Freeswitch core dump after upgrade to latest version
Hello, We are running 1.0.5pre10 for a while, and today I tried to move to the latest tarball (from January 22nd). The software crashes with a core dump after a few seconds. The core dump. The two relevant lines (to my opinion) are: #4 0xb7547d2d in nua_bye_server_report (sr=0xb6c910b0, tags=0x0) at nua_session.c:3938 #5 0xb7541bb3 in nua_server_report (sr=0x6) at nua_server.c:643 line 3938 is an assert() statement. Any idea? Thanks! __Yehavi:
Please collect the sip trace, console logs and everything you can up till the crash. Do not make decisions on what you think is relevant collect everything you can and open a jira. Also did you do a fresh checkout? or a make current? Thanks, Brian On Jan 23, 2010, at 11:06 AM, Yehavi Bourvine wrote:
Hello,
We are running 1.0.5pre10 for a while, and today I tried to move to the latest tarball (from January 22nd). The software crashes with a core dump after a few seconds. The core dump. The two relevant lines (to my opinion) are:
#4 0xb7547d2d in nua_bye_server_report (sr=0xb6c910b0, tags=0x0) at nua_session.c:3938 #5 0xb7541bb3 in nua_server_report (sr=0x6) at nua_server.c:643 line 3938 is an assert() statement. Any idea?
Thanks! __Yehavi:
Could you confirm that you have an issue described in http://jira.freeswitch.org/browse/SFSIP-197 ? Seems like you're not the only unlucky... On Saturday 23 January 2010, Yehavi Bourvine wrote:
Hello,
We are running 1.0.5pre10 for a while, and today I tried to move to the latest tarball (from January 22nd). The software crashes with a core dump after a few seconds. The core dump. The two relevant lines (to my opinion) are:
#4 0xb7547d2d in nua_bye_server_report (sr=0xb6c910b0, tags=0x0) at nua_session.c:3938 #5 0xb7541bb3 in nua_server_report (sr=0x6) at nua_server.c:643 line 3938 is an assert() statement. Any idea?
Thanks! __Yehavi:
Yes, it looks exactly the same, with the same code in retval. It happens just when an incoming INVITE arrives. Since it has already a jira issue opened, do I still have to provide the traces? About how I upgrade: I've downloded the tarball of the latest version into a fresh directory, built it, and in order to install it: deleted everyhting in bin, mod and lib, and then made "make install". Thanks, __Yehavi: 2010/1/23 Sergey Okhapkin <sos@sokhapkin.dyndns.org>
Could you confirm that you have an issue described in http://jira.freeswitch.org/browse/SFSIP-197 ?
Seems like you're not the only unlucky...
On Saturday 23 January 2010, Yehavi Bourvine wrote:
Hello,
We are running 1.0.5pre10 for a while, and today I tried to move to the latest tarball (from January 22nd). The software crashes with a core dump after a few seconds. The core dump. The two relevant lines (to my opinion) are:
#4 0xb7547d2d in nua_bye_server_report (sr=0xb6c910b0, tags=0x0) at nua_session.c:3938 #5 0xb7541bb3 in nua_server_report (sr=0x6) at nua_server.c:643 line 3938 is an assert() statement. Any idea?
Thanks! __Yehavi:
_______________________________________________ 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
Add your traces and all information you can provide to http://jira.freeswitch.org/browse/SFSIP-197 , the more information developers will get, the sooner they will be able to find and fix the issue. I wish my FS crash "in a few seconds" like yours, then I will be able to debug myself, but it crashes very rarely... On Saturday 23 January 2010, Yehavi Bourvine wrote:
Yes, it looks exactly the same, with the same code in retval. It happens just when an incoming INVITE arrives. Since it has already a jira issue opened, do I still have to provide the traces?
About how I upgrade: I've downloded the tarball of the latest version into a fresh directory, built it, and in order to install it: deleted everyhting in bin, mod and lib, and then made "make install".
Thanks, __Yehavi:
2010/1/23 Sergey Okhapkin <sos@sokhapkin.dyndns.org>
Could you confirm that you have an issue described in http://jira.freeswitch.org/browse/SFSIP-197 ?
Seems like you're not the only unlucky...
On Saturday 23 January 2010, Yehavi Bourvine wrote:
Hello,
We are running 1.0.5pre10 for a while, and today I tried to move to the latest tarball (from January 22nd). The software crashes with a core dump after a few seconds. The core dump. The two relevant lines (to my
opinion)
are:
#4 0xb7547d2d in nua_bye_server_report (sr=0xb6c910b0, tags=0x0) at nua_session.c:3938 #5 0xb7541bb3 in nua_server_report (sr=0x6) at nua_server.c:643 line 3938 is an assert() statement. Any idea?
Thanks! __Yehavi:
_______________________________________________ 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
Yes please provide traces the more information we have the clearer the picture and possibility of fixing it. /b On Jan 23, 2010, at 12:02 PM, Yehavi Bourvine wrote:
Yes, it looks exactly the same, with the same code in retval. It happens just when an incoming INVITE arrives. Since it has already a jira issue opened, do I still have to provide the traces?
About how I upgrade: I've downloded the tarball of the latest version into a fresh directory, built it, and in order to install it: deleted everyhting in bin, mod and lib, and then made "make install".
Thanks, __Yehavi:
Just uploaded. It happens when I call our voicemail number and disconnects the call. The voicemail application answers, and I disconnect the call the Freeswitch crash. Thanks, __Yehavi: 2010/1/23 Brian West <brian@freeswitch.org>
Yes please provide traces the more information we have the clearer the picture and possibility of fixing it.
/b
On Jan 23, 2010, at 12:02 PM, Yehavi Bourvine wrote:
Yes, it looks exactly the same, with the same code in retval. It happens just when an incoming INVITE arrives. Since it has already a jira issue opened, do I still have to provide the traces?
About how I upgrade: I've downloded the tarball of the latest version into a fresh directory, built it, and in order to install it: deleted everyhting in bin, mod and lib, and then made "make install".
Thanks, __Yehavi:
_______________________________________________ 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
GOOD! You have a consistent way to reproduce the problem! On Saturday 23 January 2010, Yehavi Bourvine wrote:
Just uploaded. It happens when I call our voicemail number and disconnects the call. The voicemail application answers, and I disconnect the call the Freeswitch crash.
Thanks, __Yehavi:
2010/1/23 Brian West <brian@freeswitch.org>
Yes please provide traces the more information we have the clearer the picture and possibility of fixing it.
/b
On Jan 23, 2010, at 12:02 PM, Yehavi Bourvine wrote:
Yes, it looks exactly the same, with the same code in retval. It happens
just when an incoming INVITE arrives.
Since it has already a jira issue opened, do I still have to provide the
traces?
About how I upgrade: I've downloded the tarball of the latest version
into a fresh directory, built it, and in order to install it:
deleted everyhting in bin, mod and lib, and then made "make install".
Thanks, __Yehavi:
_______________________________________________ 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
I have looked into the code and backtrace related to this issue, and the best I can see, its impossible... We will definitely need access to a box this is reproducible on to fix this issue. Mike On Jan 23, 2010, at 2:02 PM, Brian West wrote:
Best to find us on IRC when anthm is around and lets get into your box and fix this.
/b
On Jan 23, 2010, at 12:55 PM, Sergey Okhapkin wrote:
GOOD! You have a consistent way to reproduce the problem!
About getting into my box: It is a production machine with quite tight access control, so this is close to impossible... I am trying to reproduce it on a backup system. If it works, then I might be able to give access to it. I am working on it now... __Yehavi: 2010/1/23 Brian West <brian@freeswitch.org>
Best to find us on IRC when anthm is around and lets get into your box and fix this.
/b
On Jan 23, 2010, at 12:55 PM, Sergey Okhapkin wrote:
GOOD! You have a consistent way to reproduce the problem!
_______________________________________________ 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
I've tried printing the value of retval and it is 1 (and not 6...). BTW, after adding the printf()'s I cannot examine anymore the value of retval in gdb as it is optimized away. Maybe this is the reason for a value of 6 which does not exist in the code that returns it. nua_base_server_report() returns one as "terminated" variable is false. Hope this helps. Regards, __Yehavi: 2010/1/24 Yehavi Bourvine <yehavi.bourvine@gmail.com>
About getting into my box: It is a production machine with quite tight access control, so this is close to impossible... I am trying to reproduce it on a backup system. If it works, then I might be able to give access to it. I am working on it now...
__Yehavi:
2010/1/23 Brian West <brian@freeswitch.org>
Best to find us on IRC when anthm is around and lets get into your box and fix this.
/b
On Jan 23, 2010, at 12:55 PM, Sergey Okhapkin wrote:
GOOD! You have a consistent way to reproduce the problem!
_______________________________________________ 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
Yes, ssh is best. Are you able to meet up with me on irc so we can discuss this issue? On Jan 24, 2010, at 9:42 AM, Yehavi Bourvine <yehavi.bourvine@gmail.com> wrote:
I've tried printing the value of retval and it is 1 (and not 6...). BTW, after adding the printf()'s I cannot examine anymore the value of retval in gdb as it is optimized away. Maybe this is the reason for a value of 6 which does not exist in the code that returns it.
nua_base_server_report() returns one as "terminated" variable is false. Hope this helps.
Regards, __Yehavi: 2010/1/24 Yehavi Bourvine <yehavi.bourvine@gmail.com> About getting into my box: It is a production machine with quite tight access control, so this is close to impossible... I am trying to reproduce it on a backup system. If it works, then I might be able to give access to it. I am working on it now...
__Yehavi:
2010/1/23 Brian West <brian@freeswitch.org> Best to find us on IRC when anthm is around and lets get into your box and fix this.
/b
On Jan 23, 2010, at 12:55 PM, Sergey Okhapkin wrote:
GOOD! You have a consistent way to reproduce the problem!
_______________________________________________ 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 _______________________________________________ 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
Please get a full pcap trace console log and attach it. /b On Jan 23, 2010, at 12:38 PM, Yehavi Bourvine wrote:
Just uploaded. It happens when I call our voicemail number and disconnects the call. The voicemail application answers, and I disconnect the call the Freeswitch crash.
Thanks, __Yehavi:
participants (4)
-
Brian West
-
Michael Jerris
-
Sergey Okhapkin
-
Yehavi Bourvine