Production system - Freepbx 13.0.67 with Asterisk 11.19. Newly built on a physical server, ran for several weeks then was converted using vmware conversion app. The conversion was almost 2 months ago and these crashes began a week ago.
I first noticed this happening when a fax came in. The log message would indicate: T.38 timeout and then Asterisk would restart (dropping calls). The core backtrace shows this happening in res_fax_digium.so. I checked that the correct version was installed, ran benchfax again and re-installed the module.
(gdb) bt
#0 0x00007f5b7d9f166e in ?? () from /usr/lib64/asterisk/modules/res_fax_digium.so
#1 0x00007f5b7d9c98db in ?? () from /usr/lib64/asterisk/modules/res_fax_digium.so
#2 0x00007f5b7d9c9ccb in ?? () from /usr/lib64/asterisk/modules/res_fax_digium.so
#3 0x0000000000000000 in ?? ()
(gdb)
One interesting thing to note - a fax tried to come in and asterisk crashed. After it restarted the fax retried and this time it came in successfully. So this is not fully repeatable and seems to happen more when the system has been up a few days.
Yesterday the crash happened without a fax. Backtrace below.
(gdb) bt
#0 0x000000000049ed3b in ?? ()
#1 0x000000000049f0f5 in ?? ()
#2 0x00000000004a0c02 in ast_cli_command_full ()
#3 0x00000000004a0e7e in ast_cli_command_multiple_full ()
#4 0x00000000004432b5 in ?? ()
#5 0x0000000000597bdc in ?? ()
#6 0x00007f8c9f9919d1 in start_thread (arg=0x7f8c3da60700) at pthread_create.c:301
#7 0x00007f8ca0f5e8fd in clone () at …/sysdeps/unix/sysv/linux/x86_64/clone.S:115
The vm has 4GB memory and never gets anywhere close. All other statistics such as disk latency, etc are very acceptable so while I’m open to a vm-related issue I’m not really leaning that way. I could use some ideas. I searched the forums with no real clues.