Hi,
Have any of you folks seen an abnormal CPU usage issue with rtpprpxy? At times, "top" shows rtpproxy uses 100% CPU (sometimes 99%, sometimes even 101%!), and it stops responding to openser. I have to restart everything now.
A small thread in Serdev list recommends upgrading rtpproxy to latest in CVS:
http://lists.iptel.org/pipermail/serdev/2006-June/007459.html
Mine was fairly old, so I upgraded, but without much progress. What really bugs is I don't know what triggers it.
Backtraces suggest that this occurs in different places in rtpproxy, before and after upgrading.
Before upgrade:
(gdb) where #0 0xb7f49410 in ?? () #1 0xbffd6d88 in ?? () #2 0x00000076 in ?? () #3 0xbffd2410 in ?? () #4 0x4a8d9fd1 in sendto () from /lib/tls/libc.so.6 #5 0x0804ac0c in main (argc=7, argv=0xbffd6e14) at main.c:1467
After upgrade:
(gdb) where #0 0xb7f0a410 in ?? () #1 0xbfcb0258 in ?? () #2 0xbfcadfc0 in ?? () #3 0xbfcab8e0 in ?? () #4 0x00bb8dd1 in recvfrom () from /lib/tls/libc.so.6 #5 0x0804a568 in main (argc=7, argv=0xbfcb02e4) at main.c:1364
Any clues? Is this a known problem and is there a patch or some sort ofworkaround? Is there anything I need to upgrade? (This is CentOS 4.4, FWIW) Anything that should be done in openser configuration?
Thanks much in advance.
- Sajith.
Sajith T S wrote:
Backtraces suggest that this occurs in different places in rtpproxy, before and after upgrading.
Oops, my mistake. In both these cases, rtpproxy seems to be in a loop, actually - it is not really "stuck" in one place, so those backtraces are not exactly correct.
In any case, I am still looking for a clue.
- Sajith.
Hi Sajith,
I encountered something similar when just changing the kernel version (linux). According to my tests, rtpproxy was blocking in recvfrom() - poll was reporting the IN event on the socket, but recvfrom blocked on READ.....
I was not able to find a solution to it...
regards, bogdan
Sajith T S wrote:
Hi,
Have any of you folks seen an abnormal CPU usage issue with rtpprpxy? At times, "top" shows rtpproxy uses 100% CPU (sometimes 99%, sometimes even 101%!), and it stops responding to openser. I have to restart everything now.
A small thread in Serdev list recommends upgrading rtpproxy to latest in CVS:
http://lists.iptel.org/pipermail/serdev/2006-June/007459.html
Mine was fairly old, so I upgraded, but without much progress. What really bugs is I don't know what triggers it.
Backtraces suggest that this occurs in different places in rtpproxy, before and after upgrading.
Before upgrade:
(gdb) where #0 0xb7f49410 in ?? () #1 0xbffd6d88 in ?? () #2 0x00000076 in ?? () #3 0xbffd2410 in ?? () #4 0x4a8d9fd1 in sendto () from /lib/tls/libc.so.6 #5 0x0804ac0c in main (argc=7, argv=0xbffd6e14) at main.c:1467
After upgrade:
(gdb) where #0 0xb7f0a410 in ?? () #1 0xbfcb0258 in ?? () #2 0xbfcadfc0 in ?? () #3 0xbfcab8e0 in ?? () #4 0x00bb8dd1 in recvfrom () from /lib/tls/libc.so.6 #5 0x0804a568 in main (argc=7, argv=0xbfcb02e4) at main.c:1364
Any clues? Is this a known problem and is there a patch or some sort ofworkaround? Is there anything I need to upgrade? (This is CentOS 4.4, FWIW) Anything that should be done in openser configuration?
Thanks much in advance.
- Sajith.
Users mailing list Users@openser.org http://openser.org/cgi-bin/mailman/listinfo/users
-
Bogdan-Andrei Iancu -
Sajith T S -
Sajith T S