Summary:ASTERISK-15384: Wrong Caller ID on Sip Channel when have a lot of hanging up call
Reporter:vutamhoan (vutamhoan)Labels:
Date Opened:2010-01-01 03:48:15.000-0600Date Closed:2010-01-25 08:43:37.000-0600
Versions:Frequency of
Description:Our system working as a simple IVR & ACD:

PSTN <======> MediaGateway <==SIP==> Asterisk <===SIP===> IPPhone

CDR Table has more than 2.5M records. And today (01/01) our crontab runs the cleaning task that leads Mysql get very slow. So when call hangup, it takes long time to write CDR and system have a lot of dead call (with load of 80 call concurrent, after 2 hour we got about 3000 dead call that insert CDR)
And when system reach 500 dead calls, all callerid send to IPPhone is wrong.

Agents sometime report the wrong callerid when system in high load (200 calls), but only today I can reproduce the bug so I suppose this is the glue.
I thought this relate to https://issues.asterisk.org/view.php?id=16054, but I set session-timers=refuse and "sip show channels" show the normal number of dialogs.


02 Server: Asterisk & Database with the same configuration.
Supermicro 4 core, 8GB RAM Cenos5.3 64bit
Comments:By: vutamhoan (vutamhoan) 2010-01-01 03:58:32.000-0600

I'm using Asterisk from tar ball source.
Core show channels got a lot of dead call
[root@mediasv01 ~]# asterisk -rx "core show channels"
SIP/xx.xx.xx.xx-07f s@auto_ivr:13        Up      BackGround(/sounds/ivr
SIP/xx.xx.xx.xx-07f s@callcenter:4       Up      Queue(ivr_queue,t,,,29
SIP/xx.xx.xx.xx-07f s@auto_ivr:13        Up      BackGround(/sounds/ivr
SIP/xx.xx.xx.xx-cc0 s@auto_ivr:13        Up      BackGround(/sounds/ivr
3591 active channels
3586 active calls
13096 calls processed

But sip has normal number of dialogs
[root@mediasv01 ~]# asterisk -rx "sip show channels"
xx.xx.xx.xx     209838068   D146B6BD-BF1F11  0x4 (ulaw)       No       Rx: ACK
xx.xx.xx.xx     209760592   DB485F2D-BF1F11  0x4 (ulaw)       No       Rx: ACK
53 active SIP dialogs

xx.xx.xx.xx is ip of media gateway

By: Leif Madsen (lmadsen) 2010-01-04 10:23:26.000-0600

Can you reproduce this on the latest 1.6.1? If you can't run this on your production machine, please reproduce on a test system.

By: vutamhoan (vutamhoan) 2010-01-04 10:49:04.000-0600

Yeap, I'm thinking about upgrade to latest 1.6.1 - Will run and report in test system first.

By: vutamhoan (vutamhoan) 2010-01-24 23:25:30.000-0600

I upgraded to and it seems gone away. Wish it never happens again.
Thank you very much

By: Leif Madsen (lmadsen) 2010-01-25 08:43:36.000-0600

Closed per reporter.