your swap space is 0.5GB 1291416766 M * mkiez yup 1291416827 M * mkiez Bertl: i just do not know how it happen now but never happend before ;p 1291416900 M * Bertl if you sum up the VSZ of all processes on the machine, like this 1291416920 M * Bertl vps auxwww | gawk '{sum+=$7; mem+=$8} END { printf "%s %s\n",sum,mem }' 1291416944 M * Bertl (the first one is the sum over the virtual memory allocations, the second one the real ones) 1291416973 M * mkiez mhm 1291417020 M * Bertl now the first one is what applications 'allocate' with the kernel 1291417032 M * Bertl which is a lot more than you actually have available 1291417045 M * Bertl this is called overcommitting memory 1291417163 M * Bertl now the second sum is the actually used memory/swap 1291417192 M * Bertl i.e. those parts of the allocations are actually filled with data 1291417300 M * mkiez hmm 1291417340 M * Bertl of course, you are probably using an older kernel with Linux-VServer and grsec, maybe even some other patches 1291417355 M * mkiez Bertl: 2.6.36 1291417376 M * Bertl okay, that's at least a good one IMHO 1291417438 M * Bertl so, my suggestion would be: 1291417438 M * mkiez i change it now to 36.1 1291417445 M * mkiez yeah? 1291417467 M * Bertl create some additional swap space, if possible on a separate device 1291417482 M * Bertl but if you cannot do that, use a swap file on one of your filesystems 1291417508 Q * ghislain Quit: Leaving. 1291417512 M * Bertl let's say 2GB mem and 2GB swap (total) 1291417528 M * mkiez how i can do that ? 1291417530 M * Bertl then monitor memory usage and especially swap 1291417535 M * mkiez do i have to reinstall os? 1291417550 M * Bertl nah, just create a file with 1.5GB on one of your partitions 1291417581 M * mkiez ok 1291417586 M * Bertl initialize it with mkswap and add it with swapon 1291417589 M * mkiez how i can monitor usgae of memory? 1291417664 M * Bertl simplest way would be 'cat /proc/meminfo' every 10 seconds or so :) 1291417690 M * Bertl a better approach would be some monitoring setup ala munin or rrdcollect with a round robin database as backend 1291417703 M * Bertl (and maybe a graphing solution as frontend) 1291417710 M * mkiez kk 1291417726 M * mkiez i will let you know in progress ;} 1291417732 M * mkiez thanks for nice chat 1291417749 M * Bertl you're welcome! 1291418557 J * FireEgl FireEgl@2001:470:e056:9:21bc:5a46:f1b4:dc9b