I have just checked the file you have created: [flashsv @ 0x28fcd60] Frame width or height differs from first frame! [flashsv @ 0x28fcd60] fh = 810, fv 1440 vs ch = 1080, cv = 1920 Error while decoding stream #0:0: Invalid data found when processing input
According to the file header: Stream #0:0: Video: flashsv ([3][0][0][0] / 0x0003), bgr24, 1440x800, q=2-31, 200 kb/s, 1k tbn, 10 tbc so the header says: Frame height: 800 The frame itself says: Frame height: 810 No sure how this possible. According to flvrecording database table: mysql> select height, width, flv_height, flv_width from flvrecording where id = 226; +--------+-------+------------+-----------+ | height | width | flv_height | flv_width | +--------+-------+------------+-----------+ | 810 | 1440 | 800 | 1440 | +--------+-------+------------+-----------+ why do we have both height and flv_height what is the difference? On Thu, Mar 6, 2014 at 3:10 AM, Maxim Solodovnik <solomax...@gmail.com>wrote: > I'll ask Vasiliy to share his recording and/or create another one. > Will try to check your recording in the morning > seems like it have no videodata (or broken videodata) :( > Not sure what can be done here, but i'll check > > > On Thu, Mar 6, 2014 at 3:03 AM, seba.wag...@gmail.com < > seba.wag...@gmail.com> wrote: > >> Does it take for you also that long when you hit the stop button until the >> recording UI does really stop and switch the button state ? >> >> Can we have a look at the 1 hour recording, can you put that in the public >> folder too ? >> >> I created another video on demo.dataved.ru using a Windows machine. I put >> it into the public folder. >> This time the recording has not visual, only audio. I also cannot see any >> kind of error. The AVI file when I download it has less then 1 MB. >> Would you mind checking what was wrong with it? It is flvRecordingId 226. >> >> Thanks, >> Sebastian >> >> >> >> >> 2014-03-06 1:05 GMT+13:00 Maxim Solodovnik <solomax...@gmail.com>: >> >> > I haven't monitored Red5 CPU usage >> > But the system is very responsive (Maybe because it is core i7 8 cores, >> > 8Gb) >> > I see no performance degradation. >> > >> > The stream is still being written to the disc asynchronously (this code >> was >> > slightly optimized) >> > I have put the recording made just now on my machine to the same public >> > group, you can check it. >> > >> > Vasiliy was able to create recording with duration ~1 hour (while >> > performing test plan for this RC) >> > >> > >> > On Wed, Mar 5, 2014 at 2:52 PM, seba.wag...@gmail.com < >> > seba.wag...@gmail.com >> > > wrote: >> > >> > > Hi Maxim, >> > > >> > > The other topics to be covered in another conversation, I think this >> > issue >> > > is worth a separated thread. >> > > >> > > "Could you please move your recording into Public (Openmeetings) group >> > so I >> > > can take a look at it?" >> > > => I did just that now. >> > > >> > > "All test recordings I have made were made locally, should will try >> to do >> > > it on demo." >> > > => I did try locally too, but the recordings are not really working. >> The >> > > recorded file freezes after 10-20 sec. Sound sometimes continues >> > sometimes >> > > sound stops too. >> > > >> > > If I start a recording I can see two processes jump in the CPU usage: >> > > 1) Screensharing application to 40-50% >> > > 2) The red5 server process jumps to 100%CPU (or what is leftover after >> > the >> > > screen-sharing webapp takes 40%) >> > > >> > > So did you monitor the red5 server process CPU while doing a >> > screen-sharing >> > > recording? >> > > I think the high server side CPU usage might be also the reason why >> when >> > > you click on the stop recording button for ages just nothing happens. >> > > >> > > We had a similar problem in the past where the red5 cpu started to use >> > 100% >> > > of the CPU while recording. The issue at that time was that the >> packets >> > > could not be written fast enough to disk. So I implemented the asynch >> > > stream writer with the BlockingQueue that caches the packets and >> starts a >> > > separated Thread to write the packets to disk in a reasonable >> interval. >> > > That is also why there is this wait period/check in the recordings >> > > processing that needs to wait until all data really is written to disk >> > > [see: BaseConverter#waitForTheStream()]. >> > > At that point in time that seemed to solve the problem. >> > > >> > > However the source of the issue might be also a purely red5 related >> > issue. >> > > Nevertheless I think we need to solve/verify it. >> > > >> > > At some point the screen-sharing webapp became just completely >> > > unresponsive. And only by killing the process of the screen-sharing >> app >> > the >> > > red5 server side process did also reduce CPU usage to a reasonable >> level. >> > > >> > > Do you experience a similar problem with the red5 server process while >> > > doing a recording ? >> > > I do not completely deny that it might be also a hardware problem, for >> > > instance if the disk is too slow. But I am using a SSD flash disk :) >> So >> > > that should be fast enough :) However in the future we might need to >> > > revisit the write speed for the stream writer that writes the queue to >> > > disk. >> > > >> > > The problem with this high Red5 Cpu usage in the past was also that it >> > will >> > > slow down pretty much anything, for instance all other streams even in >> > > other conference rooms will become slow-motion, funny sound effects et >> > > cetera. >> > > >> > > Sebastian >> > > >> > > >> > > >> > > 2014-03-04 23:47 GMT+13:00 Maxim Solodovnik <solomax...@gmail.com>: >> > > >> > > > Hello Sebastian, >> > > > >> > > > Thanks for detailed testing >> > > > >> > > > Could you please move your recording into Public (Openmeetings) >> group >> > so >> > > I >> > > > can take a look at it? >> > > > All test recordings I have made were made locally, should will try >> to >> > do >> > > it >> > > > on demo. >> > > > >> > > > now will try to answer in bullet list: >> > > > 1) "Can we make it that somehow that it will be displayed directly" >> I >> > can >> > > > propose to choose an icon from FamFam set (for ex. time.png, >> > > hourglass.png, >> > > > clock.png, clock_pause.png) will appreciate any help on this since I >> > > > personally not really like all the icons I have mentioned. >> > > > >> > > > 2) "Previously you did get the pop warning 'not ready' directly on a >> > > click >> > > > on the recording when it is not yet ready for viewing." I don't >> really >> > > like >> > > > those pop ups since I find it annoying, I believe correct icon is >> > better >> > > > solution. (Maybe tooltip should be added) >> > > > >> > > > 3) "but I cannot load the debug window on that server." currently >> debug >> > > is >> > > > only available if wicket is run in debug mode, this might need to be >> > > > corrected (or maybe not since room should be rewritten) >> > > > >> > > > 4) Also you send me another issue report: contact request email has >> > > fields >> > > > mixed up (from and to user names are mixed) >> > > > >> > > > 5) "What about the attribute 'length' of the recording" this fild of >> > the >> > > > recording is not being filled. What should be written in it? >> recording >> > > > duration? recording file length? >> > > > >> > > > 6) Appointments inside private messages were reported to was broken >> in >> > > > previous versions as well, so we need to file an issue and fix in in >> > > > 3.1.0/3.0.1 >> > > > >> > > > What is your resume? >> > > > Should we release or wait for RC3? what of the above are blockers? >> > > > >> > > > >> > > > >> > > > On Tue, Mar 4, 2014 at 3:27 PM, seba.wag...@gmail.com < >> > > > seba.wag...@gmail.com >> > > > > wrote: >> > > > >> > > > > "not ready status will be displayed after click on "!" icon in the >> > > > popup." >> > > > > I did not get a not ready status displayed. I see the icon on the >> > left, >> > > > but >> > > > > neither by clicking on the icon nor on the recording anything >> > happens. >> > > > > >> > > > > Can we make it that somehow that it will be displayed directly >> when >> > you >> > > > > click on the recording? >> > > > > >> > > > > Having links that you click and under some circumstances nothing >> > > happens >> > > > > are just very hard to understand for end-users. Previously you did >> > get >> > > > the >> > > > > pop warning 'not ready' directly on a click on the recording when >> it >> > is >> > > > not >> > > > > yet ready for viewing. >> > > > > >> > > > > I tried to do three recordings on demo.dataved.ru, all failed. >> > > Possibly >> > > > > because I am using Mac and/or bandwidth issues. >> > > > > I was able to produce a recording at the 06th of February on that >> > > server >> > > > > with a Mac, so it is not a general issue. >> > > > > In the error log window I can see some ffmpeg command fails (see >> > > attached >> > > > > on the bottom of the mail some example) >> > > > > >> > > > > But the screen sharer itself stayed at 50% of the CPU which is >> > > acceptable >> > > > > for the first release of 3.x.x from my point of view. >> > > > > Just to get some recordings produced with it would be something I >> > would >> > > > be >> > > > > keen to verify. >> > > > > >> > > > > Another issues was that I cannot enter a room on >> > demo.dataved.ruafter a >> > > > > while. I just stuck in the screen when the loading bar in front of >> > > > entering >> > > > > the conference room is complete. I guess there is some null >> pointer >> > > > > exception in the code but I cannot load the debug window on that >> > > server. >> > > > > I was using the "remember me" cookie, after deleting the cookie >> and >> > > > > re-login entering the room did work again. >> > > > > >> > > > > What about the attribute "length" of the recording in the UI, the >> > field >> > > > > seems to be simply not filled with any value yet, even if the >> > recording >> > > > was >> > > > > successful. Is that not implemented yet? >> > > > > >> > > > > Btw the autocomplete box in the private message center and >> calendar >> > is >> > > > > a-w-e-s-o-m-e >> > > > > I think we should leave it as is. Normally you can only send >> messages >> > > to >> > > > > your contacts, but I think like that it is even better. The old >> one >> > was >> > > > too >> > > > > restrictive to be intuitive. >> > > > > When attaching a meeting request with a date/time to a private >> > message >> > > > ... >> > > > > it did not show up in my calendar afterwards. Is that something >> that >> > > > needs >> > > > > to done still ? >> > > > > >> > > > > This is the FFMPEG error on all my recordings: >> > > > > process: previewFullFLV command: ffmpeg -i >> > > > > >> > > > > >> > > > >> > > >> > >> /opt/openmeetings-3.0.0.r1573575/webapps/openmeetings/streams/hibernate/flvRecording_210.flv >> > > > > -vcodec mjpeg -vframes 1 -an -f rawvideo -s 1072x672 >> > > > > >> > > > > >> > > > >> > > >> > >> /opt/openmeetings-3.0.0.r1573575/webapps/openmeetings/streams/hibernate/flvRecording_210.jpg >> > > > > exception: null error: ffmpeg version git-2013-10-22-c68faca >> > Copyright >> > > > (c) >> > > > > 2000-2013 the FFmpeg developers built on Oct 22 2013 12:58:25 with >> > gcc >> > > > 4.6 >> > > > > (Ubuntu/Linaro 4.6.3-1ubuntu5) configuration: >> > > --prefix=/root/ffmpeg_build >> > > > > --extra-cflags=-I/root/ffmpeg_build/include >> > > > > --extra-ldflags=-L/root/ffmpeg_build/lib --bindir=/root/bin >> > > > > --extra-libs=-ldl --enable-gpl --enable-libass --enable-libfdk-aac >> > > > > --enable-libmp3lame --enable-libopus --enable-libtheora >> > > > --enable-libvorbis >> > > > > --enable-libvpx --enable-libx264 --enable-nonfree --enable-x11grab >> > > > > libavutil 52. 47.101 / 52. 47.101 libavcodec 55. 37.102 / 55. >> 37.102 >> > > > > libavformat 55. 19.103 / 55. 19.103 libavdevice 55. 4.100 / 55. >> 4.100 >> > > > > libavfilter 3. 89.100 / 3. 89.100 libswscale 2. 5.101 / 2. 5.101 >> > > > > libswresample 0. 17.104 / 0. 17.104 libpostproc 52. 3.100 / 52. >> > 3.100 >> > > > [flv >> > > > > @ 0x3693900] Could not find codec parameters for stream 0 (Video: >> > none, >> > > > 200 >> > > > > kb/s): unspecified size Consider increasing the value for the >> > > > > 'analyzeduration' and 'probesize' options Input #0, flv, from >> > > > > >> > > > > >> > > > >> > > >> > >> '/opt/openmeetings-3.0.0.r1573575/webapps/openmeetings/streams/hibernate/flvRecording_210.flv': >> > > > > Metadata: noaudiocodec : 0 server : Red5 Server 1.0.0 RC3 Rev: >> 4393 >> > > > > creationdate : Tue Mar 04 08:04:03 CET 2014 canSeekToEnd : true >> > > encoder : >> > > > > Lavf55.19.103 Duration: 00:01:28.87, start: 0.000000, bitrate: 593 >> > kb/s >> > > > > Stream #0:0: Video: none, 200 kb/s, 1k tbr, 1k tbn, 1k tbc Stream >> > #0:1: >> > > > > Audio: mp3, 22050 Hz, mono, s16p, 32 kb/s Output #0, rawvideo, to >> > > > > >> > > > > >> > > > >> > > >> > >> '/opt/openmeetings-3.0.0.r1573575/webapps/openmeetings/streams/hibernate/flvRecording_210.jpg': >> > > > > Metadata: noaudiocodec : 0 server : Red5 Server 1.0.0 RC3 Rev: >> 4393 >> > > > > creationdate : Tue Mar 04 08:04:03 CET 2014 canSeekToEnd : true >> > > encoder : >> > > > > Lavf55.19.103 Output file #0 does not contain any stream >> exitValue: >> > > null >> > > > > out: >> > > > > >> > > > > Sebastian >> > > > > >> > > > > >> > > > > >> > > > > 2014-03-04 20:18 GMT+13:00 Maxim Solodovnik <solomax...@gmail.com >> >: >> > > > > >> > > > > > Currently "not ready" recordings are marked as "broken" (same >> "!" >> > > icon) >> > > > > > not ready status will be displayed after click on "!" icon in >> the >> > > > popup. >> > > > > > >> > > > > > interview recordings are now converted as fast as ordinary >> > > recordings. >> > > > > > >> > > > > > Icon can be changed, please suggest one :) (Do you feel it is >> > > blocker? >> > > > or >> > > > > > maybe we can release 3.0.1 will all such minor fixes?) >> > > > > > >> > > > > > >> > > > > > http://demo.dataved.ru:5080/ contains the same SNAPSHOT >> > > > > > >> > > > > > >> > > > > > >> > > > > > On Tue, Mar 4, 2014 at 2:12 PM, seba.wag...@gmail.com < >> > > > > > seba.wag...@gmail.com >> > > > > > > wrote: >> > > > > > >> > > > > > > I think there is still a slight glitch with the recordings UI. >> > > > > > > >> > > > > > > Previously, when a recording was not yet ready for viewing >> > (because >> > > > it >> > > > > is >> > > > > > > still processing) there was a message in the UI telling you >> that. >> > > In >> > > > > the >> > > > > > > case of interviews there was even a progress includes in >> > > percentages >> > > > as >> > > > > > > interviews take 1 minutes of processing time for the >> rendering of >> > > a 1 >> > > > > > > minute of video. >> > > > > > > >> > > > > > > If you at the moment are too fast in switching to the UI and >> > there >> > > > was >> > > > > > some >> > > > > > > kind of error or its kind of simply not ready for viewing you >> > click >> > > > on >> > > > > > the >> > > > > > > video and just nothing happens. >> > > > > > > >> > > > > > > What is the current mechanism? Do we catch in the recordings >> UI >> > if >> > > > the >> > > > > > > video is ready for viewing or not? >> > > > > > > >> > > > > > > Thanks, >> > > > > > > Sebastian >> > > > > > > >> > > > > > > >> > > > > > > 2014-03-04 19:54 GMT+13:00 seba.wag...@gmail.com < >> > > > > seba.wag...@gmail.com >> > > > > > >: >> > > > > > > >> > > > > > > > Well if its just me I do not want to complain. >> > > > > > > > I am using a Mac, probably that is part of the problem. >> > > > > > > > >> > > > > > > > Let me try to run a couple of tests. >> > > > > > > > >> > > > > > > > Btw: Is http://demo.dataved.ru:5080/ the same SNAPSHOT >> that is >> > > up >> > > > > for >> > > > > > a >> > > > > > > > vote? >> > > > > > > > >> > > > > > > > Thanks! >> > > > > > > > Sebastian >> > > > > > > > >> > > > > > > > >> > > > > > > > 2014-03-04 17:14 GMT+13:00 Maxim Solodovnik < >> > > solomax...@gmail.com >> > > > >: >> > > > > > > > >> > > > > > > > Hello Sebastian, >> > > > > > > >> >> > > > > > > >> I was unable to fix high CPU usage because I can't see it >> :( >> > > > > > > >> We have tried to reproduce it on our machines (me and >> Vasiliy, >> > > > > Vasiliy >> > > > > > > >> provides you with his measurements), but it is not >> > reproducible. >> > > > > > > >> No additional comments were provided, so this issue have >> been >> > > > > > postponed >> > > > > > > >> >> > > > > > > >> >> > > > > > > >> >> > > > > > > >> >> > > > > > > >> On Tue, Mar 4, 2014 at 11:08 AM, seba.wag...@gmail.com < >> > > > > > > >> seba.wag...@gmail.com> wrote: >> > > > > > > >> >> > > > > > > >> > Hi Maxim, >> > > > > > > >> > >> > > > > > > >> > Thanks, I think one of the major obstacles has been the >> CPU >> > > > usage >> > > > > of >> > > > > > > the >> > > > > > > >> > screen sharing application. >> > > > > > > >> > Did we reduce the CPU usage significantly? >> > > > > > > >> > >> > > > > > > >> > Thanks, >> > > > > > > >> > Sebastian >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > 2014-03-04 5:41 GMT+13:00 Alvaro <zurca...@gmail.com>: >> > > > > > > >> > >> > > > > > > >> > > +1 >> > > > > > > >> > > >> > > > > > > >> > > ------------------- >> > > > > > > >> > > >> > > > > > > >> > > >> > > > > > > >> > > El lun, 03-03-2014 a las 23:09 +0700, Maxim Solodovnik >> > > > escribió: >> > > > > > > >> > > > Dear OpenMeetings Community, >> > > > > > > >> > > > >> > > > > > > >> > > > I would like to start a vote about releasing Apache >> > > > > OpenMeetings >> > > > > > > >> 3.0.0 >> > > > > > > >> > > RC2 >> > > > > > > >> > > > >> > > > > > > >> > > > Main changes are covered in the >> > > > > > > >> > > > Readme: >> > > > > > > >> > > >> > > > > > >> https://svn.apache.org/repos/asf/openmeetings/tags/3.0.0RC2/README >> > > > > > > >> > > > >> > > > > > > >> > > > Full Changelog: >> > > > > > > >> > > > >> > > > > > > >> >> > > > > >> > https://svn.apache.org/repos/asf/openmeetings/tags/3.0.0RC2/CHANGELOG >> > > > > > > >> > > > >> > > > > > > >> > > > Release artefacts: >> > > > > > > >> > > > >> > > > > https://dist.apache.org/repos/dist/dev/openmeetings/3.0.0/rc2/ >> > > > > > > >> > > > >> > > > > > > >> > > > Tag: >> > > > > > https://svn.apache.org/repos/asf/openmeetings/tags/3.0.0RC2/ >> > > > > > > >> > > > >> > > > > > > >> > > > PGP release keys (signed using >> > > > > > > >> > > > C467526E): >> > > > > > > >> > > >> > > > > > >> https://dist.apache.org/repos/dist/dev/openmeetings/3.0.0/rc2/KEYS >> > > > > > > >> > > > >> > > > > > > >> > > > Vote will be open for 72 hours. >> > > > > > > >> > > > >> > > > > > > >> > > > [ ] +1 approve >> > > > > > > >> > > > [ ] +0 no opinion >> > > > > > > >> > > > [ ] -1 disapprove (and reason why) >> > > > > > > >> > > > >> > > > > > > >> > > > My vote is +1. >> > > > > > > >> > > > >> > > > > > > >> > > > >> > > > > > > >> > > >> > > > > > > >> > > >> > > > > > > >> > > >> > > > > > > >> > >> > > > > > > >> > >> > > > > > > >> > -- >> > > > > > > >> > Sebastian Wagner >> > > > > > > >> > https://twitter.com/#!/dead_lock >> > > > > > > >> > http://www.webbase-design.de >> > > > > > > >> > http://www.wagner-sebastian.com >> > > > > > > >> > seba.wag...@gmail.com >> > > > > > > >> > >> > > > > > > >> >> > > > > > > >> >> > > > > > > >> >> > > > > > > >> -- >> > > > > > > >> WBR >> > > > > > > >> Maxim aka solomax >> > > > > > > >> >> > > > > > > > >> > > > > > > > >> > > > > > > > >> > > > > > > > -- >> > > > > > > > Sebastian Wagner >> > > > > > > > https://twitter.com/#!/dead_lock >> > > > > > > > http://www.webbase-design.de >> > > > > > > > http://www.wagner-sebastian.com >> > > > > > > > seba.wag...@gmail.com >> > > > > > > > >> > > > > > > >> > > > > > > >> > > > > > > >> > > > > > > -- >> > > > > > > Sebastian Wagner >> > > > > > > https://twitter.com/#!/dead_lock >> > > > > > > http://www.webbase-design.de >> > > > > > > http://www.wagner-sebastian.com >> > > > > > > seba.wag...@gmail.com >> > > > > > > >> > > > > > >> > > > > > >> > > > > > >> > > > > > -- >> > > > > > WBR >> > > > > > Maxim aka solomax >> > > > > > >> > > > > >> > > > > >> > > > > >> > > > > -- >> > > > > Sebastian Wagner >> > > > > https://twitter.com/#!/dead_lock >> > > > > http://www.webbase-design.de >> > > > > http://www.wagner-sebastian.com >> > > > > seba.wag...@gmail.com >> > > > > >> > > > >> > > > >> > > > >> > > > -- >> > > > WBR >> > > > Maxim aka solomax >> > > > >> > > >> > > >> > > >> > > -- >> > > Sebastian Wagner >> > > https://twitter.com/#!/dead_lock >> > > http://www.webbase-design.de >> > > http://www.wagner-sebastian.com >> > > seba.wag...@gmail.com >> > > >> > >> > >> > >> > -- >> > WBR >> > Maxim aka solomax >> > >> >> >> >> -- >> Sebastian Wagner >> https://twitter.com/#!/dead_lock >> http://www.webbase-design.de >> http://www.wagner-sebastian.com >> seba.wag...@gmail.com >> > > > > -- > WBR > Maxim aka solomax > -- WBR Maxim aka solomax
