[OpenSIPS-Users] Opensips restart and hanging process (dialogs?)

[Bogdan-Andrei Iancu]

Yes indeed, it looks like flushing dialog info into DB. How many dialog 
do you have ongoing and how fast your db is ??

Regards,

Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
http://www.opensips-solutions.com

On 08.05.2015 13:05, Max Mühlbronner wrote:
>
> Here is a pastebin of the bt (attendant process, while restarting). 
> Seems to be the database operations/dialogs?
>
>
> http://pastebin.com/9zHQjMMq
>
>
> BR
>
> Max M.
>
> On 08.05.2015 11:01, Bogdan-Andrei Iancu wrote:
>> Max,
>>
>> What you can do is to try to attach with GDB to the attendant process 
>> while doing the cleanup. It will be interesting to see what is going on.
>>
>> Regards,
>>
>> Bogdan-Andrei Iancu
>> OpenSIPS Founder and Developer
>> http://www.opensips-solutions.com
>>
>> On 08.05.2015 09:58, Max Mühlbronner wrote:
>>> Hi,
>>>
>>> Thanks for your reply. No, nothing special. Even changed db_mode for 
>>> dialog module to realtime. No usrloc.
>>>
>>> Delaying the start-stop-daemon (--retry) like Aron suggested is a 
>>> workaround for restarting, although i am still wondering why the 
>>> process is taking so long to quit.
>>>
>>> Indeed, see you at the Summit.
>>>
>>>
>>> BR
>>>
>>> Max M.
>>>
>>> On 07.05.2015 18:35, Bogdan-Andrei Iancu wrote:
>>>> Hi Max,
>>>>
>>>> The attendant is responsible for doing the cleanup for all modules 
>>>> and core - this is why it may take longer to complete. Do you have 
>>>> memory logging or some expensive DB flushing on shutdown ?
>>>>
>>>> BTW: see you in Amsterdam for the OpenSIPS Summit !
>>>>
>>>> Regards,
>>>>
>>>> Bogdan-Andrei Iancu
>>>> OpenSIPS Founder and Developer
>>>> http://www.opensips-solutions.com
>>>>
>>>> On 07.05.2015 11:01, Max Mühlbronner wrote:
>>>>>
>>>>> Comparing "opensipsctl ps" and the pid of the remaining process 
>>>>> shows it's the attendant process.
>>>>>
>>>>> root at opensips1:/etc/opensips# opensipsctl ps
>>>>>
>>>>> Process:: ID=0 PID=25935 Type=attendant
>>>>>
>>>>>
>>>>>
>>>>> root at opensips1:/etc/opensips# /etc/init.d/opensips restart
>>>>>
>>>>> Restarting opensips: opensips^[[Aroot at opensips1:/etc/opensips# ps 
>>>>> uax|grep opensips
>>>>>
>>>>> opensips 25935 0.0 2.1 962160 21768 ? S 09:44 0:00 /sbin/opensips 
>>>>> -P /var/run/opensips/opensips.pid -m 896 -M 32 -u opensips -g 
>>>>> opensips
>>>>>
>>>>>
>>>>> After a few seconds the attendant process is gone, and opensips is 
>>>>> not running anymore. Did anyone experience something like this 
>>>>> before?
>>>>>
>>>>>
>>>>> BR
>>>>>
>>>>> Max M.
>>>>>
>>>>> On 05.05.2015 11:00, Max Mühlbronner wrote:
>>>>>> Hi,
>>>>>>
>>>>>> when restarting Opensips it will shut down all the processes 
>>>>>> immediately, but there is still a single process left "hanging" 
>>>>>> (cpu load) which eventually exits after some time. It is not 
>>>>>> possible to restart Opensips until this process is killed/quits.
>>>>>>
>>>>>> root at opensips1:~# /etc/init.d/opensips restart
>>>>>> Restarting opensips: opensipsroot at opensips1:~#
>>>>>>
>>>>>>
>>>>>> Opensips version is 1.11.4-tls. But we also had the same issue 
>>>>>> with 1.7.x and 1.10.X.
>>>>>>
>>>>>> There are ~3000 active dialogs, but it should not take that long 
>>>>>> to dump everything into the database. (db_mode is set to "2") right?
>>>>>>
>>>>>> Any hints/ideas are appreciated.
>>>>>>
>>>>>>
>>>>>> Best Regards
>>>>>>
>>>>>>
>>>>>> Max M.
>>>>>>
>>>>>> _______________________________________________
>>>>>> Users mailing list
>>>>>> Users at lists.opensips.org
>>>>>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Users mailing list
>>>>> Users at lists.opensips.org
>>>>> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>>>
>>>>>
>>>>
>>>
>>>
>>
>
>