crash from kamailio could be cnxcc related

[dunst0]

I was trying to set up kamailio with cnxcc, and suddenly I got this:

Nov 23 13:32:02 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=15.000000, ca2=14.000000
Nov 23 13:32:03 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=16.000000, ca2=15.000000
Nov 23 13:32:04 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=17.000000, ca2=16.000000
Nov 23 13:32:05 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=18.000000, ca2=17.000000
Nov 23 13:32:06 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=19.000000, ca2=18.000000
Nov 23 13:32:07 sip42 kamailio[4744]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=20.000000, ca2=19.000000
Nov 23 13:32:07 sip42 sems[14299]: [#7f117d358700] [onRxReqSanity, AmBasicSipDialog.cpp:293] INFO:  remote cseq lower than previous ones - refusing request
Nov 23 13:32:07 sip42 kamailio[4744]: INFO: <script>: [b6aps.CcAJpV6zEGbgAPuyvg6N6irS7l]: call killed
Nov 23 13:32:12 sip42 kamailio[4733]: : <core> [mem/q_malloc.c:121]: qm_debug_frag(): BUG: qm_*: fragm. 0x7f21b7731498 (address 0x7f21b77314c8) beginning overwritten(0)!
Nov 23 13:32:12 sip42 kamailio[4757]: CRITICAL: <core> [pass_fd.c:275]: receive_fd(): EOF on 39
Nov 23 13:32:12 sip42 kamailio[4698]: ALERT: <core> [main.c:728]: handle_sigs(): child process 4733 exited by a signal 6
Nov 23 13:32:12 sip42 kamailio[4698]: ALERT: <core> [main.c:731]: handle_sigs(): core was not generated
Nov 23 13:32:12 sip42 kamailio[4698]: INFO: <core> [main.c:743]: handle_sigs(): terminating due to SIGCHLD
Nov 23 13:32:12 sip42 kamailio[4757]: INFO: <core> [main.c:798]: sig_usr(): signal 15 received
Nov 23 13:32:12 sip42 kamailio[4754]: INFO: <core> [main.c:798]: sig_usr(): signal 15 received
Nov 23 13:32:12 sip42 kamailio[4753]: INFO: <core> [main.c:798]: sig_usr(): signal 15 received
Nov 23 13:32:12 sip42 kamailio[4698]: WARNING: tm [h_table.c:130]: free_cell_helper(): removed cell 0x7f21b77319f0 is still linked in hash table (h_table.c:440)
Nov 23 13:32:12 sip42 kamailio[4698]: : <core> [mem/q_malloc.c:121]: qm_debug_frag(): BUG: qm_*: fragm. 0x7f21b7731498 (address 0x7f21b77314c8) beginning overwritten(0)!

I was not able to reproduce it:

Nov 23 13:40:55 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=1.000000, ca2=0.000000
Nov 23 13:40:56 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=3.000000, ca2=1.000000
Nov 23 13:40:57 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=4.000000, ca2=3.000000
Nov 23 13:40:58 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=5.000000, ca2=4.000000
Nov 23 13:40:59 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=6.000000, ca2=5.000000
Nov 23 13:41:00 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=7.000000, ca2=6.000000
Nov 23 13:41:01 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=8.000000, ca2=7.000000
Nov 23 13:41:02 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=9.000000, ca2=8.000000
Nov 23 13:41:03 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=10.000000, ca2=9.000000
Nov 23 13:41:04 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=11.000000, ca2=10.000000
Nov 23 13:41:05 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=12.000000, ca2=11.000000
Nov 23 13:41:06 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=13.000000, ca2=12.000000
Nov 23 13:41:07 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=14.000000, ca2=13.000000
Nov 23 13:41:08 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=15.000000, ca2=14.000000
Nov 23 13:41:09 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=16.000000, ca2=15.000000
Nov 23 13:41:10 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=17.000000, ca2=16.000000
Nov 23 13:41:11 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=18.000000, ca2=17.000000
Nov 23 13:41:12 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=19.000000, ca2=18.000000
Nov 23 13:41:13 sip42 kamailio[4813]: INFO: cnxcc [cnxcc_check.c:93]: check_calls_by_money(): ec=0.000000, ca=20.000000, ca2=19.000000
Nov 23 13:41:13 sip42 sems[14299]: [#7f117d459700] [onRxReqSanity, AmBasicSipDialog.cpp:293] INFO:  remote cseq lower than previous ones - refusing request
Nov 23 13:41:13 sip42 kamailio[4813]: INFO: <script>: [CEMClRtq-0bIDGKdphgjaeGK5KStdRhL]: call killed

[miconda]

It looks like a buffer overflow (writing more than allocating). What is the exact version you are running (kamailio -v)?

Before starting kamailio, run:

ulimit -c unlimited

Eventually run also kamailio as root in order to get the core file.

If you didn't configure one core file per process, do:

echo "1" > /proc/sys/kernel/core_uses_pid

Once you get core files, send the 'bt full' from gdb.

[dunst0]

Using version: kamailio 4.3.3 (x86_64/linux) 2fe732
Trying to reproduce the crash but was not successful until now.

[dunst0]

I close this issue cause I am not able to reproduce it.