Supervisor: killing unresponsive child

[rodrigopassos]

Hello dears,

we have problem with yate how could be see in attached file.
Yate has generated a core file, after supervisor restart the process.
We have the same scenario without any change, and now, a problem is happening without any information into log file.
Someone could analyse the file generated by gdb and help to identify what happening?

BR,

Rodrigo Passos

[marian]

There are a lot of threads waiting in mysqldb module (wait for semaphore lock).
This indicates a very slow database.
Try increasing the 'poolsize' parameter value (try setting it to 10) in mysqldb.conf.
Check the 'database' messages (the 'query' parameter) to see who is using it (what table(s), for what purpose).

They may 'eat' all internal message dispatching threads.
When this happens the supervisor will kill and restart the process.

[rodrigopassos]

Marian,

Thank you for your help.
I made a database clear of old informations. It occurs on the cdr process, where we need to generate some information about calls. But, i think yate have problems with some data. How can I wait more time without Yate stop their process?

Regards,

Rodrigo R Passos

[marian]

Hi,
You may increase the maximum number of workers to create (yate.conf, [general] section, 'maxworkers' parameter).

If cdr update is an issue you may consider implementing some local database for it and move records when finalized to cdr database.

[rodrigopassos]

Mariam,

The process is local from the beginning. But, a have problems.
I will increase the maxworkers.
What is the default value?  How can i planning a value to optimize my ambient?

BR,

Rodrigo