[asterisk-bugs] [JIRA] (ASTERISK-21385) SIP Channel Lock

[Ross Beer (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-21385?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Ross Beer updated ASTERISK-21385:
---------------------------------

    Attachment: Asterisk 11.3.0 Lock at 10CPS 50 Concurrent Calls.txt
                Asterisk 11.2.0 Lock at 10CPS 50 Concurrent Calls.txt
                Asterisk 10.12.2 Lock at 10CPS 560 Concurrent Calls.txt
                Asterisk 1.8.21.0 Lock at 10CPS 360 Concurrnet Calls.txt
                Asterisk 1.8.19.0 Lock at 10CPS 460 Concurrent Calls.txt

I created another ticket when you closed this one which has the same information. Can you please merge the ticket ASTERISK-21386?

I have been testing with different asterisk versions and have some interesting results:

Asterisk 1.8.19.0 460 Concurrent Calls Before Lock

Asterisk 1.2.21.0 360 Concurrent Calls Before Lock

Asterisk 10.12.2 560 Concurrent Calls Before Lock

Asterisk 11.2.0 50 Concurrent Calls Before Lock

Asterisk 11.3.0 50 Concurrent Calls Before Lock


I have attached the lock status for each version. The best performing version is Asterisk 10.

All of the locks appear to be in the same place.
                
> SIP Channel Lock
> ----------------
>
>                 Key: ASTERISK-21385
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-21385
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Channels/chan_sip/General
>    Affects Versions: 1.8.21.0, 11.4.0
>         Environment: Cent OS
>            Reporter: Ross Beer
>         Attachments: Asterisk 10.12.2 Lock at 10CPS 560 Concurrent Calls.txt, Asterisk 11.2.0 Lock at 10CPS 50 Concurrent Calls.txt, Asterisk 11.3.0 Lock at 10CPS 50 Concurrent Calls.txt, Asterisk 1.8.19.0 Lock at 10CPS 460 Concurrent Calls.txt, Asterisk 1.8.21.0 Lock at 10CPS 360 Concurrnet Calls.txt, ThreadLock2.rtf, ThreadLocks.txt
>
>
> When above 10 calls per second are placed there is a thread lock which means that the SIP channel stops processing calls. There is a fix for the general SIP channel in 11.4 however this does not appear to resolve Realtime ODBC issues.
> RealTime peers only seam to use a single connection to the ODBC database even when configured to use more:
> ODBC DSN Settings
> -----------------
>   Name:   mysql
>   DSN:    MySql
>     Last connection attempt: 1970-01-01 01:00:00
>   Pooled: Yes
>   Limit:  100
>   Connections in use: 3
>     - Connection 1: in use (res_config_odbc.c:180 realtime_odbc)
>     - Connection 2: connected (:0 )
>     - Connection 3: connected (:0 )
> The RealTime ODBC process should allow as many connections as possible else the backlog causes the SIP channel to stop responding.
> I believe the SIP channel is threaded and therefore should be able to use more connections as needed.
> This is a major issue as SIP registrations stop when there are 100's of devices registered and calls are being made via the SIP channel.
> On each outbound call, the database is access to lookup the peer making the call. The peer is not stored in cache for any period of time and therefore this also increases the need for accessing the database.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.asterisk.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira