No log entries for the failures. Internet connectivity has been stable. The subscriber Flash movies stay connected and I can send other content to them without any restart on the server. I can also connect new subscribers to the session after the published stream has stopped. It appears that the published stream is lost, but no error or info is reported in the log. Any ideas anyone?
The good news is that I have not seen the Video/audio sync problems that started this thread. Video and audio have stayed in sync throughout all testing. Interalab wrote: > Failed again with same symptoms 45 minutes later. I'll look through > the log for any clues. > > Interalab wrote: > >> This is the first result of our live stream testing with the trunk - >> the svn 0.6 dev trunk is the same as 0.6RC2 isn't it?: >> >> Server: >> AMD64 3000+ >> Fedora Core 4 Linux >> 2 GIG RAM >> 1 GIG Internet connection in data center >> One live publisher and one subscriber >> Both publisher and subscriber are remote (on a cable modem) >> >> Publisher: >> 1 GIG Processor >> 512 Meg RAM >> Windows 2000 >> >> 16 fps >> 80% quality >> 320x240 >> audio 22 khz >> overall bandwidth 370 kbps >> >> Initial Stats: >> 465 Meg RAM in use >> 0 Swap in use >> virtually no CPU load >> Stream delay < 2 seconds >> >> After approximately 2 hours the stream froze and the upload and >> download bandwidth went down to 0. >> There was no significant change in RAM or CPU stats. >> At the time of freezing, the live stream had gradually built up an >> overall delay of 10 seconds. Both audio and video were still in sync >> with each other. >> We noticed a few minor hiccups during the 2 hours of successful >> streaming where the video and audio would freeze for a second and then >> come back. >> >> Restarted the publisher Flash movie. The subscriber started working >> without any intervention once the publisher was back from it's restart. >> More later . . . >> >> >>> Interalab wrote: >>> >>> >>>> Lately, we've been using a dynamic buffering scheme that can go from >>>> 2 seconds up to 10 seconds if needed. We haven't tried live >>>> streaming since changing to dynamic - we'll give it a try and report >>>> back. Thanks! >>>> >>>> Rodrigo Ordonez Licona wrote: >>>> >>>> >>>>> WE reduced the bufering time and achieved better results on a live >>>>> video >>>>> stream. >>>>> >>>>> This seems so develop with high latency connections , >>>>> >>>>> 2 -6 second buffer solved the problem for us >>>>> >>>>> And not sure if RC2 actually behaved the same >>>>> >>>>> Will try. >>>>> >>>>> HTH >>>>> >>>>> Rodrigo O >>>>> Xnet >>>>> -----Original Message----- >>>>> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On >>>>> Behalf >>>>> Of Interalab >>>>> Sent: Sábado, 17 de Febrero de 2007 11:43 a.m. >>>>> To: [email protected] >>>>> Subject: Re: [Red5] Video delays, 0.6RC2 doesn't compile >>>>> >>>>> I haven't seen anything from anyone since the delay issue was first >>>>> posted >>>>> to indicate that it has been looked at or fixed. It is a HUGE >>>>> issue related >>>>> to live streaming and is blocking a major project of ours too. >>>>> I've asked >>>>> before if there is anything we can do to help or where we should start >>>>> looking but never received any direction. Does anyone know >>>>> anything about >>>>> this issue or what we can do to help?? >>>>> >>>>> Bill >>>>> >>>>> Daniel Küstner wrote: >>>>> >>>>> >>>>>> Hi, >>>>>> >>>>>> I am working on the AV chat for Covcell (see http:// >>>>>> www.covcell.org/). During the Covcell project we are implementing >>>>>> some extensions to the open source LMS Moodle, among other things >>>>>> an audio-/video chat. The chat component consists of a flash >>>>>> client, compiled from OpenLaszlo sources and the Red5 streaming >>>>>> server. >>>>>> >>>>>> First I'd like to thank you for the great work you are doing on >>>>>> Red5. It's a big help to have such an easy to use and powerful tool. >>>>>> >>>>>> I have two issues with Red5: >>>>>> >>>>>> 1) When streaming a broadcast stream, the result at client side >>>>>> is getting a delay after some time. Sometimes the video stream >>>>>> catches up again, sometimes it's getting worse, the longer I am >>>>>> streaming. I found the track ticket >>>>>> http://mirror1.cvsdude.com/trac/osflash/red5/ ticket/177. >>>>>> >>>>>> Does anybody know the progress on this bug? Will it be fixed in >>>>>> 0.6 final? >>>>>> >>>>>> 2) In order to check, if it is fixed in 0.6 RC2, I downloaded it, >>>>>> but got some build errors in Eclipse. The errors are all >>>>>> happening at accesses to the class javax.script.ScriptEngine, >>>>>> which is delivered in js-223-1.0-pr.jar. Maybe a wrong version of >>>>>> that library was delivered? >>>>>> >>>>>> I am using Eclipse 3.2.1 on a Mac OS X 10.4.8 with Java 1.5.0. >>>>>> >>>>>> greetinx. daniel. >>>>>> >>>>>> >>>>> _______________________________________________ >>>>> Red5 mailing list >>>>> [email protected] >>>>> http://osflash.org/mailman/listinfo/red5_osflash.org >>>>> >>>>> >>>>> _______________________________________________ >>>>> Red5 mailing list >>>>> [email protected] >>>>> http://osflash.org/mailman/listinfo/red5_osflash.org >>>>> >>>>> >>>> _______________________________________________ >>>> Red5 mailing list >>>> [email protected] >>>> http://osflash.org/mailman/listinfo/red5_osflash.org >>>> >>>> >>>> >>> _______________________________________________ >>> Red5 mailing list >>> [email protected] >>> http://osflash.org/mailman/listinfo/red5_osflash.org >>> >>> > > _______________________________________________ > Red5 mailing list > [email protected] > http://osflash.org/mailman/listinfo/red5_osflash.org > _______________________________________________ Red5 mailing list [email protected] http://osflash.org/mailman/listinfo/red5_osflash.org
