SIP server stops responding

1. Brekeke Product Name and version: SIP Server 2.4.8.6/286.3

2. Java version: 1.7.0_04

3. OS type and the version: 2.6.18-274.18.1.el5.centos.plus

4. UA (phone), gateway or other hardware/software involved: Various UA's including Cisco 2821, Polycom IP 450's, Bria

5. Select your network pattern from http://www.brekeke-sip.com/bbs/network/ ... terns.html :Pattern 6

6. Your problem:

The issue is very simple but difficult to troubleshoot. The SIP server stops processing SIP packets. Upon logging into the GUI a message pops up saying the SIP server stopped responding at which point we are able to restart the SIP server via the GUI.

How can I go about troubleshooting root cause? Is there a log I can access that shed some light on the issue?

Thanks,

Isaac

Do you know how many concurrent calls at that time?
Which edition of Brekeke SIP Server are you using?
Is there any SIP UA which uses TCP for connecting to the SIP server?

Tata wrote:Do you know how many concurrent calls at that time?

It's hard to tell without a log. I can't imagine there would be more than a handful of calls, at most 5 but there could also be no active calls at the time it crashes.

Tata wrote:Which edition of Brekeke SIP Server are you using?

Standard

Tata wrote:Is there any SIP UA which uses TCP for connecting to the SIP server?

Most, if not all of the SIP UA's are connecting via UDP. There are isolated instances where mobile UA's may connect via TCP for mobile phone registrations.


It would really help if I could reference a log somewhere. Is there a log somewhere I can access?

Thanks for all your help.

How often does the problem happen?

Can you access to the SIP Server's GUI and go to the [Server Status] page when the SIP server stops responding?


If the GUI looks like a freeze, I suppose OS's TCP handling might have caused the issue.

FYI: you can find SIP server's logs at
<TOMCAT>/webapps/proxy/WEB-INF/work/sv/log/2012/

Can you find any error/exception in logs?

Tata wrote:How often does the problem happen?

The issue is intermittent.

Tata wrote:Can you access to the SIP Server's GUI and go to the [Server Status] page when the SIP server stops responding?

Tata wrote:
Yes

Tata wrote:Can you find any error/exception in logs?

In sv.20120510.1.log I found this: (ip addresses obfuscated for secuirty)

Code:

session.372894: send: to=UAC:8.8.8.8(UDP) at 05/10/12 02:54:10.072
==============================================
SIP/2.0 500 Internal Server Error
Via: SIP/2.0/UDP 8.8.8.8:5060;branch=z9hG4bK6fe57506;rport=5060
From: "Unknown" <sip:Unknown@8.8.8.8>;tag=as4da7f4df
To: <sip:2368@8.8.4.4:5060>;tag=E698BD7A-6FDFC431
CSeq: 102 NOTIFY
Call-ID: 68cbbd6e-44f460d5-48b28200@192.168.1.102
Record-Route: <sip:8.8.4.4:5060;lr>
Event: message-summary
User-Agent: PolycomSoundPointIP-SPIP_450-UA/3.2.3.1734
Accept-Language: en
Content-Length: 0


==============================================

session.372894: sipex.372893: close: result=Failure error=500 total-pkt=2 at 05/10/12 02:54:10.082

shutdown: processing.
shutdown: sender=127.0.0.1:44248
shutdown: now
listener: close all transport listeners and connections
tcp-listener: close port
listener: close: port=5060  current=05/10/12 02:54:12.146  life-length=8days 18:56:36  total-session=183393

sv: listener-phase=3
sv: exit-code=0
sv: exit-time=05/10/12 02:54:12.147
sv: total.mem=537067520  free.mem=2834784  thread=2
sv: Sayonara!

Here is a snippet from another crash: (sv.20120601.1.log)

Code:

session.403605: phase=0: OPTIONS Initializing
session.403603: became zombie at OPTIONS Finishing
session.403603: sipex.403602: close: result=Time Out error=504 total-pkt=1 at 06/01/12 14:51:41.942

sv: listener-phase=1
sv: exit-code=0
sv: exit-time=06/01/12 14:56:47.883
sv: total.mem=667090944  free.mem=371312  thread=4
sv: Sayonara!

I can see it's crashing, but that's about it.

If you are using an evaluation edition, the SIP server will stop randomly.

http://wiki.brekeke.com/wiki/sip-server ... top-resume

Yes, Eval can stop at anytime.

Just to add in here about server stops. The only reason's I have seen a server stop is if the cpu was overwhelmed by a sip hack attack or a plugin or function took too long to execute thereby stacking up sip processes until crash. The second scenario has been addressed by Brekeke with version 3 i believe. At least in the advanced sip server it was. You can specify how long Brekeke should wait for a plugin.

In 7 years that's the only reason i have personally observed a Brekeke sip server stopping.