| *** tpb has joined #timvideos | 00:00 | |
| tumbleweed | thaytan: would it be reasonable to insert a debugging server in the pipeline somewhere? (I don't know if gstreamer has such a thing) | 00:01 |
|---|---|---|
| thaytan | tumbleweed, what's a debugging server? | 00:02 |
| tumbleweed | i.e. an rtsp server that you can connect to, but doesn't hold things up if there are no clients attached | 00:02 |
| thaytan | oic | 00:02 |
| thaytan | maybe, with a bit of work | 00:02 |
| tumbleweed | btw, x264parse didn't seem to make youtube much happier, but it's what we're doing | 00:03 |
| thaytan | h264parse, you mean | 00:04 |
| thaytan | tumbleweed, also interesting to see some of that h264 output then, and see if it's really only doing keyframes every 15 seconds | 00:05 |
| tumbleweed | that | 00:05 |
| tumbleweed | that's easier to get a test of | 00:05 |
| tumbleweed | oh, if we can reproduce the desync problems in the test room, then I'm sure we can add all sorts of debugging in there | 00:05 |
| thaytan | yes | 00:06 |
| thaytan | but aiui, it's not predictable | 00:06 |
| xfxf | tumbleweed: I tried LACP before with vocto and had some wierd issues, I forget what | 00:21 |
| xfxf | i'll have to reproduce sometime | 00:21 |
| xfxf | jea: scrolling back... is A/V sync still off? that's super difficult to fix in post - we should be restarting voctocore :/ | 00:22 |
| xfxf | okay, a bunch of rooms are WAY off | 00:23 |
| xfxf | jea: are you monitoring? who is monitoring? | 00:23 |
| jea | so when we restarted it yesterday, we noticed no change | 00:32 |
| CarlFK | thaytan: is there an audio test source that beeps every second? based on wall time... :D | 00:33 |
| jea | and rooms reported it was fine in front of them, just me noticing things wrong on the stream. makes me wonder if it is to do with the vaapi pipeline we are using | 00:33 |
| xfxf | so i just went to all rooms | 00:33 |
| xfxf | and sync seems fine in vocto | 00:33 |
| xfxf | so i think it's isolated to streaming | 00:33 |
| thaytan | CarlFK, of course not :) | 00:33 |
| CarlFK | thaytan: patch coming. in a few... weeks. | 00:34 |
| mithro | CarlFK: talking about slots in Tasman B/C :-P | 00:58 |
| CarlFK | mithro: yay slots! | 00:59 |
| thaytan | o/~ Slot to the heart, and you're to blame... o/~ | 01:02 |
| tumbleweed | anyone know how /srv/gw1/video/lca/lca2017/dv/spare/2017-01-19/ keeps appearing? it's not me | 01:03 |
| * tumbleweed deletes it again | 01:04 | |
| tumbleweed | CarlFK: there seems to be a race in enc.py. I've seen the same video picked up by two nodes, twice in 5 mins | 01:05 |
| CarlFK | tumbleweed: I just noticed that. make that 3 encXs | 01:05 |
| tumbleweed | it should probably check that it got the lock | 01:06 |
| CarlFK | tumbleweed: it might be because I reset 12114 to Encode, hit save, then tweeked it, hit save again, talked to ryan, couldn't remember if I had hit save, hit save again | 01:06 |
| * tumbleweed goes looking for the locking code | 01:07 | |
| tumbleweed | enc2 and 3 had that one | 01:07 |
| CarlFK | im not sure how I can prevent me from stepping on the locks | 01:07 |
| tumbleweed | yeah | 01:07 |
| tumbleweed | kill the old encode when you want to step on one? | 01:08 |
| tumbleweed | or don't let you step on one when its locked | 01:08 |
| CarlFK | then how to do I ever unlock one when the process that locked it dies? | 01:08 |
| tumbleweed | admin / --unlock | 01:08 |
| CarlFK | hmm... you may be onto something... | 01:09 |
| tumbleweed | systemd is currently encoding on two nodes | 01:10 |
| tumbleweed | I'll kill the enc6 one | 01:11 |
| CarlFK | systemd is encoding? | 01:11 |
| tumbleweed | 12114 | 01:11 |
| CarlFK | oh.. that systemd | 01:12 |
| jea | how many systemds do you want? | 01:12 |
| CarlFK | I thought you had made enc.py a systemd thing | 01:12 |
| tumbleweed | :P | 01:13 |
| tumbleweed | systemd and veyepar both of tentacles | 01:13 |
| tumbleweed | we'll see which one wins | 01:13 |
| tumbleweed | s/of/have/ | 01:13 |
| tumbleweed | CarlFK: time to run a make_public loop? | 01:17 |
| CarlFK | probably.. | 01:18 |
| CarlFK | poking my home Internet right now... | 01:18 |
| tumbleweed | CarlFK: aah, spare keeps coming back because of a weird symlink on the box | 01:20 |
| tumbleweed | I'm guessing that was you | 01:20 |
| jea | you run the make public script from your home computer? | 01:20 |
| tumbleweed | lrwxrwxrwx 1 videoteam videoteam 29 Jan 16 23:01 /srv/video/lca/lca2017/dv -> /srv/gw1/video/lca/lca2017/dv | 01:20 |
| CarlFK | hmm, possible ;) | 01:20 |
| * tumbleweed fixes it | 01:21 | |
| xfxf | copying from gw1 > enc6 is taking forever | 01:22 |
| xfxf | rsync has been going since yesterday evening :/ | 01:22 |
| xfxf | ~23MB/s | 01:22 |
| tumbleweed | should be a lot faster than that | 01:23 |
| tumbleweed | there's a 10gbps link between the two | 01:23 |
| xfxf | yeah, something's up | 01:23 |
| xfxf | maybe we're starving disk I/O on the array? | 01:23 |
| xfxf | jea: i don't ask CarlFK questions... veyepar things just happen :P | 01:24 |
| *** tpb has joined #timvideos | 01:28 | |
| mithro | jea: So the bot hasn't been announcing the schedule on the IRC channels | 01:30 |
| mithro | jea: I forgot to enable the RSS announcement module... | 01:30 |
| mithro | jea: The formatting of the front page has gone weird | 01:32 |
| mithro | jea: And the talks are not "moving forward" | 01:32 |
| jea | mithro: I didn't know there was an announcement module. I can look into that | 02:35 |
| mithro | jea: The bot is just suppose to read the rss feed and announce them | 02:35 |
| jea | front page is different because I changed from using a table to bootstrap columns, so it will be responsive | 02:36 |
| tumbleweed | xfxf: err, you're using -z with rsync, try without that | 02:40 |
| xfxf | oohhh | 02:41 |
| xfxf | oops, trigger finger | 02:41 |
| xfxf | that'll be bottlenecking CPU on one of this macjines | 02:41 |
| * tumbleweed noticed it was pegging a CPU core or gw1 | 02:41 | |
| xfxf | Good spot | 02:41 |
| xfxf | Will fix | 02:41 |
| xfxf | ~95MB/sec now | 02:43 |
| xfxf | slightly better :P | 02:44 |
| tumbleweed | :) | 02:44 |
| xfxf | iirc there's an alt ssh crypto which is computationally less expensive (but cryptographically less secure) which is good for this | 02:45 |
| tumbleweed | looks like you are I/O bound now | 02:45 |
| xfxf | but i suspect the bottleneck is now the I/O | 02:45 |
| xfxf | indeed | 02:45 |
| tumbleweed | so, that won't make a huge difference | 02:45 |
| xfxf | ack | 02:45 |
| *** Sewar_ has joined #timvideos | 02:48 | |
| *** hyadez has quit IRC | 02:49 | |
| *** Sewar has quit IRC | 02:49 | |
| *** Sewar_ is now known as Sewar | 02:49 | |
| *** CARAM__ has quit IRC | 02:52 | |
| *** cr1901_modern has quit IRC | 02:52 | |
| *** CARAM__ has joined #timvideos | 02:53 | |
| *** cr1901_modern has joined #timvideos | 02:54 | |
| *** hyadez has joined #timvideos | 03:11 | |
| mithro | CarlFK: This is a reminder that you want to test the new firmware on the Atlys at the conference otherwise it might never happen | 03:11 |
| CarlFK | mithro: thanks | 03:27 |
| xfxf | tumbleweed: 1,596,555,264 27% 170.84MB/s 0:00:23 | 03:57 |
| xfxf | this rsync is too slow :P | 03:57 |
| *** karan has joined #timvideos | 04:41 | |
| karan | hello world | 04:43 |
| CarlFK | /nick World Hello karan! | 04:44 |
| * CarlFK rolls off | 04:45 | |
| tumbleweed | CarlFK: don't forget to grab the sync_rax patch | 04:47 |
| *** tvCommitBot has joined #timvideos | 04:58 | |
| -tvCommitBot- [lca2017-av] joeladdison pushed 1 new commit to master: https://git.io/vMMiX | 04:58 | |
| -tvCommitBot- lca2017-av/master 0d8aa28 Joel Addison: Add prometheus node_exporter for monitoring | 04:58 | |
| *** tvCommitBot has left #timvideos | 04:58 | |
| mithro | jea: I'm guessing you don't have a login to logs.timvideos.us? | 05:01 |
| jea | mithro: unless it is the same thing, no | 05:01 |
| mithro | jea: Adding you now | 05:05 |
| mithro | jea: Should be able to login to [email protected]? | 05:30 |
| jea | mithro: yep, that worked for me | 05:31 |
| xfxf | shenki: i say with some confidence if you bricked the unit in your car, mazda would probably fix it for free as a 'warranty job' | 05:32 |
| xfxf | dealers updating the cars computer isn't new (ECU updates during services have been a thing for at least a decade) | 05:32 |
| xfxf | the dealer isn't going to know anything about the unit apart from if it doesn't work they probably have some tool to make it work again (which probably just reflashes it) | 05:33 |
| xfxf | just tell them it stopped working and you don't know why :P | 05:33 |
| *** karan has quit IRC | 05:34 | |
| *** karan has joined #timvideos | 05:44 | |
| karan | mithro: can you suggest me some bugs based on verilog that don't require fpga | 05:45 |
| *** thaytan has quit IRC | 05:47 | |
| *** tvCommitBot has joined #timvideos | 05:49 | |
| -tvCommitBot- [lca2017-av] stefanor pushed 2 new commits to master: https://git.io/vMM1U | 05:49 | |
| -tvCommitBot- lca2017-av/master 9308dfc Stefano Rivera: Avoid repeatedly unpacking fonts | 05:49 | |
| -tvCommitBot- lca2017-av/master a24e2d7 Stefano Rivera: Get the monitoring role up | 05:49 | |
| *** tvCommitBot has left #timvideos | 05:49 | |
| *** karan has quit IRC | 05:52 | |
| *** CarlFK has quit IRC | 05:54 | |
| *** karan has joined #timvideos | 05:58 | |
| *** karan has left #timvideos | 05:58 | |
| jea | tumbleweed: https://github.com/docker/compose/issues/3525 | 05:58 |
| tpb | Title: Missing dependency on Ubuntu 16.04 · Issue #3525 · docker/compose · GitHub (at github.com) | 05:58 |
| *** CarlFK has joined #timvideos | 06:08 | |
| *** ChanServ sets mode: +v CarlFK | 06:08 | |
| *** karan_ has joined #timvideos | 06:14 | |
| karan_ | mithro:ping | 06:42 |
| *** karan_ has quit IRC | 07:02 | |
| *** cr1901_modern has quit IRC | 07:20 | |
| *** cr1901_modern has joined #timvideos | 07:40 | |
| *** CarlFK has quit IRC | 08:29 | |
| *** aps has joined #timvideos | 08:49 | |
| *** cr1901_modern has quit IRC | 10:18 | |
| *** cr1901_modern has joined #timvideos | 10:24 | |
| *** CarlFK has joined #timvideos | 11:38 | |
| *** ChanServ sets mode: +v CarlFK | 11:38 | |
| xfxf | http://veyepar.nextdayvideo.com/main/C/lca/S/lca2017/s/8/E/12047/ | 13:14 |
| tpb | Title: veyepar: Advances in CPU Frequency Management (at veyepar.nextdayvideo.com) | 13:14 |
| *** rohitksingh has joined #timvideos | 13:35 | |
| *** aps has quit IRC | 15:46 | |
| *** sb0 has quit IRC | 16:19 | |
| *** rohitksingh has quit IRC | 17:22 | |
| *** hyades has joined #timvideos | 19:57 | |
| *** nueces has joined #timvideos | 20:04 | |
| CarlFK | up | 20:35 |
| *** CarlFK has quit IRC | 21:00 | |
| *** CarlFK has joined #timvideos | 21:57 | |
| *** ChanServ sets mode: +v CarlFK | 21:57 | |
| *** hyades has quit IRC | 22:01 | |
| * tumbleweed must have accidentally turned his alarm off, yesterday | 22:19 | |
| tumbleweed | althoguh I had been waking up before it, but that's how it goes | 22:19 |
| CarlFK | no worries, things are going pretty good. | 22:24 |
| * CarlFK eats a cookie | 22:24 | |
| * tumbleweed presumed as much | 22:24 | |
| xfxf | OMG EVERYTHING IS BLOWING UP | 22:32 |
| xfxf | all good | 22:32 |
| xfxf | we only need you if systemd breaks. systemd never breaks, right? :P | 22:32 |
| xfxf | (we're fine, things going wel) | 22:33 |
| *** tpb has joined #timvideos | 22:37 | |
| *** tvCommitBot has joined #timvideos | 23:10 | |
| -tvCommitBot- [lca2017-av] stefanor pushed 1 new commit to master: https://git.io/vMypj | 23:10 | |
| -tvCommitBot- lca2017-av/master 611bbbf Stefano Rivera: Exclude outside hours videos | 23:10 | |
| *** tvCommitBot has left #timvideos | 23:10 | |
| *** nueces has quit IRC | 23:11 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!