| 00:44.48 | snuff-work | mm.. |
| 00:44.55 | snuff-work | anyone seen an issue like this before.. |
| 00:45.06 | snuff-work | [Nov 21 11:40:37] WARNING[5005][C-00000013]: chan_pjsip.c:948 chan_pjsip_write_stream: Channel PJSIP/aapt_syd-out-0000002d asked to send alaw frame when native formats are (g729) (rd:alaw->slin16;(alaw@8000)->(slin@8000)->(slin@16000) wr:slin16->alaw;(slin@16000)->(slin@8000)->(alaw@8000)) |
| 00:45.24 | snuff-work | now i know my pjsip doesn't allow g729 anywhere.. |
| 00:45.31 | snuff-work | so no idea where it is coming from |
| 04:38.07 | *** join/#asterisk-dev sahmed (~sahmed@cpe-70-114-236-63.austin.res.rr.com) |
| 07:10.16 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01:e0a:d:1111:b749:768d:ed77:7c9a) |
| 07:19.46 | *** join/#asterisk-dev pchero_work (~pchero@87.213.247.82) |
| 07:20.18 | *** join/#asterisk-dev pchero_work (~pchero@87.213.247.82) |
| 07:29.24 | *** join/#asterisk-dev jkroon (~jkroon@41.113.35.102) |
| 08:07.08 | *** join/#asterisk-dev spatel (~spatel@pool-96-237-230-175.bstnma.fios.verizon.net) |
| 09:09.02 | *** join/#asterisk-dev jkroon (~jkroon@41.113.122.97) |
| 09:10.32 | *** join/#asterisk-dev CELYA_ (~Thunderbi@185.14.178.189) |
| 12:46.46 | *** join/#asterisk-dev pc-m (~pcm@modemcable094.94-70-69.static.videotron.ca) |
| 13:26.25 | *** join/#asterisk-dev sahmed (~sahmed@66.64.44.146) |
| 13:41.37 | *** join/#asterisk-dev gtjoseph (~gtjoseph@unaffiliated/gtj) |
| 13:41.46 | *** mode/#asterisk-dev [+o gtjoseph] by ChanServ |
| 14:13.12 | *** join/#asterisk-dev sduthil (~sduthil@modemcable094.94-70-69.static.videotron.ca) |
| 15:08.58 | *** join/#asterisk-dev spatel (~spatel@static-71-174-102-210.bstnma.fios.verizon.net) |
| 15:15.58 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-dxpddoasyxqttcan) |
| 15:15.58 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 15:21.21 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-mjgxznnazxlsdzvt) |
| 15:21.21 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 15:31.15 | elguero | snuff-work: Maybe related to https://issues.asterisk.org/jira/browse/ASTERISK-28139 ... just a guess |
| 16:24.04 | *** join/#asterisk-dev aliasbex (~aliasbex1@80-123-31-162.adsl.highway.telekom.at) |
| 16:45.23 | *** join/#asterisk-dev jkroon (~jkroon@165.16.204.110) |
| 18:33.25 | *** join/#asterisk-dev sahmed (~sahmed@cpe-70-114-236-63.austin.res.rr.com) |
| 18:36.26 | *** join/#asterisk-dev CELYA_ (~Thunderbi@2a01cb0583d31d00f4d7ddd1bc0987bc.ipv6.abo.wanadoo.fr) |
| 18:52.12 | seanbright | for ASTERISK-28625, the patch is fine, but i'm more curious why it takes 15-30 seconds for a cache lookup |
| 18:57.36 | file | seanbright: an excellent ponderance |
| 19:20.56 | *** join/#asterisk-dev CELYA_1 (~Thunderbi@anantes-655-1-124-21.w90-59.abo.wanadoo.fr) |
| 19:37.08 | seanbright | maybe the files are on an NFS export that is on an SMB share that is exposed by amazon S3 |
| 20:48.06 | *** join/#asterisk-dev spatel (~spatel@static-71-174-102-211.bstnma.fios.verizon.net) |
| 21:06.25 | *** join/#asterisk-dev mvanbaak (~mvanbaak@asterisk/contributor-and-bug-marshal/mvanbaak) |
| 21:06.25 | *** mode/#asterisk-dev [+o mvanbaak] by ChanServ |
| 21:08.51 | *** join/#asterisk-dev tzafrir (~tzafrir@lemon.tzafrir.org.il) |
| 21:08.51 | *** mode/#asterisk-dev [+o tzafrir] by ChanServ |
| 21:19.10 | *** join/#asterisk-dev wdoekes (~walter@wjd.osso.nl) |
| 21:19.10 | *** mode/#asterisk-dev [+o wdoekes] by ChanServ |
| 21:36.57 | seanbright | don't try to load 25k items from astdb into the media cache on startup |
| 21:37.01 | seanbright | you'll have a bad time |
| 21:43.29 | kharwell | heh I'll take your word for it |
| 21:48.01 | snuff-work | elguero: cheers.. that sounds EXTACTLY like my problem |
| 21:48.08 | snuff-work | going to try that setting |
| 22:02.08 | seanbright | so |
| 22:02.22 | seanbright | how do we feel about creating additional tables in astdb.sqlite? |
| 22:03.00 | seanbright | leave the astdb table and the astdb API in place |
| 22:03.38 | seanbright | but - as an example - create a media_cache table that has appropriate columns, data types, etc. |
| 22:05.17 | seanbright | fwiw, a Playback of a non-URI filename happens instantly for me with 25k items in the media_cache |
| 22:05.44 | seanbright | granted my dev VM is not under any load |
| 22:05.51 | seanbright | but it doesn't even stutter |
| 22:22.40 | seanbright | is hoping that "load average hovering around 20-30" means "20-30% CPU and not 1 minute load average of 20-30" |
| 22:32.16 | kharwell | would it be optional to use astdb for that? |
| 22:32.40 | seanbright | i think it would populate both |
| 22:33.05 | seanbright | insert into media_cache table in nice way, then insert into astdb in crappy way |
| 22:33.34 | seanbright | assuming we care about backwards compat |
| 22:34.25 | kharwell | I'm not too familiar with the media_cache actually. don't think I've really looked at the code. |
| 22:34.54 | kharwell | glancing at it appears to already use the astdb? Or do you mean redo the tabling structure and layout in a different way? |
| 22:37.26 | seanbright | it does use astdb currently which makes it's performance pretty bad on startup |
| 22:37.41 | seanbright | its* |
| 22:37.58 | seanbright | so i am suggesting we create a new table in the sqlite3 database and use that instead of astdb |
| 22:39.52 | kharwell | aah I think I understand now. In my head I usually just translate astdb as sqlite3 as it's backed by that, but it's the actual singular monolithic table. Yeah I'd be fine having it broken out |
| 22:40.10 | seanbright | ok. awesome. |
| 22:40.36 | seanbright | will try to get additional buy-in from the other key decision makers |
| 22:40.44 | seanbright | synergize a new paradigm |
| 22:40.57 | kharwell | makes sense |
| 22:41.28 | kharwell | I guess the media index cache is something completely different? https://blogs.asterisk.org/2019/01/30/rip-media-index-cache/ |
| 22:41.38 | seanbright | it is |
| 22:42.07 | seanbright | the media index is an index of stuff in /var/lib/asterisk/sounds |
| 22:42.24 | kharwell | ah |
| 22:42.32 | seanbright | the media cache is like a browser cache for media that comes over http |
| 22:43.17 | kharwell | does it store the actual audio in the db then? |
| 22:43.24 | seanbright | no, just metadata |
| 22:43.26 | kharwell | or on the filesystem and use links |
| 22:43.28 | kharwell | aah okay |
| 22:43.33 | seanbright | but in pretty much the worst way possible |
| 22:43.43 | kharwell | hard to believe |
| 22:43.46 | kharwell | :-D |
| 22:44.02 | seanbright | so there is a row in the table for each item, which looks like this: |
| 22:44.32 | seanbright | /MediaCache/http://seanbright.com/some-file.ulaw => <sha1 hash of URI> |
| 22:44.42 | seanbright | and then: |
| 22:45.00 | seanbright | /<sha1 hash of URI>/path => /tmp/foo-bar.ulaw |
| 22:45.13 | seanbright | /<sha1 hash of URI>/content-type => audio/pcmu |
| 22:45.15 | seanbright | etc |
| 22:45.30 | seanbright | so if you have 25k rows (which is ridiculous and no one should) |
| 22:45.50 | seanbright | it queries all of the /MediaCache/* rows, and then for each one, queries the /<sha1 hash of URI> rows |
| 22:46.06 | kharwell | yeah ouch. can see how that might take time |
| 22:48.27 | seanbright | and the abstraction provided by db.h is crappy |
| 22:48.45 | seanbright | "i want a list of all things that start with X" |
| 22:49.05 | seanbright | "ok, i will run a query for all things that start with X, iterate over the results, building a linked list, and then return the linked list to you" |
| 22:49.22 | jsmith | couldn't agree more :-) |
| 22:49.31 | seanbright | "great, once i get that back from you, i will iterate over that linked list and do something with it" |
| 22:49.38 | seanbright | "then we'll just throw the whole damn thing away" |
| 22:50.26 | seanbright | anyway, i'm babbling |
| 22:59.11 | coreyfarrell | is it possible to configure a localhost media proxy and disable media cache (leave it to the proxy)? |
| 23:04.36 | seanbright | that's a good question |
| 23:05.16 | seanbright | res_http_media_cache would have to let you configure that somehow i assume |
| 23:05.37 | seanbright | unless you did some magic on your router |