00:17.44 | josePHPagoda | yeah |
00:18.01 | josePHPagoda | i thought you could do vm.oom-kill=0 in sysctl, but i guess that's more up to date distros than what my friend is using |
00:18.03 | josePHPagoda | :-/ |
00:27.48 | fugue88 | josePHPagoda: If your machine runs out of physical memory, and an app asks for more, *something* has to die. |
00:27.55 | josePHPagoda | yeah |
00:27.58 | fugue88 | oom-killer tries to pick an app with some intelligence. |
00:28.01 | josePHPagoda | i think this was an overcommit issue |
00:28.08 | fugue88 | Otherwise, whatever asked for the allocation dies. |
00:28.13 | fugue88 | Which can appear completely random. |
00:28.21 | josePHPagoda | it only happens when running backups |
00:28.21 | fugue88 | Not necessarily the memory-gobbling app. |
00:28.28 | josePHPagoda | so we've made logging more verbose |
00:28.37 | josePHPagoda | the weird thing is that swap never gets used |
00:28.45 | josePHPagoda | you would expect swap to be used to some extent |
00:28.49 | fugue88 | Why do you suspect oom-killer? |
00:29.40 | fugue88 | josePHPagoda: ^ |
00:29.47 | josePHPagoda | because the logs say that oom-killer killed it |
00:29.49 | josePHPagoda | :P |
00:30.15 | fugue88 | Well, that's a fairly good indicator... |
00:30.19 | fugue88 | :) |
00:30.39 | fugue88 | josePHPagoda: 32 or 64 bit arch? |
00:31.20 | fugue88 | wonders if you get ever get an oom-killer log entry from exhausting a process's virtual memory space |
00:33.09 | josePHPagoda | 64bit |
00:33.09 | fugue88 | josePHPagoda: Also, does the output of "swapon -s" appear sane? Or the output of "free"? |
00:33.16 | josePHPagoda | they do |
00:33.20 | josePHPagoda | 16GB of swap free |
00:33.22 | fugue88 | josePHPagoda: Sure, just destroy all of my ideas, okay! |
00:33.24 | josePHPagoda | using like 2 M |
00:33.25 | josePHPagoda | :P |
00:33.35 | fugue88 | josePHPagoda: What distro? |
00:33.40 | josePHPagoda | proxmox |
00:33.43 | josePHPagoda | (it's build on debian) |
00:33.45 | fugue88 | Well there you go! |
00:33.49 | fugue88 | j/k ;P |
00:33.54 | josePHPagoda | but the guy who set it up didn't follow best practice |
00:34.00 | josePHPagoda | so i've found all sorts of a broken configs |
00:34.11 | josePHPagoda | like he didn't leave any room for lvm snapshots |
00:34.13 | josePHPagoda | which was annoying |
00:34.27 | fugue88 | k sry cant halp |
00:34.44 | josePHPagoda | it's aight |
00:35.00 | josePHPagoda | fugue88: so you making it to the conf this year? |
00:35.50 | fugue88 | Uh, heheh, oops. |
00:36.01 | fugue88 | pleads the "new baby" defense |
00:36.22 | fugue88 | checks the conf website |
00:36.41 | fugue88 | Oh yeah. Orem. |
00:37.30 | fugue88 | josePHPagoda: May 2-4, yeah I might could will be there. |
00:38.45 | fugue88 | can't imagine how an Orem location with be popular with anybody (by Condorcet ranking) except those in that valley. |
00:40.15 | fugue88 | s/with/would/ |
00:46.24 | fugue88 | Wow. Javascript runs through the entire schedule. |
01:11.55 | *** join/#utos eighty9vision (~89vision@2601:7:7180:11:222:fbff:fe12:7f9a) |
02:39.20 | *** join/#utos redbeard2 (~jphall@techhat.dsl.xmission.com) |
04:01.41 | *** join/#utos jalbretsen (~jalbretse@c-24-10-128-246.hsd1.ut.comcast.net) |
09:23.00 | *** join/#utos infobot (~infobot@rikers.org) |
09:23.00 | *** topic/#utos is OpenWest Conference Community voting now open! Go vote for your favorite presentations at http://openwest.org/vote |
14:30.42 | *** join/#utos redbeard2 (~jphall@63.230.12.160) |
15:24.07 | *** join/#utos jalbretsen (~jalbretse@sonic.itransact.com) |
16:22.19 | *** join/#utos sinuhe (~sinuhe@38.73.135.226) |
18:07.35 | *** join/#utos bgerber (~gerberb@168.103.176.50) |
20:21.15 | *** join/#utos cro (~Adium@204-228-149-217.ip.xmission.com) |
21:14.57 | *** join/#utos cro (~Adium@204-228-149-217.ip.xmission.com) |
21:33.40 | *** join/#utos sinuhe (~sinuhe@38.73.135.226) |
23:23.28 | *** join/#utos mindjuju (~mindjuju@173-10-26-193-BusName-utah.ut.hfc.comcastbusiness.net) |
23:23.28 | *** mode/#utos [+o mindjuju] by ChanServ |