segfault v 2.2 on reactor init #779
Comments
|
All log: Feb 2 11:05:51 front--voice-4 opensips: INFO:core:fix_poll_method: using epoll_lt as the IO watch method (auto detected) |
|
@cepehutu , could you extract a backtrace from the core file ? |
|
Couldn't find core file. LOG removed as long and irrelevant ( @bogdan-iancu ) |
|
The strace log is not helpfull here. But I see you get a core file: Feb 2 11:05:51 front--voice-4 /-voice/sbin/opensips[59731]: INFO:core:handle_sigs: child process 59738 exited by a signal 11 So, look for the core file into the working dir of opensips try the / or /root or home dir. |
|
Yeap. Core generated with strace only: Feb 2 11:49:17 front--voice-4 opensips_log[10457]: INFO:core:handle_sigs: child process 10473 exited by a signal 11 Without strace: Core tar at url: |
|
@cepehutu , hard to get info from a core file generated on another machine. Please extract the backtrace on your server by using gdb /path/to/opensips /path/to/core and run "bt full" |
|
like this? 13:52+00 root@front--voice-4:/home/ilhovsky]# gdb /-voice/sbin/opensips /core.* bt full |
|
@cepehutu , once inside gdb, run "bt full" and paste the output. |
|
(gdb) bt full |
|
2.1 version started good at same config. |
|
@cepehutu , it is really strange, I cannot reproduce it. |
|
Problem on 2.2 branch,which marked a dev.version: opensips 2.2-dev (x86_64/linux) wri Machine on dmz zone, only local access. |
|
ok, let's try some remote debugging. In gdb, run "p *h" . Post only the output of that command. I assume you will do it over the same core file as above. |
|
(gdb) p *h next? |
|
p _worker_io |
|
(gdb) p _worker_io |
|
ups.....fd_hash is NULL there 😲 |
|
What next? |
|
first, please recompile everything: |
|
make clean Feb 2 16:51:48 front--voice-4 opensips_log[46356]: NOTICE:usrloc:preload_udomain: load start time [1454431908] |
|
|
ok, could you please apply this patch and provide the logs again: |
|
applied patch. Feb 8 13:09:05 front--voice-4 kernel: opensips[28159]: segfault at 70 ip 00000000005669da sp 00007ffc79083430 error 4 in opensips[400000+1df000] have no core fles :( |
|
@cepehutu , I do not see the logs I inserted into the code - be sure you run in debug level 3 and that you copy here the whole set of logs from opensips starting attempt. |
|
Hi @cepehutu , if you do not have any update on this, I will have to close the report as invalid (so far, nobody else was reported this really strange error. |
|
Closing as there is no follow up here. Also the reported issue seems to be related to some particularities/local issues of the built machine - there is no other similar report and the C code does not confirm such an crash. |
eb 2 11:02:05 front--voice-4 kernel: opensips[59678]: segfault at 70 ip 000000000056655a sp 00007ffc3153d5f0 error 4 in opensips[400000+1df000]
Feb 2 11:02:05 front--voice-4 kernel: opensips[59682]: segfault at 70 ip 000000000056655a sp 00007ffc3153d5f0 error 4 in opensips[400000+1df000]
Feb 2 11:02:05 front--voice-4 opensips_log[59670]: INFO:core:handle_sigs: child process 59678 exited by a signal 11
Feb 2 11:02:05 front--voice-4 opensips_log[59670]: INFO:core:handle_sigs: core was not generated
Feb 2 11:02:05 front--voice-4 opensips_log[59670]: INFO:core:handle_sigs: child process 59682 exited by a signal 11
Feb 2 11:02:05 front--voice-4 opensips_log[59670]: INFO:core:handle_sigs: core was not generated
Feb 2 11:02:05 front--voice-4 opensips_log[59670]: INFO:core:handle_sigs: terminating due to SIGCHLD
version: opensips 2.2-dev (x86_64/linux)
flags: STATS: On, EXTRA_DEBUG, DISABLE_NAGLE, USE_MCAST, SHM_MMAP, PKG_MALLOC, F_MALLOC, DBG_MALLOC, FAST_LOCK-ADAPTIVE_WAIT
ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16, MAX_URI_SIZE 1024, BUF_SIZE 65535
poll method support: poll, epoll_lt, epoll_et, sigio_rt, select.
main.c compiled on 10:29:53 Feb 2 2016 with gcc 4.4.7
The text was updated successfully, but these errors were encountered: