| 00:19.53 | *** join/#asterisk infobot (ibot@208.53.50.136) |
| 00:19.53 | *** topic/#asterisk is #asterisk The Open Source PBX and Telephony Platform (asterisk.org) -=- LTS: 13.22.0 (2018/07/12), Standard: 15.5.0 (2018/07/12); DAHDI: DAHDI-linux 2.11.1 (2016/03/01), DAHDI-tools 2.11.1 (2016/03/01); libpri 1.6.0 (2017/01/27) -=- Wiki: wiki.asterisk.org -=- Code of Conduct: bit.ly/1hH6P22 -=- Logs: bit.ly/1s4AKKu |
| 00:34.22 | *** join/#asterisk puzzola (~puzzola@unaffiliated/puzzola) |
| 01:11.45 | *** join/#asterisk kfife (~Miranda@home.chicagoventure.com) |
| 01:16.19 | kfife | Is running Asterisk (i.e. Dahdi) on FreeBSD a well supported configuraiton? |
| 02:19.35 | *** join/#asterisk forgotmynick (uid24625@gateway/web/irccloud.com/x-wspfyirzlwjaljcd) |
| 02:21.46 | *** join/#asterisk znoteer_ (~znoteer_@104-222-127-176.cpe.teksavvy.com) |
| 02:26.59 | *** join/#asterisk MrC0de (60510431@gateway/web/freenode/ip.96.81.4.49) |
| 02:28.14 | MrC0de | if I use set_var=myVar=123 in pjsip.conf, can it be read back anywhere other than a channel (ie from the endpoint?) |
| 02:49.03 | Samot | No |
| 02:49.10 | Samot | It sets channel vars |
| 03:32.51 | *** join/#asterisk jjrh (~weechat12@ppp-199-167-117-191.storm.ca) |
| 03:41.16 | *** join/#asterisk tuxd00d (~tuxd00d@unaffiliated/tuxd00d) |
| 05:18.44 | *** join/#asterisk tehgooch (tehgooch@unaffiliated/tehgooch) |
| 05:23.28 | *** join/#asterisk gerhard7 (~gerhard7@ip5657ee30.direct-adsl.nl) |
| 07:07.45 | *** join/#asterisk Sepultura (~Sepultura@unaffiliated/sepultura) |
| 07:19.31 | *** join/#asterisk SunTsu (miyamoto@unaffiliated/suntsu) |
| 07:34.02 | *** join/#asterisk SunTsu (miyamoto@unaffiliated/suntsu) |
| 07:38.34 | *** join/#asterisk miralin (~Thunderbi@194.8.128.63) |
| 07:38.56 | *** join/#asterisk drathir (~kamiljk8@unaffiliated/drathir) |
| 07:50.17 | *** join/#asterisk jkroon (~jkroon@165.16.203.126) |
| 08:00.38 | *** join/#asterisk hehol (~hehol@gatekeeper.loca.net) |
| 08:02.31 | MrC0de | ty Samot,there isnt a way to set it directly on the endpoint likein chan_sip is there? |
| 08:45.23 | *** join/#asterisk tomaluca95 (~quassel@kde/developer/tomaluca) |
| 08:55.51 | *** join/#asterisk tomaluca95 (~quassel@kde/developer/tomaluca) |
| 08:59.03 | *** join/#asterisk pvoigt (~Linux@unaffiliated/pvoigt) |
| 09:06.30 | *** join/#asterisk sibyakin (~sibyakin@83.234.69.145) |
| 09:11.36 | *** join/#asterisk tomaluca95 (~quassel@kde/developer/tomaluca) |
| 09:48.04 | *** join/#asterisk Eloy (~Eloy@195.191.28.250) |
| 10:14.55 | *** join/#asterisk sekil (~sekil@nat-73.net011.net) |
| 10:43.52 | *** join/#asterisk lankanmon (~LKNnet@CPE64777d632383-CM64777d632380.cpe.net.cable.rogers.com) |
| 10:45.59 | jkroon | is it possible to have asterisk offer encryption on outbound INVITEs but not require it? ie, allow the remote end to encrypt if it supports it, but not force it. I get that this can be man-in-the-middled, but I need migration path, not all endpoints currently support encryption, so I'd like to utilize optimistic encryption, and then later only force it once I know all endpoints are fine with it. |
| 10:54.04 | *** join/#asterisk war9407 (war@pool-70-106-250-134.clppva.fios.verizon.net) |
| 11:08.07 | *** join/#asterisk Oatmeal (Suzeanne@gateway/vpn/privateinternetaccess/suzeanne) |
| 11:18.32 | jkroon | MrC0de, setvar=var=val, same as in chan_sip.conf |
| 11:27.48 | file | jkroon: in chan_pjsip yes, in chan_sip no (re: optimistic) |
| 11:28.59 | jkroon | file, set that, but it still doesn't seem to offer any non-encrypted options. asterisk 13.21.0 - possible bug? media_encryption : dtls, media_encryption_optimistic : true |
| 11:29.10 | file | DTLS can't be optimistic |
| 11:29.26 | jkroon | oh crap. |
| 11:29.38 | file | well, the code doesn't allow it, and last I checked there wasn't a spec for it - but that could have changed |
| 11:29.39 | jkroon | ok, and for webrtc one has to use dtls? |
| 11:29.42 | file | yes. |
| 11:29.57 | file | have to use ICE, have to use DTLS |
| 11:31.03 | jkroon | ok ... that presents a few problems. So should probably have two profiles, one where webrtc will function, and one where everything else will. odd, so far it's only some older Grandstream phones that gave me issues. |
| 11:31.36 | file | yes, that is what people do |
| 11:32.12 | jkroon | yea, i was just hoping to avoid that. never needed that kind of distinction with chan_sip ... but alas. |
| 11:32.53 | file | people have brought it up, but after explaining it, and them looking into it, they never go further |
| 11:33.00 | file | it's not an easy problem to solve for everything |
| 11:33.30 | jkroon | using realtime, is the UA string perhaps available somehow as part of the query issued to odbc? |
| 11:33.39 | file | no. |
| 11:33.49 | jkroon | ok. there goes that idea. |
| 11:34.43 | jkroon | ok, the "to be avoided" profile option will unfortunately have to be added, only really two options, "default" (which enables dtls et al) and "legacy" for stuff that can't do dtls. |
| 11:35.16 | jkroon | thanks file. |
| 11:35.28 | file | aye |
| 11:35.36 | *** join/#asterisk nighty- (~nighty@s229123.ppp.asahi-net.or.jp) |
| 11:39.12 | *** join/#asterisk miralin (~Thunderbi@194.8.128.63) |
| 11:45.24 | jkroon | file, since it's in-db, whenever a call goes out one could detect usupported ua strings, and switch to the legacy profile automatically, from time to time cleaning up again if those UA strings haven't been seen for an endpoint for a few days. anyway, for now just need to get it working. |
| 11:46.06 | file | if someone wants to contribute something to do this which doesn't break anything else, has defined use cases and tests, then we'll review it |
| 11:46.09 | file | but so far noone has |
| 11:47.08 | jkroon | file, i'm not going to either. simpler solution is to have multiple profiles. |
| 11:47.14 | file | exactly |
| 11:47.20 | file | which is why noone has really invested the time |
| 11:47.51 | jkroon | yea, and i suspect most users of asterisk control the environment (phones) more closely, so they can dictate what the configuration should look like on the phone itself. |
| 11:48.18 | file | even if you don't, you generally control the WebRTC portion |
| 11:52.27 | jkroon | so my use case requires some hard phones along with webrtc ... so i'll just say "technical limitation, if you wan't webrtc you can't use phones that doesn't support dtls" |
| 11:52.40 | jkroon | or they just need to be on different extensions ... |
| 11:54.09 | file | user accounts and extensions in dialplan are separate things. |
| 11:54.41 | file | if it's WebRTC then you likely control the WebRTC part as I said, so if they provide a username... you can place -webrtc at the end internally or something |
| 12:09.22 | *** join/#asterisk hehol (~hehol@gatekeeper.loca.net) |
| 12:22.31 | jkroon | thank you file. |
| 12:28.34 | *** join/#asterisk qxork (~qxork@unaffiliated/qxork) |
| 12:42.22 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 12:50.39 | *** join/#asterisk depate (~depate@dyn-054810efff6911a095001000.eduroam.ipv6.ruhr-uni-bochum.de) |
| 12:50.39 | *** join/#asterisk Iamnach0 (~Iamnacho@ip68-102-129-52.ks.ok.cox.net) |
| 12:50.39 | *** join/#asterisk ghoti_ (ghoti@dev-160.experiencepoint.com) |
| 12:51.44 | *** join/#asterisk woleium (~woleium@bc.io) |
| 12:51.48 | *** join/#asterisk _nexxus_ (~bwg@leon.generalamalgamated.com) |
| 12:53.40 | *** join/#asterisk yoavz (~yoavz@82.166.176.37) |
| 12:54.35 | *** join/#asterisk juvenal (Elite21271@gateway/shell/elitebnc/x-eyeenyziangktced) |
| 12:58.37 | *** join/#asterisk emk (~emk@unaffiliated/emk) |
| 13:08.05 | *** join/#asterisk znoteer (~znoteer@104-222-127-176.cpe.teksavvy.com) |
| 13:09.52 | *** part/#asterisk znoteer_ (~znoteer_@104-222-127-176.cpe.teksavvy.com) |
| 13:11.35 | *** join/#asterisk Oatmeal (Suzeanne@gateway/vpn/privateinternetaccess/suzeanne) |
| 13:22.11 | *** join/#asterisk znoteer_ (~znoteer_@104-222-127-176.cpe.teksavvy.com) |
| 13:27.46 | *** join/#asterisk Asgaroth (~Asgaroth@212.2.172.228) |
| 13:28.45 | *** join/#asterisk [TK]D-Fender (~joe@216.191.106.165) |
| 13:37.59 | MrC0de | thanks file |
| 13:40.16 | *** join/#asterisk gerhard7 (~gerhard7@ip5657ee30.direct-adsl.nl) |
| 13:42.50 | file | what what |
| 13:43.03 | MrC0de | thanks jkroon, file u replied to him lol |
| 13:43.15 | MrC0de | its early lol |
| 13:43.41 | jkroon | yea, that was rather confusing. |
| 13:44.35 | MrC0de | i feel like setvar=val=val gave me the 'no suitable category' but ill try again... after coffee |
| 13:46.32 | Samot | 4:02:32 AM <MrC0de> ty Samot,there isnt a way to set it directly on the endpoint likein chan_sip is there? |
| 13:46.39 | Samot | ^^ I have no idea what you mean. |
| 13:46.57 | Samot | setvar and set_var do the same thing on each of their respective drivers. |
| 13:47.14 | Samot | If I setvar something on Chan_SIP it doesn't mean the endpoint will have it. |
| 13:47.21 | Samot | It's a channel variable. |
| 13:47.27 | Samot | It exists as long as the channel exists. |
| 13:47.54 | MrC0de | when I used it in chan_sip, modified code to copy it into peer, makes it show up under 'sip show peers' |
| 13:48.02 | Samot | OK.. |
| 13:48.09 | Samot | So when you are saying "endpoint" |
| 13:48.14 | Samot | You mean in the peer configuration |
| 13:48.17 | Samot | Not an actual device. |
| 13:48.56 | *** join/#asterisk jeffspeff (~jeffspeff@12.49.160.131) |
| 13:49.00 | MrC0de | trying to do same in pjsip, thought it would be endpoint or aor or contact |
| 13:50.11 | MrC0de | the real problem is probably that I've been doing pjsip for about a day lol, hopefully will figure what 'they' want |
| 13:50.29 | Samot | Where are you putting the set_var ? |
| 13:51.06 | Samot | You are putting it in the [endpoint] section? |
| 13:51.23 | file | set_var goes in the PJSIP endpoint and configures a channel variable that is set on channels created using it, they do appear in "pjsip show endpoint" as the name of the variable and the value (set_var=test=1234 would show as test: 1234) |
| 13:52.30 | Samot | Yup. |
| 13:52.37 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 13:53.41 | MrC0de | ah it wasnt setvar for the peer, just remembered, trying to replicate the effect of 'myvar=123' that we had set directly on the thingamajiggy in sip.conf |
| 13:53.57 | Samot | setvar is not a PJSIP variable. |
| 13:54.00 | Samot | set_var is |
| 13:54.38 | *** join/#asterisk khamsin (khamsin@lowtek.net) |
| 13:55.26 | MrC0de | its a wierd setup and wierd request, and doesnt seem to work in pjsip, i still dont know what they dont want to use accountcode instead |
| 13:55.39 | Samot | Show what you have |
| 13:55.40 | Samot | ~pb |
| 13:55.41 | infobot | i heard pastebin is a web-based service where you should paste anything over 3 lines so you don't flood the channel. Here are links to a few: http://pastebin.ca, http://channels.debian.net/paste, http://paste.lisp.org, http://bin.cakephp.org/; or install pastebinit with yum or aptitude. |
| 13:56.01 | file | you can't put arbitrary configuration fields in pjsip.conf entries, I'll say that |
| 13:56.10 | Samot | ^^^ 100% |
| 13:56.12 | file | it considers that a broken/invalid configuration and rejects it |
| 13:56.15 | MrC0de | thats what they want to do.. yep.. |
| 13:56.21 | MrC0de | and thats exactly what happens |
| 13:56.27 | Samot | Show what you've got |
| 13:56.33 | MrC0de | i gtg to work :P I dont have it here |
| 13:56.38 | Samot | OK |
| 13:56.42 | Samot | Well without looking at it |
| 13:56.48 | Samot | We don't know if it really is incorrect or not |
| 14:04.19 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 14:07.13 | imcdona | "sip show peers" will give you the total number of peers along with the number of offline and online peers. "pjsip show endpoints" provides the number of endpoints, but not the number of online endpoints. Anyone know how I can obtain the number of online/offline pjsip endpoints? |
| 14:08.27 | imcdona | I'm thinking "pjsip show contacts" is best. Answered my own question. |
| 14:14.45 | *** join/#asterisk brad_mssw (~brad@66.129.88.50) |
| 14:15.50 | *** part/#asterisk Billy2 (~Billy2@208-85-2-206.static.as40244.net) |
| 14:21.14 | *** join/#asterisk salviadud (~ralfalfa@187-167-69-132.static.axtel.net) |
| 14:32.10 | *** join/#asterisk kharwell (kharwell@nat/digium/x-xpsrklmyschmqejc) |
| 14:32.10 | *** mode/#asterisk [+o kharwell] by ChanServ |
| 14:36.41 | *** join/#asterisk jamesaxl (~James_Axl@41.248.195.234) |
| 14:38.02 | *** join/#asterisk bford (uid283514@gateway/web/irccloud.com/x-wiofxmtxpyjuwvwz) |
| 14:38.02 | *** mode/#asterisk [+o bford] by ChanServ |
| 14:44.12 | *** join/#asterisk rmudgett (rmudgett@nat/digium/x-hfwvcikyxmdhpbdi) |
| 14:44.12 | *** mode/#asterisk [+o rmudgett] by ChanServ |
| 14:53.08 | *** join/#asterisk cresl1n (uid299068@asterisk/libpri-and-libss7-expert/Cresl1n) |
| 14:53.08 | *** mode/#asterisk [+o cresl1n] by ChanServ |
| 14:58.47 | *** join/#asterisk tuxd00d (~tuxd00d@unaffiliated/tuxd00d) |
| 14:58.55 | *** join/#asterisk exuberocity (~exuberoci@rrcs-162-155-56-66.central.biz.rr.com) |
| 15:27.03 | *** join/#asterisk craigify (~craigify@ip184-186-18-123.no.no.cox.net) |
| 15:54.34 | *** join/#asterisk gruetzkopf (gruetzkopf@bnc.dont-follow.us) |
| 16:00.15 | *** join/#asterisk SebastienThiry (~Thunderbi@109.134.17.12) |
| 16:11.55 | *** join/#asterisk exuberocity (~exuberoci@rrcs-162-155-56-66.central.biz.rr.com) |
| 16:29.59 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 16:31.06 | *** part/#asterisk mitrax (mitrax@LFbn-NCY-1-325-94.w83-196.abo.wanadoo.fr) |
| 16:44.47 | *** join/#asterisk sinaowolabi (~Sina@41.58.118.98) |
| 17:09.09 | *** join/#asterisk jkroon (~jkroon@105.4.6.160) |
| 17:19.35 | *** join/#asterisk jkroon (~jkroon@105.4.6.160) |
| 17:21.35 | *** join/#asterisk degenerate (~degenerat@S0106cc2de0099182.no.shawcable.net) |
| 17:22.51 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 17:31.19 | *** join/#asterisk jcongdon (~jcongdon@cpe-76-176-105-117.san.res.rr.com) |
| 17:33.38 | jcongdon | Hello, I recently upgraded from v11 to v15 with hopes of using the Media Cache. We have 4 asterisk boxes, and noticed that the load average is now running much much higher. One of the four happened to have a great load average and the difference in modules came down to SPEEX and Media Cache. On one server we disabled all of those modules and it appears to be behaving MUCH better today. Any advice, or am I SOL on usi |
| 17:33.38 | jcongdon | ng the media cache? |
| 17:34.44 | *** join/#asterisk sekil (~sekil@87.116.191.36) |
| 17:43.33 | *** join/#asterisk jameswf (uid27319@gateway/web/irccloud.com/x-zlendvxwrdnkqsmf) |
| 17:49.21 | kfife | Is running Asterisk (DAHDI) on FreeBSD a well supported configuration? |
| 17:49.28 | kfife | or is that |
| 17:49.32 | kfife | fringy |
| 17:53.19 | igcewieling | jcongdon: how does it look if you only disable speex? |
| 17:53.45 | jcongdon | We did that on a second server, and it didn't change the results :( |
| 17:54.04 | file | kfife: the project does not test or develop for or on FreeBSD in any capacity |
| 17:54.12 | jcongdon | So we had 3 servers with a 176 modules. And one with 167. The one with 167 performs the best. |
| 17:54.16 | kfife | That's all I need to hear. |
| 17:54.42 | kfife | Bummer that ZFS-on-root isn't well supported in Linux yet. That's the magic combination. |
| 17:54.48 | kfife | Thanks file |
| 17:54.54 | Samot | When you say "performs the best" what does that mean? |
| 17:55.05 | jcongdon | So on one server we disabled the other 9 modules, and now it seems to perform well. On another we only removed the 3 speex modules, and it didn't change. |
| 17:55.10 | Samot | What is the load average you are seeing? |
| 17:55.27 | jcongdon | We are only noticing issues on recorded prompts. Actual voice conversation seems to be fine. |
| 17:55.50 | Samot | What is the load average you are seeing? |
| 17:55.54 | jcongdon | When load is around 10, it seems fine. 30+ gets choppy. But we are seeing upwards of 100+ at some times |
| 17:56.03 | Samot | 10? |
| 17:56.15 | Samot | What are you using to determine the load? |
| 17:56.19 | jcongdon | uptime |
| 17:56.35 | Samot | So you mean 1.0, 3.0, etc? |
| 17:56.37 | jcongdon | The system load average |
| 17:56.38 | jcongdon | yes |
| 17:56.42 | Samot | The 5/10/15 minute averages.. |
| 17:56.43 | Samot | OK |
| 17:56.51 | Samot | So how many cores/CPUs are in these machines? |
| 17:57.50 | jcongdon | varies, 24 in the older machines, 32 in the newer ones |
| 17:58.02 | Samot | OK |
| 17:58.12 | Samot | Show an example |
| 17:58.24 | Samot | Show the output of the 'w' command from the system cli. |
| 17:58.42 | kfife | will Dahdi run on 32-bit Linux? |
| 17:58.48 | jcongdon | ââ[#]⺠w |
| 17:58.48 | jcongdon | <PROTECTED> |
| 17:59.26 | Samot | And is this a 32 or 24 core system? |
| 17:59.29 | kfife | (embedded application) |
| 17:59.30 | jcongdon | 32 |
| 18:00.07 | jcongdon | From the one with 24 that performs the best |
| 18:00.07 | jcongdon | 12:59:41 up 7 days, 11:39, 3 users, load average: 1.80, 5.32, 6.35 |
| 18:01.09 | jcongdon | It really does seem to come down to one of these modules... res_curl, res_config_curl, res_http_media_cache, func_curl |
| 18:01.16 | Samot | And what is happening on that first machine right now? |
| 18:01.26 | jcongdon | We are going to disable those 4 on the "bad" server tonight after hours and try again tomorrow |
| 18:01.49 | jcongdon | They all have between 100-250 sip calls |
| 18:02.03 | Samot | And what else? |
| 18:02.20 | Samot | Recordings since you've mentioned that. |
| 18:02.27 | Samot | Are there queues? |
| 18:02.30 | jcongdon | no |
| 18:02.36 | Samot | Parking lots? |
| 18:02.41 | Samot | A lot of calls on hold? |
| 18:02.43 | jcongdon | recordings are just the voice prompts |
| 18:02.46 | jcongdon | no calls on hold |
| 18:02.58 | Samot | So you're not recording the calls.. |
| 18:03.02 | jcongdon | It's all live traffic. Occasionally transferring / bridging calls |
| 18:03.09 | jcongdon | Yes, we do record some calls |
| 18:03.09 | Samot | You just playing back system recordings? |
| 18:03.12 | Samot | OK |
| 18:03.16 | Samot | That's what I am asking. |
| 18:03.26 | Samot | What is happening on the server when you have 100-250 calls going.. |
| 18:03.27 | jcongdon | Playing back system recordings is when we notice the issue |
| 18:04.00 | jcongdon | All of these things. It's all the same thing happening on all 4 servers. |
| 18:04.32 | jcongdon | We use AGI scripts for lots of things, but that hasn't changed in years |
| 18:04.42 | jcongdon | We use AMI to control calls from our web service |
| 18:05.05 | Samot | Are all the servers using the same location for the media cache? |
| 18:05.06 | jcongdon | It's mostly our user's calling in. And from there we connect them to their contacts |
| 18:05.14 | jcongdon | We aren't using the media cache yet |
| 18:05.19 | jcongdon | We want to in the near future |
| 18:05.31 | jcongdon | That's why we upgraded from v11 to v15 |
| 18:05.35 | Samot | OK. |
| 18:08.07 | Samot | When do the issues start? |
| 18:08.26 | Samot | Do you have issues on those systems when there's 1 call, 10 calls, 50, 100? |
| 18:09.00 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 18:10.09 | *** join/#asterisk joako (~joako@opensuse/member/joak0) |
| 18:10.29 | jcongdon | When load is low, we don't see the issue |
| 18:10.42 | Samot | What is "low"? |
| 18:10.44 | jcongdon | It's normally during the peak of our day, I would estimate 100+ calls |
| 18:10.55 | jcongdon | Sorry, overusing the word "Load" |
| 18:11.09 | jcongdon | When call volume is low (maybe < 50) everything seems fine |
| 18:11.58 | Samot | Now are you only counting PSTN calls? |
| 18:12.07 | jcongdon | But we aren't constantly calling into the server to know exactly when it starts. Just something we noticed recently after a couple of complaints. Then we started digging and only thing we found was the system load average as being a key indicator. |
| 18:12.10 | Samot | Or are you counting PSTN + Internal calls? |
| 18:12.25 | jcongdon | We don't have internal calls. Everything is done via sip |
| 18:12.34 | Samot | Uhm. |
| 18:12.45 | Samot | I mean are there "extension" to "extension" calls. |
| 18:12.49 | jcongdon | no |
| 18:12.53 | Samot | OK. |
| 18:13.11 | Samot | How many endpoints are on the system? |
| 18:13.26 | jcongdon | 47 sip peers |
| 18:13.31 | jcongdon | Is that what you mean? |
| 18:13.36 | Samot | Yes. |
| 18:17.18 | jcongdon | At this point I'm willing to just remove media cache modules tonight on the additional server and report back findings tomorrow |
| 18:17.45 | jcongdon | It worked for one server, if it works for the second... then that has to be what's causing my pain |
| 18:18.38 | jcongdon | Samot: I really appreciate your attempt at fully understanding and helping. It means a ton. |
| 18:20.00 | Samot | Well I'm running 15 and I don't have those problems. |
| 18:20.34 | Samot | And so far one machine has hit 800 calls, all using remote Playback/Background, aka media cache |
| 18:21.10 | Samot | So I'm just trying to figure out what else the machine is doing. |
| 18:21.10 | jcongdon | you definitely have res_http_media_cache.so loaded? |
| 18:21.19 | Samot | Yes otherwise it wouldn't work. |
| 18:21.27 | jcongdon | Why is that? |
| 18:21.46 | jcongdon | Works for me without it. |
| 18:22.02 | jcongdon | Our goal is to eventually playback media that is stored on S3 instead of the individual machines |
| 18:22.25 | Samot | Right |
| 18:22.36 | Samot | That's what the HTTP(S) playback is for. |
| 18:22.41 | jcongdon | Right now everything is served from an NFS mount, but our complete infrastructure is changing making that harder to maintain |
| 18:22.42 | Samot | It was introduced in v14 |
| 18:23.00 | jcongdon | Exactly, but you don't need the media cache for normal playback/background |
| 18:23.06 | Samot | Right |
| 18:23.11 | jcongdon | That's why I was asking if you definitely had it enabled. |
| 18:23.11 | Samot | I didn't say I was doing that |
| 18:23.13 | Samot | I said REMOTE |
| 18:23.23 | Samot | AKA HTTP/S |
| 18:23.33 | Samot | Exactly what you are looking to do. |
| 18:23.51 | Samot | Which is was res_http_media_cache is for. |
| 18:23.56 | Samot | what* |
| 18:24.10 | jcongdon | oh, you did say remote, I missed that :( |
| 18:24.28 | *** join/#asterisk forgotmynick (uid24625@gateway/web/irccloud.com/x-aqqhfeicxyartvyc) |
| 18:24.36 | jcongdon | Can you tell me the machine specs you are using for 800+ calls? |
| 18:25.04 | jcongdon | We are currently trying to limit to 200-250 calls per machine |
| 18:25.17 | jcongdon | Adding 3 more physical servers to handle call volume |
| 18:25.41 | Samot | I'm using a 8core server. |
| 18:26.00 | Samot | And so far have only tested up to 800 calls with the media cache.. |
| 18:26.17 | Samot | Local playback back that bad boy has been tested for about 1600 |
| 18:26.23 | *** join/#asterisk Jesterboxboy (~Thunderbi@84-115-150-8.cable.dynamic.surfer.at) |
| 18:27.06 | jcongdon | Are those 800 bridged calls? Or just doing playback? |
| 18:27.13 | Samot | Uhm. |
| 18:27.25 | Samot | You have to be bridge to do the plackback |
| 18:27.41 | Samot | What good is playback over a non-bridged/answered channel? |
| 18:27.42 | jcongdon | I meant, are you doing actual phone calls too? Person -> Person talking |
| 18:27.48 | Samot | No. |
| 18:27.56 | Samot | These are stripped down alert calls. |
| 18:28.41 | jcongdon | Ours are a person call in, listens to prompts, enters a pin, then we use AMI to start dialing contacts for them to talk to. Actual talking is fine, prompts sound like they're under water. |
| 18:29.12 | Samot | And if you're doing local playback then it could be the recordings |
| 18:29.13 | sibiria | are you transcoding the prompts from a complicated format, maybe |
| 18:29.17 | Samot | ^^^^ |
| 18:29.26 | Samot | What format are the recordings? |
| 18:29.28 | jcongdon | not that I'm aware of |
| 18:29.31 | Samot | What format are the recordings? |
| 18:29.36 | jcongdon | sln I believe |
| 18:29.44 | jcongdon | Again, this didn't start until asterisk 15 |
| 18:29.49 | jcongdon | It was perfectly fine on asterisk 11 |
| 18:29.55 | Samot | OK |
| 18:30.05 | Samot | But there's been 3 releases since. |
| 18:30.14 | Samot | Did you look at the changes from 11 to 12, to 13 to 14? |
| 18:30.36 | jcongdon | I lied, it's a wav file |
| 18:30.59 | Samot | And what codecs are you using for the users? |
| 18:31.01 | sibiria | we keep all our audio files in both ulaw and alaw, since those are the two codecs all are calls end up as, whether incoming or outgoing |
| 18:31.05 | sibiria | and it makes a big difference |
| 18:31.07 | jcongdon | ulaw |
| 18:31.56 | Samot | Well I think you need to take some time and look at the changelogs. |
| 18:32.19 | Samot | There have been numerous updates/changes/new features that have been introduced in 12 and 14 |
| 18:32.24 | Samot | And I think a few in 15. |
| 18:32.26 | jcongdon | ok, I did look through them when upgrading. Looking for issues, but I could have missed something for sure |
| 18:32.59 | Samot | One thing to check for is deprecation. |
| 18:33.16 | Samot | Figure out what you are using in v11 that is deprecated in v15 and stop using that method. |
| 18:34.17 | Samot | Asterisk generally doesn't remove settings/features, they deprecate them. At that point, they are no longer part of the development map. |
| 18:34.52 | jcongdon | Right, I get that. I don't think I'm using any deprecated features. But will look through again. |
| 18:35.17 | jcongdon | At this point, I'm removing media cache tonight, and seeing if that solves my issues. If so, I will find another solution for remote playback. |
| 18:35.29 | Samot | There isn't. |
| 18:35.35 | Samot | That *is* the solution. |
| 18:35.57 | jcongdon | I meant my own solution. Bringing down the files I need before starting a call, then deleting after the call. |
| 18:36.03 | jcongdon | And storing locally |
| 18:36.12 | jcongdon | Only during the time I need them |
| 18:36.29 | Samot | That's going to delay the calls. |
| 18:37.21 | jcongdon | Nah... I know before my user dials in what I need. While they are dialing my server, I can have a job that gets the 3-10 files they will use |
| 18:37.29 | jcongdon | When they hangup I can delete those files |
| 18:37.51 | Samot | Yes, you can do that. |
| 18:38.04 | Samot | But while you are "transferring" those files, the call halts. |
| 18:38.08 | Samot | It waits for that to be done. |
| 18:38.33 | Samot | It's not a factor of figuring out what they need, it's the factor of transferring it. |
| 18:38.42 | jcongdon | This would be done outside of asterisk. So I'm not too concerned about it |
| 18:39.13 | Samot | How would it be done outside of Asterisk? |
| 18:39.19 | jcongdon | I was already doing it with the media cache. Before they actually got connected to my server, I was having the media cache pull files down locally |
| 18:40.05 | jcongdon | My users are on my website. They click a button to start what we call a dial session. At that point we give them a phone number to call in. As they are doing it, I can start some code to download the files they need to the server they are calling. |
| 18:40.26 | Samot | OK |
| 18:40.40 | Samot | so how will the call progress? |
| 18:41.07 | Samot | I think you just need to look at the system configs |
| 18:41.12 | jcongdon | Like normal... |
| 18:41.42 | Samot | "Like normal"? |
| 18:41.45 | jcongdon | I agree, I will have to look through all of them. I can't see any reason the http media cache would cause this issue... |
| 18:41.56 | Samot | v11 configs might not have the same current settings as v15 configs. |
| 18:42.07 | Samot | Functions may have been changed |
| 18:42.14 | Samot | It shouldn't. |
| 18:42.33 | Samot | Specially since you're not even using it, as you said. |
| 18:42.43 | Samot | It's not even putting media into the cache. |
| 18:43.12 | jcongdon | I think there's a disconnect on how we use Asterisk vs how you do. That's why I'm so confident that the call will progress without blocking. |
| 18:43.56 | jcongdon | We control the flow ourselves through AMI. There's some dialplan stuff, but not a ton. |
| 18:44.07 | Samot | That's just one use of Asterisk I use for one client. |
| 18:44.54 | Samot | I run entire ITSPs off Asterisk setups. |
| 18:47.36 | jcongdon | Ok, thanks again for all the advice, you're awesome. I have to head out for a while. |
| 18:48.45 | kfife | Anyone know if there's any meaningful difference between the older AEX800 cards and the newer A8B ? |
| 18:49.32 | kfife | I don't see the old molex connector. Are they able to get enough juice from the PCIe bus to generate a ring without 'jumper cables' directly to the 12v bus? |
| 18:52.24 | kfife | Actually the A8 also has the molex connector. Just hidden until you look in the manual. |
| 18:52.38 | *** part/#asterisk jcongdon (~jcongdon@cpe-76-176-105-117.san.res.rr.com) |
| 18:52.38 | kfife | Any meaningful differences? Reliability? |
| 19:11.43 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 19:29.50 | *** join/#asterisk BlackBishop (dexter@ipv6.d3xt3r01.tk) |
| 19:48.16 | *** part/#asterisk BlackBishop (dexter@ipv6.d3xt3r01.tk) |
| 19:50.44 | *** join/#asterisk troyt (~troyt@c-73-65-211-70.hsd1.ut.comcast.net) |
| 19:53.09 | [TK]D-Fender | <kfife> Actually the A8 also has the molex connector. Just hidden until you look in the manual. <- or the card |
| 19:54.15 | *** join/#asterisk jkroon (~jkroon@165.16.204.168) |
| 19:55.45 | *** join/#asterisk troyt (~troyt@c-73-65-211-70.hsd1.ut.comcast.net) |
| 20:03.23 | *** join/#asterisk BlackBishop (dexter@ipv6.d3xt3r01.tk) |
| 20:03.27 | BlackBishop | asterisk: malloc.c:4030: _int_malloc: Assertion `(unsigned long) (size) >= (unsigned long) (nb)' failed. |
| 20:03.29 | BlackBishop | [1] 20077 abort asterisk -vvvvvvv |
| 20:03.38 | BlackBishop | when doing "asterisk -vvvvvvvvv" .. |
| 20:03.51 | BlackBishop | any easy way to find what I'm doing wrong ? |
| 20:04.25 | BlackBishop | I think it's something to do with res_odbc |
| 20:07.32 | *** join/#asterisk scgm11_ (~scgm11@r186-50-149-43.dialup.adsl.anteldata.net.uy) |
| 20:09.46 | BlackBishop | adding -dddddd dies like this: https://paste.fedoraproject.org/paste/~Elm2gw0Z5HDtu4GDj-9Yg |
| 20:13.53 | BlackBishop | hah, if I run multiple times it dies with various other messages |
| 20:15.27 | Samot | What are you trying to do? |
| 20:15.41 | BlackBishop | 'free(): invalid pointer' 'segmentation fault' 'free(): invalid next size (fast)' ... |
| 20:15.48 | BlackBishop | Samot: right now ? just start the cli |
| 20:15.58 | BlackBishop | and test my extensions.ael |
| 20:16.08 | Samot | OK so that's an issue in AEL |
| 20:16.21 | Samot | Which is python-esque, right? |
| 20:16.45 | BlackBishop | Samot: nope, it's not ael's bug .. I think it's related to myodbc |
| 20:16.45 | Samot | Or LUA |
| 20:17.02 | BlackBishop | and asterisk dies for some reason |
| 20:17.19 | BlackBishop | I just upgraded asterisk .. haven't used it in a long while now |
| 20:17.52 | Samot | OK but AEL isn't a standard roll out |
| 20:18.11 | BlackBishop | yeah, but I didn't get to the ael part :) |
| 20:18.26 | Samot | So I'm trying to get some ideas here because that error is about running out of memory or going "out of bounds" based on the quick search I just did... |
| 20:18.31 | BlackBishop | mv res_odbc.conf res_odbc.conf.backup seems to allow it to start |
| 20:19.34 | BlackBishop | and there I only some connection details set |
| 20:20.24 | *** join/#asterisk scgm11__ (~scgm11@r186-52-156-184.dialup.adsl.anteldata.net.uy) |
| 20:20.28 | BlackBishop | so I think the link between asterisk and myodbc fails somehow |
| 20:21.19 | *** join/#asterisk miralin (~Thunderbi@194.8.128.63) |
| 21:49.32 | *** join/#asterisk Eloy (~Eloy@177.red-81-32-207.dynamicip.rima-tde.net) |
| 22:07.26 | *** join/#asterisk tomaluca95 (~quassel@kde/developer/tomaluca) |
| 22:07.41 | *** join/#asterisk qxork (~qxork@unaffiliated/qxork) |
| 22:39.24 | *** join/#asterisk tehgooch (tehgooch@unaffiliated/tehgooch) |
| 22:56.36 | *** join/#asterisk u0m3 (~u0m3@188.25.10.217) |
| 23:13.21 | *** join/#asterisk davlefou (~davlefou@unaffiliated/davlefou) |
| 23:37.48 | *** part/#asterisk kharwell (kharwell@nat/digium/x-xpsrklmyschmqejc) |