mem leak and core dump
Hello Freeswitch'rs, I recently began using the latest freeswitch beta code as a pure B2BUA for the purpose of topology hiding of IP - PSTN calls. running env: Linux b2bua 2.6.18 #6 SMP Thu Nov 16 10:35:45 EST 2006 i686 Intel(R) Xeon(R) CPU 5140 @ 2.33GHz GNU/Linux Freeswitch version: FreeSwitch Version 1.0.pre1 (6220M) The modules I am using are 711 and 729 codecs, sip, directory, xml_rpc, cdr, xml_cdr, console and syslog. I am seeing what appears to be a memory leak. Consumed memory continues to drop linearly based on number of calls. Also had one core dump. What can I do to help troubleshoot these issues? Today I plan to stop using the CDR, rpc and logging modules to see if it makes a difference. Also considering running on one of my fedora core 7 boxes instead of gentoo, again to see if anything changes. Ideas appreciated. Regards, Tom
Tom, You should really be running svn trunk as its very close to being tagged beta2 and has fixed some of the issues you're outlining in this email. This specific issue has already been fixed and was a bug in libsofia. /b On Nov 19, 2007, at 7:47 AM, Freeswitch Freeswitch wrote:
Hello Freeswitch'rs,
I recently began using the latest freeswitch beta code as a pure B2BUA for the purpose of topology hiding of IP - PSTN calls.
running env: Linux b2bua 2.6.18 #6 SMP Thu Nov 16 10:35:45 EST 2006 i686 Intel(R) Xeon(R) CPU 5140 @ 2.33GHz GNU/Linux Freeswitch version: FreeSwitch Version 1.0.pre1 (6220M)
The modules I am using are 711 and 729 codecs, sip, directory, xml_rpc, cdr, xml_cdr, console and syslog.
I am seeing what appears to be a memory leak. Consumed memory continues to drop linearly based on number of calls. Also had one core dump.
What can I do to help troubleshoot these issues?
Today I plan to stop using the CDR, rpc and logging modules to see if it makes a difference.
Also considering running on one of my fedora core 7 boxes instead of gentoo, again to see if anything changes.
Ideas appreciated.
Regards,
Tom _______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
On 11/19/07, Freeswitch Freeswitch <freeswitch@2ride.com> wrote:
Hello Freeswitch'rs,
I recently began using the latest freeswitch beta code as a pure B2BUA for the purpose of topology hiding of IP - PSTN calls.
running env:
Linux b2bua 2.6.18 #6 SMP Thu Nov 16 10:35:45 EST 2006 i686 Intel(R) Xeon(R) CPU 5140 @ 2.33GHz GNU/Linux Freeswitch version: FreeSwitch Version 1.0.pre1 (6220M)
The modules I am using are 711 and 729 codecs, sip, directory, xml_rpc, cdr, xml_cdr, console and syslog.
I am seeing what appears to be a memory leak. Consumed memory continues to drop linearly based on number of calls. Also had one core dump.
Can you provide details of the call scenarios you are testing, call rates, number of concurrent calls and any other details? What can I do to help troubleshoot these issues?
Today I plan to stop using the CDR, rpc and logging modules to see if it makes a difference.
We have tested quite heavily with mod_console running, that should not be an issue. Also considering running on one of my fedora core 7 boxes instead of gentoo,
again to see if anything changes.
We have had repeated issues with gentoo due to patches made either to gcc or to libc or the kernel. It is the only os/distro that I highly suggest never using. Mike
Hello, Now I have updated to the latest trunk as of Nov 21 and moved to a fedora 7 system. Still chasing random crashes of FS. First time this happened last night there were around 1000 sessions up. Crit error was logged referring to a signal caught on an unmapped thread. I wasn't able to catch the specific signal in the console before it scrolled off. This morning I had another crash. This time around 200 sessions running. There was no logging reported during the second crash. Calling scenario is about as simple as it gets. gw1 -> freeswitch -> gw2 all sip codecs, 729 and 711 transcoding off I have been playing with the various rtp knobs in the sofia conf. Sometimes running with rewriting rtp timerstamps, sometimes not. Also tried with soft timer and off. ( can someone point me to an explanation of what the soft timer does? ). Also notable that I compiled the code with debug symbols. No core has been generated during crashes. Is there some way to enable core dumps? Overall I am very impressed with FS - would love to get the cause of these crashes nailed so I can press forward. Any help is very much appreciated. Best regards, Tom On 11/19/07, Michael Jerris <mike@jerris.com> wrote:
On 11/19/07, Freeswitch Freeswitch <freeswitch@2ride.com> wrote:
Hello Freeswitch'rs,
I recently began using the latest freeswitch beta code as a pure B2BUA for the purpose of topology hiding of IP - PSTN calls.
running env:
Linux b2bua 2.6.18 #6 SMP Thu Nov 16 10:35:45 EST 2006 i686 Intel(R) Xeon(R) CPU 5140 @ 2.33GHz GNU/Linux Freeswitch version: FreeSwitch Version 1.0.pre1 (6220M)
The modules I am using are 711 and 729 codecs, sip, directory, xml_rpc, cdr, xml_cdr, console and syslog.
I am seeing what appears to be a memory leak. Consumed memory continues to drop linearly based on number of calls. Also had one core dump.
Can you provide details of the call scenarios you are testing, call rates, number of concurrent calls and any other details?
What can I do to help troubleshoot these issues?
Today I plan to stop using the CDR, rpc and logging modules to see if it makes a difference.
We have tested quite heavily with mod_console running, that should not be an issue.
Also considering running on one of my fedora core 7 boxes instead of
gentoo, again to see if anything changes.
We have had repeated issues with gentoo due to patches made either to gcc or to libc or the kernel. It is the only os/distro that I highly suggest never using.
Mike
_______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Caught signal 11 for unmapped thread! On 11/23/07, Freeswitch Freeswitch <freeswitch@2ride.com> wrote:
Hello,
Now I have updated to the latest trunk as of Nov 21 and moved to a fedora 7 system.
Still chasing random crashes of FS. First time this happened last night there were around 1000 sessions up. Crit error was logged referring to a signal caught on an unmapped thread. I wasn't able to catch the specific signal in the console before it scrolled off. This morning I had another crash. This time around 200 sessions running. There was no logging reported during the second crash.
Calling scenario is about as simple as it gets.
gw1 -> freeswitch -> gw2 all sip codecs, 729 and 711 transcoding off
I have been playing with the various rtp knobs in the sofia conf. Sometimes running with rewriting rtp timerstamps, sometimes not. Also tried with soft timer and off. ( can someone point me to an explanation of what the soft timer does? ).
Also notable that I compiled the code with debug symbols.
No core has been generated during crashes. Is there some way to enable core dumps?
Overall I am very impressed with FS - would love to get the cause of these crashes nailed so I can press forward.
Any help is very much appreciated.
Best regards,
Tom
On 11/19/07, Michael Jerris < mike@jerris.com> wrote:
On 11/19/07, Freeswitch Freeswitch <freeswitch@2ride.com> wrote:
Hello Freeswitch'rs,
I recently began using the latest freeswitch beta code as a pure B2BUA for the purpose of topology hiding of IP - PSTN calls.
running env:
Linux b2bua 2.6.18 #6 SMP Thu Nov 16 10:35:45 EST 2006 i686 Intel(R) Xeon(R) CPU 5140 @ 2.33GHz GNU/Linux Freeswitch version: FreeSwitch Version 1.0.pre1 (6220M)
The modules I am using are 711 and 729 codecs, sip, directory, xml_rpc, cdr, xml_cdr, console and syslog.
I am seeing what appears to be a memory leak. Consumed memory continues to drop linearly based on number of calls. Also had one core dump.
Can you provide details of the call scenarios you are testing, call rates, number of concurrent calls and any other details?
What can I do to help troubleshoot these issues?
Today I plan to stop using the CDR, rpc and logging modules to see if it makes a difference.
We have tested quite heavily with mod_console running, that should not be an issue.
Also considering running on one of my fedora core 7 boxes instead of
gentoo, again to see if anything changes.
We have had repeated issues with gentoo due to patches made either to gcc or to libc or the kernel. It is the only os/distro that I highly suggest never using.
Mike
_______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE: http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
Morning all - Where's mod-cdr gone? Cheers -- Dave
svn co http://svn.freeswitch.org/svn/unsupported/mod_cdr /b On Nov 23, 2007, at 11:27 PM, David Knell wrote:
Morning all -
Where's mod-cdr gone?
Cheers --
Dave
_______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
participants (4)
-
Brian West
-
David Knell
-
Freeswitch Freeswitch
-
Michael Jerris