| 00:19.02 | *** join/#asterisk-dev cmendes0101 (~cmendes01@wtnl.corp.tierra.net) |
| 00:32.26 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 00:32.26 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 02:10.05 | *** join/#asterisk-dev greenwolf (42570079@gateway/web/freenode/ip.66.87.0.121) |
| 02:43.06 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 02:43.07 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 03:45.55 | *** join/#asterisk-dev greenwolf (425702a9@gateway/web/freenode/ip.66.87.2.169) |
| 04:44.23 | *** join/#asterisk-dev drmessano (~nonya@pdpc/supporter/active/drmessano) |
| 04:45.08 | *** join/#asterisk-dev dpilon (~dpilon@c-50-138-178-238.hsd1.ct.comcast.net) |
| 04:59.06 | *** join/#asterisk-dev greenwolf (425702a9@gateway/web/freenode/ip.66.87.2.169) |
| 06:48.10 | *** join/#asterisk-dev jmls (~somefake@77.107.171.82) |
| 06:56.30 | snuff-work | nice work wedhorn.. |
| 06:57.22 | wedhorn | thanks, getting there ... |
| 07:02.07 | snuff-work | mm with that immed dial in the global settings.. only 1 line to add.. since we dont support showing global conf via manager.. |
| 07:03.10 | wedhorn | you could always add global conf via manager if you want more than 1 line ;) |
| 07:04.31 | snuff-work | if chan_sip doesn't.. why should we ;) |
| 07:05.23 | wedhorn | why follow sip? Shouldn't the question be whether you want it or not? |
| 07:05.55 | snuff-work | i think part of this sort of stuff will come out with the new data api |
| 07:06.54 | *** join/#asterisk-dev drmessano (~nonya@pdpc/supporter/active/drmessano) |
| 07:08.23 | snuff-work | rather stick to status quo.. also slight issue with ur sample conf for skinny and the immed dial i noticed |
| 07:09.08 | snuff-work | you put it in as immeddialkey="#" .. i just check for the '#' sign.. not '"#"' |
| 07:10.06 | wedhorn | isn't there an asterisk string function that strips the quotes |
| 07:10.27 | snuff-work | probably.. i dont recall top of my head |
| 07:11.13 | snuff-work | <PROTECTED> |
| 07:11.20 | wedhorn | depends what ast_jb_read_conf does with the string |
| 07:14.06 | snuff-work | i dont think it touches it at all.. since all other conf options dont use strip quotes etc |
| 07:15.39 | wedhorn | probably easier to just test it rather than trying to trace the code |
| 07:15.51 | snuff-work | well i know what works for me.. |
| 07:17.18 | snuff-work | as is.. with just immeddialkey=# .. unless i also put in a call to the ast_strip_quoted() to remove the extra "'s |
| 07:17.27 | wedhorn | yeah, happy for you to change the conf file (it's probably better to be similar to others, even if the quotes didn't impact) |
| 07:17.48 | wedhorn | okay, need to fix the conf then |
| 07:18.18 | snuff-work | 2 lines in 2 file patch.. |
| 07:18.23 | snuff-work | i gotta get going to my train |
| 07:52.33 | *** join/#asterisk-dev tzafrir_laptop (~tzafrir@59.157-240-81.adsl-dyn.isp.belgacom.be) |
| 07:52.33 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 09:05.52 | *** join/#asterisk-dev bulkorok (~bulkorok@85.183.36.36) |
| 09:11.23 | *** join/#asterisk-dev askin (~askn@85.105.28.77) |
| 09:56.11 | *** join/#asterisk-dev askin (~askn@85.105.28.77) |
| 10:06.50 | *** join/#asterisk-dev sgimeno (~sgimeno@163.117.206.10) |
| 10:13.11 | *** join/#asterisk-dev jmls (~somefake@77.107.171.82) |
| 10:13.11 | *** join/#asterisk-dev newtonr (~newtonr@64.34.219.47) |
| 10:13.11 | *** join/#asterisk-dev pnlarsson (~niklas@fw1.gml.g.icnet.infracom.se) |
| 10:13.11 | *** join/#asterisk-dev aimka (~aimka@unaffiliated/aimka) |
| 10:13.11 | *** mode/#asterisk-dev [+o newtonr] by wright.freenode.net |
| 10:16.09 | *** join/#asterisk-dev newtonr (~newtonr@64.34.219.47) |
| 10:16.09 | *** mode/#asterisk-dev [+o newtonr] by ChanServ |
| 12:54.08 | leifmadsen | wdoekes: ping? |
| 12:54.11 | *** join/#asterisk-dev anonymouz666 (~anonymouz@189-25-136-163.user.veloxzone.com.br) |
| 12:56.26 | wdoekes | pong |
| 12:57.01 | leifmadsen | wdoekes: hey! Jim was using my account yesterday so was the one dealing with the maxfiles thing, but I'm looking at it now too, and I'm not sure what we should be recommending here.... |
| 12:57.16 | leifmadsen | I understand we should raise it, but to what value? How do we determine what that value should be? |
| 12:58.02 | wdoekes | calls*5 could be a heuristic |
| 12:58.24 | leifmadsen | looking at output of "core show settings" on a 1.8 box, the 'maximum open files handles' has the value "not set" by default, so it's not 1000 by default it seems |
| 12:58.31 | wdoekes | if you assume 2 udp ports per leg (rtp/rtcp) |
| 12:58.54 | leifmadsen | so basically channels * 2 ? |
| 12:59.02 | wdoekes | + overhead |
| 12:59.04 | leifmadsen | right |
| 12:59.07 | leifmadsen | hence the 5 |
| 12:59.23 | wdoekes | sort of.. I ran into the limit once |
| 12:59.47 | wdoekes | and it was at somewhere between 100-300 calls |
| 12:59.57 | leifmadsen | I'm looking at the output of the location you showed... I have this system set at 63536 max open files it seems |
| 13:00.17 | wdoekes | in that case you'd be fine.. let me check |
| 13:00.33 | leifmadsen | does asterisk then set that value I guess with the maxfiles setting then? |
| 13:00.51 | leifmadsen | seems like a dangerous thing to have asterisk set if you had other services on the same box :) |
| 13:02.40 | leifmadsen | ok I think I do have enough information now to add something of use |
| 13:02.54 | leifmadsen | I'll also talk about setting it system wide outside of asterisk using sysctl |
| 13:03.17 | leifmadsen | wish we had the time to have written a system tuning section with a bunch of advanced knowledge |
| 13:03.32 | leifmadsen | btw thanks for your feedback, it has been most useful |
| 13:04.33 | wdoekes | core show settings is evil. it shows what's in the .conf, not what it's actually using |
| 13:04.52 | wdoekes | (on 11.x) |
| 13:05.19 | leifmadsen | wdoekes: ah perfect, I'll ignore core show settgings then :) |
| 13:08.21 | wdoekes | if I provide no defaults, I get 1024 soft + 4096 hard by default. but this all depends on the os (and the /etc/security/limits*) and optional re-setting in the init script |
| 13:08.37 | leifmadsen | indeed |
| 13:08.45 | leifmadsen | adding some info, then will update from SVN for your review |
| 13:09.27 | leifmadsen | btw I usually use sysctl.conf for this.... any reason not to? |
| 13:09.49 | wdoekes | sysctl? that would be a global setting then |
| 13:10.12 | leifmadsen | oh gotcha |
| 13:14.24 | *** join/#asterisk-dev Rico29 (~rico@oceanet-telecom-fttb-129-2.olm.fr) |
| 13:14.33 | Rico29 | hi |
| 13:16.08 | wdoekes | (oh.. the core show settings wasn't the one printing the max files. it was set_ulimit() from the rasterisk.) |
| 13:16.21 | wdoekes | (in any case: "Not Set" means "Not Modified") |
| 13:22.20 | leifmadsen | +1 |
| 13:55.58 | *** join/#asterisk-dev serafie (~erin@nat/digium/x-pawdtibtkboarfbm) |
| 13:56.53 | seanbright | so latest asterisk 1.8, taking calls from sip clients with dtmf over rfc2833 and sending them out over iax2 to another box |
| 13:57.01 | seanbright | the dialplan is basically: |
| 13:57.15 | seanbright | exten => 1,1,Answer(1500) |
| 13:57.27 | seanbright | exten => 1,2,Dial(IAX2/foo/baryay) |
| 13:57.59 | seanbright | just turned on debug and a DSP got added to those channels for some reason |
| 13:58.23 | seanbright | getting these spewing through the logs: |
| 13:58.25 | seanbright | [Feb 1 07:56:03] DEBUG[17791]: dsp.c:585 tone_detect: tone 1100, Ew=9.63E+05, Et=2.07E+09, s/n= 0.00 |
| 13:59.18 | seanbright | i'm assuming chan_sip is the culprit (it's one of the few places that calls ast_dsp_new), but i have no idea why it would do that when inband dtmf isn't being used |
| 13:59.19 | opticron | huh, I don't think IAX2 is even capable of sending DTMF inband |
| 13:59.57 | opticron | do you know what version this started in? |
| 14:00.38 | seanbright | i don't |
| 14:00.43 | seanbright | 1.2 |
| 14:00.44 | seanbright | heh |
| 14:00.50 | opticron | har har |
| 14:01.52 | seanbright | we're sporadically seeing high load on these machines so i've been tweaking things to try and tease more information out asterisk |
| 14:02.07 | seanbright | out of* |
| 14:02.19 | seanbright | and i just came across it... seems weird |
| 14:05.54 | seanbright | hmm, unless i am reading something wrong, it looks like chan_sip creates a dsp unconditionally |
| 14:07.18 | seanbright | chan_sip.c:7513 |
| 14:13.37 | seanbright | oh, i see. enable_dsp_detect peels away features before creating the DSP and if none are left just bails out. |
| 14:13.55 | seanbright | err, the other way around, but same diff |
| 14:14.30 | opticron | right |
| 14:14.43 | opticron | do you have fax detection enabled? |
| 14:15.53 | opticron | faxdetect=yes will do it |
| 14:16.37 | seanbright | un-fucking-believeable |
| 14:16.43 | seanbright | thank you, opticron. |
| 14:17.05 | opticron | try faxdetect=t38 instead :) |
| 14:17.14 | opticron | if you need fax stuff, anyway |
| 14:18.28 | seanbright | we don't. i have no idea who set it or why it was set. |
| 14:18.33 | opticron | heh |
| 14:19.27 | leifmadsen | wdoekes: most of your feedback I think has been addressed and the changes are now up on the ofps site |
| 14:19.40 | leifmadsen | let me know if screwed up anything :) |
| 14:39.14 | wdoekes | nice |
| 14:42.14 | wdoekes | "The former replace[s] instead of overwrites." |
| 14:45.04 | wdoekes | (but I'm not fond of the phrasing) |
| 14:48.14 | *** join/#asterisk-dev w9sh (~chatzilla@64.238.96.125) |
| 14:51.45 | leifmadsen | wdoekes: which chapter is that? |
| 14:51.52 | leifmadsen | ya, I don't think I like it either |
| 15:18.18 | Qwell | wdoekes: This build system stuff is making me go gray. |
| 15:18.28 | wdoekes | hehe |
| 15:22.24 | wdoekes | ao2_locks are recursive right? |
| 15:23.09 | wdoekes | I have a chanspy lockup, and I don't immediately see someone else using this channel from the backtrace |
| 15:24.33 | Qwell | wdoekes: Do you remember at what point you canceled the build, when you got that failure? |
| 15:25.29 | wdoekes | early, when using make -j 7 |
| 15:25.35 | wdoekes | the error doesn't show up until later |
| 15:26.17 | Qwell | I'm wondering if it's not remnants from the previous patch. |
| 15:27.39 | wdoekes | possible.. I'll go fetch a clean checkout.. again ;) |
| 15:27.51 | Qwell | sorry.. this stuff is being a PITA. |
| 15:27.55 | Qwell | It needs to be re-written - a lot. |
| 15:30.43 | Qwell | okay, reproduced. somehow. |
| 15:30.56 | Qwell | HA! |
| 15:31.03 | Qwell | wdoekes: You're going to love this. |
| 15:31.10 | wdoekes | not sure |
| 15:31.15 | Qwell | cc1: warnings being treated as errors |
| 15:31.15 | Qwell | conftest.c:12: error: function declaration isn't a prototype |
| 15:31.16 | Qwell | aconfigure:2835: error: C compiler cannot create executables |
| 15:31.58 | Qwell | in res/Makefile, kill the $(_ASTCFLAGS) on the ./configure line. |
| 15:32.10 | Qwell | (then rm res/pjproject/build.mak) |
| 15:32.33 | Qwell | nevermind, don't rm it - it won't be there |
| 15:33.00 | Qwell | --enable-dev-mode adds -Werror. pjproject has lots of warnings (which I'm also going to fix at some point) |
| 15:33.42 | wdoekes | crap.. |
| 15:33.53 | wdoekes | good find :) |
| 15:34.14 | Qwell | I ran into that before... I thought I removed it. I don't know what happened there. |
| 15:34.52 | wdoekes | whoah! a build without errors! |
| 15:34.57 | Qwell | ! |
| 15:38.26 | Qwell | wdoekes: PS, next AstriCon, you will not be going without beers and/or apple juices. |
| 15:38.43 | wdoekes | ok.. 2 builds without fail.. you get the ship-it |
| 15:38.44 | Qwell | (assuming you're there. and you totally should be.) |
| 15:38.55 | Qwell | awesome! would you mind doing so on the review? |
| 15:41.09 | Qwell | It's no wonder our build system is in such a bad shape though. That stuff is brain-melting. |
| 15:42.24 | wdoekes | ya.. the -include magic which may fail once doesn't make things clearer either |
| 15:42.44 | Qwell | it shouldn't fail |
| 15:42.48 | Qwell | oh, in the top-level |
| 15:42.49 | Qwell | yeah |
| 15:43.03 | wdoekes | I see exit, it doesn't exit, what |
| 15:43.33 | Qwell | on the make clean before ./configure, you mean? |
| 15:43.37 | wdoekes | yea |
| 15:44.50 | Qwell | if there weren't so many targets where makeopts was necessary, I'd do similar magic that I did for pjproject's -include |
| 15:55.50 | *** part/#asterisk-dev Rico29 (~rico@oceanet-telecom-fttb-129-2.olm.fr) |
| 16:07.44 | seanbright | possibly a dumb question, but if you have faxdetect enabled, shouldn't it stop trying to recognize fax tones at some point? |
| 16:07.52 | seanbright | like... 30 seconds into answered call? |
| 16:13.48 | *** join/#asterisk-dev pabelanger (~pabelange@asterisk/contributor-and-bug-marshal/pabelanger) |
| 16:13.48 | *** mode/#asterisk-dev [+o pabelanger] by ChanServ |
| 16:18.36 | Juggie | seanbright, that would make sense |
| 16:53.24 | seanbright | putnopvut: question - if i set ast_threadpool_options.initial_size to 10, and idle_timeout to 30 and create a pool |
| 16:53.37 | seanbright | assuming nothing happens in 30 seconds will all of those threads die? |
| 16:53.53 | seanbright | yes, i am too lazy to look at the code |
| 16:53.54 | seanbright | :-) |
| 16:53.55 | putnopvut | I believe they should. |
| 16:54.05 | seanbright | gotcha |
| 16:54.16 | putnopvut | The 10 threads will start active and immediately go idle. After 30 seconds of idleness they should die off. |
| 16:54.26 | seanbright | would you be opposed to me adding a min_size option? |
| 16:54.32 | putnopvut | Not at all. |
| 16:54.35 | seanbright | alrighty |
| 16:56.24 | file | I oppose everything |
| 17:26.21 | *** join/#asterisk-dev tzafrir_laptop (~tzafrir@91.183.118.5) |
| 17:26.21 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 17:32.58 | leifmadsen | file: do you oppose opposable thumbs? |
| 18:38.28 | *** join/#asterisk-dev tzafrir_laptop (~tzafrir@59.157-240-81.adsl-dyn.isp.belgacom.be) |
| 18:38.28 | *** mode/#asterisk-dev [+o tzafrir_laptop] by ChanServ |
| 19:10.02 | *** join/#asterisk-dev anonymouz666 (~anonymouz@189-25-136-163.user.veloxzone.com.br) |
| 19:11.53 | anonymouz666 | how hard would be to run let's say AMD in EARLY media state (before answer)? |
| 21:49.34 | opticron | anyone interested in reviewing integration tests? https://reviewboard.asterisk.org/r/2268/ |
| 22:01.20 | *** join/#asterisk-dev mjordan (~mjordan@28.254-240-81.adsl-static.isp.belgacom.be) |
| 22:01.20 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |
| 23:11.29 | *** join/#asterisk-dev mjordan (~mjordan@28.254-240-81.adsl-static.isp.belgacom.be) |
| 23:11.29 | *** mode/#asterisk-dev [+o mjordan] by ChanServ |