[
https://issues.apache.org/jira/browse/OPENMEETINGS-928?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14087929#comment-14087929
]
Maxim Solodovnik commented on OPENMEETINGS-928:
-----------------------------------------------
Is this issue reproducible using latest 3.0.3 builds (we going to release it in
a week)
https://builds.apache.org/view/M-R/view/OpenMeetings/job/Openmeetings%203.0.x/
> audio mesh not establishing some connections
> --------------------------------------------
>
> Key: OPENMEETINGS-928
> URL: https://issues.apache.org/jira/browse/OPENMEETINGS-928
> Project: Openmeetings
> Issue Type: Bug
> Components: Audio/Video
> Affects Versions: 2.2.0 Apache Release
> Environment: Server is running: Windows 7 x64, openmeetings 2.2.0
> Client: not sure about the others but I was running chrome (I can find out
> what browsers others were using if that may be relevent)
> Reporter: Mike Sop
> Priority: Critical
> Attachments: openmeetings.log, red5.log
>
>
> Not sure all of this info is necessary but figure it cant hurt
> I run openmeetings from windows via the "red5.bat" (not sure if you need this
> run with the red5-debug.bat, but hte logs seem to be pretty verbose so
> hopefully they can point ot the problem)
> I sent a conference invite with max 5 users, 4 of us joined and had a wierd
> mesh of connections i.e. user 1 was able to hear user 2 and 3, but only user
> 3 could hear him. User 2 could only hear user 1 and 4, user 4 could only
> hear user 2, etc. On retry we had a different broken mesh of connections
> (wasnt consistently same users able to hear same users)
> Is the audio a peer to peer connection, or does the server act as a
> "reboradcaster"? Either way I suspect the issue is with establishing those
> initial connections, since if it were a bandwidth issue I wouldnt expect the
> disconnect to persist for the entire meeting (i.e. the same users can hear
> same subset of users through duration of the meeting) however like I
> mentioned, on a 2nd try (after server restart, rejoin a room) we had a
> different ordering of disconnects, but still not a perfect mesh of
> connections (seems like a 50/50 chance of connection (each direction) being
> preoperly established, since we could all speak to about half the people, and
> hear from about half the people, but that set of people you can speak to is
> not the same as those you can hear (so the "to" and "from" connections are
> also distinct I'm guessing)
> Even with the audio problems, we were all able to use the group chat, and
> were all able to see each otehrs webcams, just seemed to be a problem with
> the audio initilizing. I've attached my logs in case it helps, unforutnately
> I dont recall start/end times of attempts, but the server was started/stopped
> shortly before/after our meeting attempt (I tried a reboot of the server to
> see if it would fix the issue but no luck)
> Also FWIW I did a testrun with 2 users a day before our group meeting which
> was fine, we both joined a public room from the dashboard (not a meeting via
> created conference i.e. from link in conference invite email) so not sure if
> that could be related, or if is just an issue with > 2 users (I suspect the
> latter)...
> Anyway let me know if you need more info (if this isnt easy to reproduce on
> your end) Seems like would be a pretty obvious bug for other users but not
> seeing it listed (sorry if this is a dup)
--
This message was sent by Atlassian JIRA
(v6.2#6252)
