Summary: | ASTERISK-08938: Crash under load from multiple SIP calls | ||
Reporter: | Rus Rus (harbour) | Labels: | |
Date Opened: | 2007-03-05 09:39:13.000-0600 | Date Closed: | 2008-03-07 15:13:21.000-0600 |
Priority: | Critical | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ( 0) all_bt_full ( 1) bt_full ( 2) bt.2 ( 3) bt.SVN-branch-1.4-r105409 ( 4) full ( 5) threads | |
Description: | executing 'sipp -r 30 -l 30' got segmentation faults. traces attached ****** ADDITIONAL INFORMATION ****** can supply any additional info | ||
Comments: | By: Joshua C. Colp (jcolp) 2007-03-05 12:01:23.000-0600 Please confirm that your build is unoptimized by going into menuselect via make menuselect, then Compiler Flags, and check that DONT_OPTIMIZE is set. If not then set it, recompile, and retry your tests. We will also need to know more about what these calls are doing, the dialplan, and console output. By: Rus Rus (harbour) 2007-03-06 02:15:35.000-0600 Attaching crashes of the unoptimized build. Asterisk and kernel is compiled by gcc-3.4.6, running Slackware-11.0 By: Rus Rus (harbour) 2007-03-06 02:27:06.000-0600 Simple setup is : ISDN PBX NT mode --- PRI E1 ---> Asterisk with single HFC-E1 card <--- sipp The PBX has phone group with the same number, this number is dialed by sipp. Dialplan is : .... exten => _6600XX,1,Dial(mISDN/1/${EXTEN}) exten => _6600XX,2,Congestion exten => _6600XX,3,Hangup .... As there is only 30 lines to PBX, they are occupied one by one. When all E1 lines are busy asterisk has to reject calls from sipp. When some of the lines to PBX become free - asterisk can pass sipp call to PBX again. The rate parameter for sipp can be lower, but the time needed for crash to occure, when this parameter (rate > 30) && (rate < 100) - the crash occured after severel seconds of the test. My hardware is ASUS M2NPV-VM with Athlon X2 4800 CPU, 2GB RAM. I've noticed that such crashes can't be reproduced on single CPU machine, only SMP variant triggers this bug. By: Russell Bryant (russell) 2007-06-06 12:02:18 This should be fixed in 1.2, 1.4, and trunk in revisions 67717, 67716, and 67715. Let us know if you have any further problems. Thanks! By: Rus Rus (harbour) 2008-02-29 07:46:03.000-0600 Tried Asterisk SVN-branch-1.4-r104334 - sigsegv in another place with the same test. By: Russell Bryant (russell) 2008-02-29 17:11:11.000-0600 this is a duplicate of 11573 By: Rus Rus (harbour) 2008-03-01 01:44:18.000-0600 SIGSEGV log for SVN-branch-1.4-r105409 (optimized one) attached. If needed I can post DONT_OPTIMIZE'ed version of the backtrace. By: Russell Bryant (russell) 2008-03-07 15:12:58.000-0600 As I said before, this issue is a duplicate. Please post further info there |