Summary: | ASTERISK-08053: High volume calling results in Asterisk core dump | ||
Reporter: | Matt Florell (mflorell) | Labels: | |
Date Opened: | 2006-11-02 12:24:00.000-0600 | Date Closed: | 2006-11-16 12:32:45.000-0600 |
Priority: | Critical | Regression? | No |
Status: | Closed/Complete | Components: | Core/General |
Versions: | Frequency of Occurrence | ||
Related Issues: | |||
Environment: | Attachments: | ( 0) crash_gdb_2006-10-19_1.2.12.1.txt ( 1) crash_gdb_2006-10-27_1.2.7.1.sv.txt ( 2) crash_gdb_2006-11-01_1.2.13.txt ( 3) crash_gdb_2006-11-01_1.2.7.1.sv.txt ( 4) crash_gdb-valgrind_2006-11-02_1.2.13.txt | |
Description: | I am testing a quad processor, dual core(eight cores total) server with VICIDIAL and I am getting a core dump from 2-10 hours into high-volume calling(20,000-30,000 calls per hour). These calls are mostly going through Local/ channels and many go into one of 100 meetme rooms. The channels are SIP and Zap. I have 4 backtraces from four different crashes on four different versions of Asterisk. Each crashed more than once, but I only kept one GDB backtrace. | ||
Comments: | By: Matt Florell (mflorell) 2006-11-02 12:26:27.000-0600 I forgot to mention that there are usually 400-600 Asterisk channels in use when the crash occurs. By: Serge Vecher (serge-v) 2006-11-09 08:20:17.000-0600 wow, that's some volume you've got going there. Just for general reference, have you tested 1.4-beta3? By: Matt Florell (mflorell) 2006-11-09 09:02:47.000-0600 FastAGI(which I need to do these tests at this level) does not work with Asterisk 1.4b3 (it is fixed in SVN 1.4) and I had done some tests and had other previously documented bug crashes with 1.4b3. I was going to wait for another 1.4beta release before trying it again. By: Joshua C. Colp (jcolp) 2006-11-16 12:32:45.000-0600 Should be fixed in 1.2 as of revision 47750, 1.4 as of revision 47751, and trunk as of revision 47752. |