| 00:19.13 | *** part/#asterisk-dev kharwell (kharwell@nat/digium/x-gkwbsuopafbyewce) |
| 00:29.03 | jpsharp | coreyfarrell and file, thanks for the help yesterday. I was finally able to find my refcount leak. |
| 01:40.37 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 01:49.36 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 04:33.53 | *** join/#asterisk-dev CELYA_ (~Thunderbi@pub9-47.mobius.fr) |
| 05:53.31 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 06:04.20 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 06:36.00 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 07:14.56 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 07:17.17 | *** join/#asterisk-dev stevedav_ (~stevedavi@197.155.252.3) |
| 07:26.42 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 07:39.44 | *** join/#asterisk-dev pchero_work (~pchero@87.213.240.121) |
| 07:40.10 | *** join/#asterisk-dev jkroon (~jkroon@41.114.74.43) |
| 08:01.17 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 08:03.02 | *** join/#asterisk-dev jkroon (~jkroon@41.114.74.43) |
| 08:04.58 | *** join/#asterisk-dev jkroon (~jkroon@41.114.74.43) |
| 08:53.21 | *** join/#asterisk-dev pchero_work (~pchero@87.213.240.121) |
| 09:48.58 | *** join/#asterisk-dev hehol (~hehol@gatekeeper.loca.net) |
| 10:11.29 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 10:45.39 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 10:53.10 | *** join/#asterisk-dev pchero_work (~pchero@51.247.195.35.bc.googleusercontent.com) |
| 10:56.34 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 10:57.17 | *** join/#asterisk-dev CELYA_ (~Thunderbi@pub9-47.mobius.fr) |
| 11:38.30 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 11:48.45 | *** join/#asterisk-dev pchero_work (~pchero@87.213.240.121) |
| 11:59.46 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 12:32.46 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 12:51.59 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 13:27.49 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 13:37.30 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 14:15.31 | *** join/#asterisk-dev cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 14:15.31 | *** mode/#asterisk-dev [+o cresl1n] by ChanServ |
| 14:39.49 | cresl1n | waves |
| 14:39.57 | seanbright | file wobbles |
| 14:53.26 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 14:53.36 | tsearle | waves |
| 15:17.24 | *** join/#asterisk-dev kharwell (kharwell@nat/digium/x-kzotzhwdvvrinqxl) |
| 15:17.24 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:37.37 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 15:45.55 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-yexonpztqwzhmlpj) |
| 15:46.05 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 16:09.52 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 16:19.27 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 18:20.17 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 18:31.40 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 18:40.39 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 18:56.59 | *** join/#asterisk-dev elcontrastador (~textual@70-90-215-98-BusName-ca.sacra.hfc.comcastbusiness.net) |
| 19:08.45 | *** join/#asterisk-dev DivideBy0 (~DivideBy0@unaffiliated/divideby0x0) |
| 19:08.45 | *** mode/#asterisk-dev [+o DivideBy0] by ChanServ |
| 19:14.09 | *** join/#asterisk-dev DivideBy0 (~DivideBy0@unaffiliated/divideby0x0) |
| 19:14.09 | *** mode/#asterisk-dev [+o DivideBy0] by ChanServ |
| 20:17.11 | seanbright | bford: so... strtok |
| 20:18.02 | seanbright | are there thread-safety concerns there? |
| 20:18.56 | seanbright | i guess i could put this question on gerrit |
| 20:22.45 | kharwell | typically you'd want to use strtok_r to be safe |
| 20:23.35 | seanbright | right. just thought i may have been missing something. |
| 20:24.42 | seanbright | there are a few places in asterisk where a general purpose delimited string walker would be useful |
| 20:25.40 | seanbright | returned pj_str_t-like structures so you didn't need to dup the source string, etc. |
| 20:25.56 | seanbright | can someone get on that? |
| 20:25.58 | seanbright | kthx |
| 20:39.41 | coreyfarrell | should we have a define for `strtok` like we do for some other system functions to block it from being used (throw a compile error)? |
| 20:41.15 | seanbright | wow... it's actually used |
| 20:41.39 | seanbright | pbx_dundi.c, chan_sip.c, and res_fax.c |
| 20:41.45 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 20:52.36 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 21:00.15 | bford | seanbright: i'll take a look |
| 21:02.55 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 21:04.54 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 21:14.13 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 21:15.14 | seanbright | coreyfarrell: i added that to my review |
| 21:15.30 | bford | seanbright: strtok_r is the right choice, fixed |
| 21:27.18 | seanbright | bford: NICE |
| 21:27.23 | seanbright | SHIP IT |
| 21:59.11 | seanbright | tm1000: i have to ask... why in god's name are 29 calls causing CPU to get anywhere close to 60% CPU utilization? |
| 21:59.58 | tm1000 | lots of system AGIs |
| 22:00.36 | seanbright | was your customer talking about overall CPU or asterisk process CPU? |
| 22:04.08 | jpsharp | We shaved about half of our CPU usage by switching to FastAGI |
| 22:04.58 | seanbright | well, yeah, but still |
| 22:05.06 | seanbright | 60% seems really high to me |
| 22:05.27 | seanbright | i have no clue what the AGIs are doing, so perhaps i am speaking out of school |
| 22:09.50 | *** join/#asterisk-dev sh_smith (~sh_smith@cpe-76-174-26-91.socal.res.rr.com) |
| 22:47.09 | *** join/#asterisk-dev pchero (~pchero@dhcp-077-249-058-090.chello.nl) |
| 22:49.20 | cresl1n | maybe lots of forking? |
| 22:49.33 | cresl1n | but it seems weird that there'd be so much CPU |
| 22:49.46 | cresl1n | if it's lots of tiny AGIs happening all over the place... |
| 22:49.58 | cresl1n | combined with lots of forking |
| 22:50.00 | cresl1n | maybe? |
| 22:50.03 | cresl1n | idk though |
| 22:50.27 | jpsharp | That's what killed our stuff. forking like there was no tomorrow. Killed the CPU and made calls very choppy. |
| 22:51.49 | seanbright | well with fastagi there is no forking |
| 22:51.54 | cresl1n | yeah |
| 22:52.10 | seanbright | i guess the logic in the agis eats a lot of cpu |
| 22:52.16 | cresl1n | maybe |
| 22:52.22 | seanbright | i'm happy to fix if you want to pay me |
| 22:52.26 | seanbright | :P |
| 22:52.50 | seanbright | for the first hour, i only want $0.01 |
| 22:52.59 | seanbright | and then double it for every subsequent hour... so cheap |
| 22:53.03 | cresl1n | last time we had someone with AGI issues, we told them they needed to change to FastAGI due to forking complications with AGI |
| 22:53.36 | seanbright | i think we can make standard AGIs faster by forking a shim process and then forking _that_ process for each agi request |
| 22:53.44 | seanbright | it's on my long list of fun-future-projects |
| 22:54.13 | seanbright | i think the expense with fork in the freepbx case was dup'ing all of the timerfd fds |
| 22:55.01 | seanbright | if that only happened once when forking the shim process (which then closed them all), it would be better |
| 23:15.07 | *** join/#asterisk-dev stevedavies (~stevedavi@197.155.252.3) |
| 23:50.20 | *** join/#asterisk-dev pchero_work (~pchero@dhcp-077-249-058-090.chello.nl) |
| 23:51.25 | tm1000 | what I blogged about is basically exactly what jpsharp said. same issues |