fop2 service dying
We installed fop2 about 2 weeks ago. In that time frame the fop2 service has died 3 times. We are not seeing anything logged in /var/log/messages. We are running Asterisk 1.4.21.2 on CentOS 5.3. Has anybody else experienced this? Is there a way to enable logging so we can try to figure out what is happening?
Comments
There is a way and I will very much appreciate if you can capture the output. You have to run fop2 on the command line enabling debug, capturing the output. The best way, in my opinion is to:
Stop fop2 as a service
And then use gnu screen and script to start fop2_server:
Then you have to dettach the screen console with ctrl-D
Verify that fop2_server is running:
You can see the debug output while it is being wrote to disk:
I have fixed some of this crashes already, maybe updating your setup (just replacing fop2_server) will cure the problem, but it would be nice anyways to capture the error as described above. Once fop2_server dies, you can reattach the screen:
You will see the last lines and error on your screen. Type "exit" to quit scripts from capturing to output.log, Type "exit" again to quit screen.
127.0.0.1 <- Event: RTCPSent
127.0.0.1 <- Privilege: reporting,all
127.0.0.1 <- To: 192.168.11.129:19435
127.0.0.1 <- OurSSRC: 1859564915
127.0.0.1 <- SentNTP: 1260993389.2531479552
127.0.0.1 <- SentRTP: 1552160
127.0.0.1 <- SentPackets: 9701
127.0.0.1 <- SentOctets: 1552160
127.0.0.1 <- ReportBlock:
127.0.0.1 <- FractionLost: 0
127.0.0.1 <- CumulativeLoss: 0
127.0.0.1 <- IAJitter: 0.0004
127.0.0.1 <- TheirLastSR: 0
127.
The complete log is 126mb because today it lasted over 8 hours without crashing. Some days it can crash every few hours.
The last lines of the log should be some kind of error like "Can't call method XXX" on line YYYY
Please try to catch that, the output you posted seems incomplete. If you run fop2 under gnu screen as mentioned it must show something like that before crashing.
I have fixed some of this issues in the next release, but I might need to squash some more of them, and the only way to find the problems is by user feedback, as usually crashes happen under certain conditions that are hard or impossible to replicate/reproduce (unconfigured buttons, extremely loaded servers, etc)
Best regards,