Friday, 2013-12-13

*** tpb has joined #timvideos00:00
skayI can't remember exactly what I do... I think I might log in, start the ttycast in another session, then start up tmux00:00
mithro-workskay, no, I mean that every terminal in tmux was automagically broadcast00:00
skayso cool00:00
skayoh oh that would be neat00:00
skayI bet someone has done that00:00
skaybtw, have you seen multitail?00:00
skayI am not sure I like it more than having multiple pains open with tailing logs, but it is kind of cool00:01
mithro-workskay, yeah but I always forget it exists00:01
iiiefor Carl when he comes on ec2-54-196-24-89.compute-1.amazonaws.com00:01
iiieskay "ssh -t [email protected] tmux attach"00:04
skayspiffy!00:06
skayhopefully my big term is not annoying00:06
skaydots everywhere00:06
iiiedownsizes to smallest viewer00:06
iiie^b n for next ^b p for previous ^b c for create00:07
skayoh that will be confusing. I've rebound that to ^a00:08
skaybecause of screen00:08
iiiewith b you can have screen or emacs inside of tmux00:08
iiie:)00:08
skayI do ^a ^a00:08
iiieI'm on 1:bash now00:08
iiiefair enough00:09
skayI think it followed you00:09
iiiehe he he00:09
iiieyeah00:09
skayI rotated around anyway00:09
skayhmm, if i can remember ^b it would probably be better to type than ^a00:09
skaydo you know about ^r00:10
skay^r and then start typing a search term00:10
iiieSometimes :)00:11
skayI would never remember everything in the loop. I'd make an alias00:11
skaymy brain is fried00:11
iiietimvideos.us/chipy will be up now00:11
iiieblank of course00:12
skaydo you have any cool videos you can send to it? I don't00:12
iiieoh and ^b d if you want to disconnect00:12
iiiehm, I should00:13
skayI detached00:13
iiieI saw, the dots resized :D00:13
iiiehm, I think my desktop probably is still a configured collector00:14
skaystream the chipy irc00:14
mithro-workskay, you can configure tmux to use ^a of course :)00:15
skayyeah, but that would not be very nice to iiie00:16
skaywho is expecting ^b to work00:16
iiieI'd adapt00:16
skaydo you guys do00:16
skaybind | split-window -h00:16
skaybind - split-window -v00:16
skayI really like that one00:16
iiieI've done it, but mostly I don't need to monitor that many different things (cssh for homogeneous)00:18
mithro-workso the settings for tmux are per connection not per server00:24
mithro-workso you can have one tmux connected with ^b and one tmux connected with ^a to the same tmux session00:24
lukejohnCarlFK: ahh i didn't think of that, i've readded the emails with an auth check on them, it looks like urllib2 which is currently being used to get the schedule handles basic authentication http://docs.python.org/2/howto/urllib2.html quite well.00:38
tpbTitle: HOWTO Fetch Internet Resources Using urllib2 Python v2.7.6 documentation (at docs.python.org)00:38
mithro-worklukejohn, you want the json feed to be accessible without authentication00:39
lukejohnmithro-work, so the Json feed is still accessible it just won't show emails without auth. If this is not workable how about I give you speaker ID's in that feed and provide speakers emails to you separately. I really dislike having those emails public.00:43
mithro-worklukejohn, no thats okay00:43
*** CarlFK has joined #timvideos00:46
iiiewelcome CarlFK00:47
CarlFKhi!~00:47
CarlFKsec.. let me plug mouse into trist00:47
iiieec2-54-196-24-89.compute-1.amazonaws.com is todays magic number00:47
CarlFKwhats the xml tagy thing it goes in?00:51
CarlFKvim  /usr/local/etc/flumotion/managers/default/planet.xml00:51
CarlFKsearch for ?00:51
CarlFKand is tim around?00:51
CarlFK      <property name="host">54.205.122.37</property>00:52
CarlFKthat.00:52
iiiesounds, right checking..00:53
iiielooks right too me00:53
iiieupdate to <property name="host">ec2-54-196-24-89.compute-1.amazonaws.com</property>00:54
iiieor the IP <property name="host">54.196.24.89</property>00:54
CarlFK(veyepar)[email protected]:~$ sudo service flumotion restart * Restarting Flumotion Streaming Server flumotion                       [ OK ]00:57
CarlFKstream up?00:57
iiieencoder says hungry00:58
iiieCarlFK: encoder says hungry01:01
CarlFKrestarted01:05
CarlFKproducer-audio-video: happy ....  all happy01:05
iiiewaking (restarted)01:05
iiiehm01:06
iiiehungry again01:06
* iiie goes and looks at the IP again01:06
CarlFK      <property name="host">c2-54-196-24-89.compute-1.amazonaws.com</property>01:06
iiiee201:06
iiienot c201:06
CarlFKhow bout  both :)01:07
skayec2-54-196-24-89.compute-1.amazonaws.com01:07
iiie...01:07
iiieec201:07
iiiewhat she said01:07
iiiesigh01:07
iiiehappy!01:07
* iiie has video!01:08
lukejohnso as far as I can see in the streaming system the only files that need to be updated to get it working with a new events data are {"streaming-system":  {"production": ["config.json", "website/frontend/externalschedule2internal.py"], "testing": ["tools/register/fake_register.py#L26"]}}01:31
lukejohnso website/[eventfeed]2internal.py is adding a conf_url and guid, i've grepped and grepped and found no use of these attributes. are they required?02:14
CarlFKlukejohn: github url?02:15
CarlFKI might recognize it02:15
lukejohnCalrFK, https://github.com/timvideos/streaming-system/blob/master/website/pycon2internal.py#L14202:17
tpbTitle: streaming-system/website/pycon2internal.py at master · timvideos/streaming-system · GitHub (at github.com)02:17
CarlFKlukejohn: my guess is it is a unique id .. and more guess is that it has to do with updating some down stream datastore02:34
mithro-worklukejohn, fake_register.py is only used for testing02:41
mithro-worklukejohn, did you check the javascript / templates?02:42
lukejohnmithro-work, yep, i grepped the frontend directory and got no matches for "guid"02:45
mithro-workoh02:46
lukejohnmithro-work, whilst av fake_register.py is only used for testing if you change the testing rooms id in the config.json you need to update the room_id in fake_register.py. probably not something that should ever happen anyway.02:47
mithro-worklukejohn, oh02:47
mithro-worklukejohn, the system is designed to have a config.json and config.json.private02:47
mithro-workyou can do a python config.py to look at the merged file02:48
mithro-workso ideally config.json should only have testing / example rooms in it and config.json.private should have the actual rooms02:49
lukejohnokay, that makes sense, i'll add that to my documentation branch.02:50
lukejohnthanks for clarifying02:51
mithro-workthe django app does the same thing02:52
mithro-workif you want to set it up locally, copy the website/private/settings.py.example to website/private/settings.py02:53
lukejohnyep, got it running locally thanks02:53
lukejohnI just didn't notice what was going on with the config.json, sorry for my confusion.02:54
mithro-worklukejohn, nah it's fine as we have pycon stuff all through config.json02:54
CarlFKmithro-work: if scp says: dc10bc.dv                                       0%   36MB   1.8MB/s 1:55:47 ETA03:03
CarlFKis 1.8MB/s under what is needed for a stream?03:03
CarlFKhuh, it went up to 2.7 now03:04
mithro-workI never know if scp is displaying MegaBytes or MegaBits03:04
mithro-workyou need ~1.5Mbits03:04
CarlFKany idea how I can test the bandwidth?03:05
CarlFKencoder kept dropping, like every min or so.  which seems consistent with the bandwidth floating between 1.7 and 2.503:06
mithro-workencoder kept dropping - please tell me what you actually mean03:07
mithro-workthe encoder kept going hungry?03:07
CarlFKiiie: ping03:07
CarlFKiiie was monitering that part03:09
iiiepong03:09
iiieyes the encoder kept going hungry03:09
mithro-workiiie, which part of the pipeline?03:10
iiieI think audio first, but I was mostly looking to see if it was hungry.  The watchdog seemed to hit it even if some of the pipeline was working (I suspect that's desirable)03:11
mithro-workIt would be good to know03:12
iiieI've exited out, I don't think those make it into the log, they might.03:13
iiieworker log says not connected some, and so does manager log, but I think that's from stopping the service or the restart from the watchdog03:17
CarlFKwant me to start the collector ?03:17
iiiesure03:18
CarlFKscp is reproting  11.0MB/s03:21
CarlFK(veyepar)[email protected]:~/Videos/2013-06-03$03:22
iiiewoah, the watch dog is having some sort of problem03:22
CarlFKam am still flooding the connection with that scp03:22
iiie..., and the disk is full03:22
CarlFKdoh03:22
iiieat least it's the first thing I look at no...03:23
iiie*now03:23
CarlFKI just killed the scp03:23
mithro-workthe server running out of disk space from logs is a problem I have encountered before03:23
iiiethis time it was CarlFK uploading a 3.7+ GB file03:24
iiieI've deleted it03:24
iiieall happy, to all hungry03:25
iiieI couldn't tell faster than the update, as the lines updated they became hungry03:25
iiieall happy again03:26
CarlFKjust talk to the python dev that works here (so not a Network guy) "we just upgraded to a bigger fiber.. I would expect everything to be fine"03:27
iiieCarlFK: for future reference try scp [email protected]:/dev/null03:27
CarlFKiiie: I was wondering about that03:27
iiiecome to think of it scp /dev/zero [email protected]:/dev/null03:28
iiiejust make sure compression is off03:28
iiieall are happy and I'm watching the stream03:28
iiiewhich I'm not clear on how this is working...03:28
CarlFK?03:29
iiiesince I don't have register running03:29
iiieI've reloaded the webpage, I would expect it to lose the registration of the encoder03:29
iiiethat or we don't have to run the register in a loop03:29
iiiesound is cutting out but the admin on encoder still says happy03:30
iiievideo has gone to hungry03:30
iiieand so has audio03:30
iiiestill unclear that it's not just the refresh through the admin03:30
iiienow all are hungry, and the watchdog is restarting the service03:31
iiieall happy again03:31
iiiewe're at the funny edge03:31
iiieall hungry, restart, wake, happy,03:32
CarlFKI am going to try the scp.. see what it says03:32
iiiewebm is hungry first03:32
iiiehigh and low03:32
iiieflv high too03:32
CarlFK[email protected]:~/Videos/roadtrips$ scp /dev/zero [email protected]:/dev/null03:33
CarlFK/dev/zero: not a regular file03:33
CarlFKstarted at 3, dropping ...03:34
iiieproducer decoder was happy even after src isn't03:34
CarlFKjust dropped bellow 203:34
CarlFKnow climbing03:34
iiiewebm seems to stay hungry and flv03:34
iiieflv seem to not make it back to happy03:34
iiiemithro-work: is any of this helpful?03:35
iiiethey're all hungry03:36
iiieexcept for the porter, which doesn't count03:36
CarlFKdc10bc.dv                                       4%  581MB   6.4MB/s   30:29 ETA03:36
mithro-workiiie, there are some buffers in there03:37
mithro-workthe flv thing tends to crash on buffers going missing03:37
CarlFKiiie: do you have a collecotor VM or something?03:38
iiieCarlFK: I have a collector on my desktop (we never uninstalled it) but IIRC we used a collector on VM that's gone away.03:39
iiieWhile I tried to bring it up before you came online at the event I did stop the service before trist started connecting.03:40
CarlFKIll stop trist, you bring yours up and see if the encoder stays happy?03:40
iiieha03:40
iiiemine hasn't worked yet03:40
iiieworth a try though03:41
CarlFKgetting kicked ou03:41
CarlFKgetting kicked out03:42
iiieyeah, I'll see if I can do something from here03:42
CarlFKdropping connection here03:59
CarlFKsee ya03:59
iiiebye03:59
*** CarlFK has quit IRC04:04
iiieI'm calling it a night, I can't get my local to run by itself, nevermind talking to mission04:18
iiieI'm taking down mission too04:18
iiiemission reports off in the aws api04:19
iiiefound an article on flumotion streaming at a conference http://newstar.stage.typethink.com/blog/04:20
iiie...fail04:20
iiiehttp://wingolog.org/archives/2006/07/07/so-you-want-to-stream-a-conference04:20
tpb<http://ln-s.net/:zbu> (at wingolog.org)04:20
*** tvCommitBot has joined #timvideos04:21
tvCommitBot[streaming-system] lukejohnosmahi opened pull request #49: Removed an assert from config.py (master...master) http://git.io/ncK_sw04:21
*** tvCommitBot has left #timvideos04:21
*** iiie has quit IRC04:21
*** CarlFK has joined #timvideos05:08
*** CarlFK has quit IRC05:09
*** CarlFK has joined #timvideos05:51
*** CarlFK has quit IRC05:52
*** CarlFK has joined #timvideos06:29
*** CarlFK has quit IRC06:43
*** tvCommitBot has joined #timvideos06:44
tvCommitBot[streaming-system] lukejohnosmahi opened pull request #50: Added config.private.json.example (master...add/master/config.private.json.example) http://git.io/VidQ_A06:44
*** tvCommitBot has left #timvideos06:44
*** 17WAAKRIC has joined #timvideos07:01
17WAAKRIC[streaming-system] mithro closed pull request #49: Removed an assert from config.py (master...master) http://git.io/ncK_sw07:01
*** 17WAAKRIC has left #timvideos07:01
*** 65MAAEJBP has joined #timvideos07:01
65MAAEJBP[streaming-system] mithro pushed 3 new commits to master: http://git.io/-f2L8A07:01
65MAAEJBPstreaming-system/master 39fc0d4 Luke John: Removed assert from config.py which was erroring if config.json did not match config.private.json07:01
65MAAEJBPstreaming-system/master 332089f Luke John: Removed leftover function used by an assert which was removed by commit 39fc0d4eb581c3d9e271c45e77738ede97a8fa8f07:01
65MAAEJBPstreaming-system/master cb4e834 Tim Ansell: Merge pull request #49 from lukejohnosmahi/master...07:01
*** 65MAAEJBP has left #timvideos07:01
*** mithro has joined #timvideos12:12
*** hyades has joined #timvideos13:56
*** hyades has quit IRC15:14
*** skay has quit IRC16:34
*** skay has joined #timvideos16:34
*** CarlFK has joined #timvideos18:18
*** iiie has joined #timvideos19:53

Generated by irclog2html.py 2.12.1 by Marius Gedminas - find it at mg.pov.lt!