Bogdan,I do periodic cache stores also in the background. But that's a rather new thing I do now. Even before I was doing those I got these hangs.. I'll try to get another bt that shows it hanging but without the cache_store. However, I think the same underlying problem is affecting both.
On Tue, Sep 22, 2009 at 7:43 AM, Bogdan-Andrei Iancu <[email protected] > wrote: > Hi Brett, > > hmm...that is rather strange as the backtrace you show here definitely > is from the fifo process and the run command is clear to be the > "cache_store". > > Maybe there is some other app that (in the same time with your tests) do > run some caching commands?? > > BTW, can you simply run via fifo the the cache_store command via > memcached backend? Just to see if that is blocking or not. > > Regards, > Bogdan > > Brett Nemeroff wrote: > > Bogdan, > > Here's my bt full. Towards the top there is a cache_store line. The > > fifo operation I'm waiting on is a profile_get_values.. Interesting > > that I don't call cache_store in any of these hung opensipsctl calls. > > > > # ps ax|grep opensipsctl > > 17929 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_values trunk > > 17974 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_values trunk > > 17976 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_values trunk > > 27643 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_size inbound > > 27687 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_size inbound > > 27688 ? S 0:00 /bin/bash /usr/local/sbin/opensipsctl fifo > > profile_get_size inbound > > > > > > I took the mi_fifo ps for the gdb.. > > > > 0x00000034c1ac54f0 in __read_nocancel () from /lib64/libc.so.6 > > (gdb) bt full > > #0 0x00000034c1ac54f0 in __read_nocancel () from /lib64/libc.so.6 > > No symbol table info available. > > #1 0x00000034c1a6cbb7 in _IO_new_file_underflow () from /lib64/libc.so.6 > > No symbol table info available. > > #2 0x00000034c1a6d57e in _IO_default_uflow_internal () from > > /lib64/libc.so.6 > > No symbol table info available. > > #3 0x00000034c1a62664 in _IO_getline_info_internal () from > > /lib64/libc.so.6 > > No symbol table info available. > > #4 0x00000034c1a61509 in fgets () from /lib64/libc.so.6 > > No symbol table info available. > > #5 0x00002b68c3c2e619 in mi_read_line (b=0x79cd50 ":cache_store", > > max=1024, stream=0x1e81b110, read=0x7fffe834bad4) at fifo_fnc.c:255 > > retry_cnt = 0 > > len = <value optimized out> > > __FUNCTION__ = "mi_read_line" > > #6 0x00002b68c3c2f52f in mi_fifo_server (fifo_stream=0x1e81b110) at > > fifo_fnc.c:400 > > mi_cmd = (struct mi_root *) 0x7a1568 > > mi_rpl = (struct mi_root *) 0x7a17c0 > > hdl = (struct mi_handler *) 0x0 > > line_len = 35 > > file_sep = 0x79cd5c "" > > command = 0x79cd51 "cache_store" > > file = 0x79d160 "/tmp/opensips_receiver_1372" > > f = (struct mi_cmd *) 0x7995e8 > > reply_stream = (FILE *) 0x1e82e7e0 > > __FUNCTION__ = "mi_fifo_server" > > #7 0x00002b68c3c31001 in fifo_process (rank=<value optimized out>) at > > mi_fifo.c:212 > > fifo_stream = (FILE *) 0x1e81b110 > > __FUNCTION__ = "fifo_process" > > #8 0x000000000046b4bc in start_module_procs () at sr_module.c:585 > > m = (struct sr_module *) 0x76ca88 > > n = 0 > > l = 0 > > x = <value optimized out> > > __FUNCTION__ = "start_module_procs" > > #9 0x000000000042745d in main (argc=5, argv=0x7fffe834bcf8) at > main.c:823 > > cfg_log_stderr = <value optimized out> > > cfg_stream = (FILE *) 0x1e80c010 > > c = <value optimized out> > > r = 8 > > tmp = 0x7fffe834dbe4 "" > > tmp_len = <value optimized out> > > port = <value optimized out> > > proto = <value optimized out> > > ret = <value optimized out> > > seed = 2557210813 > > rfd = <value optimized out> > > __FUNCTION__ = "main" > > > > > > On Thu, Sep 17, 2009 at 11:10 AM, Bogdan-Andrei Iancu > > <[email protected] <mailto:[email protected]>> wrote: > > > > Hi Brett, > > > > if you get a hang, do try to attach with gdb to the fifo proc (you > can > > do a "fifo ps" in the beginning to see the PID of the fifo process). > > > > A backtrace of the fifo proc will help in understanding the issue. > > > > BTW - what about cpu load ? > > > > Regards, > > Bogdan > > > > Brett Nemeroff wrote: > > > Hello list, > > > I periodically call opensipsctl fifo profile_get_values <profile > > > name>. Manually, it always works great.. but called from cron every > > > minute, it frequently hangs indefinitely until I kill it. Often > > I run > > > ps ax and see about 5 of them in there. > > > > > > Running 1.5.3. Any ideas? > > > -Brett > > > > > > > > > ------------------------------------------------------------------------ > > > > > > _______________________________________________ > > > Users mailing list > > > [email protected] <mailto:[email protected]> > > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > > > > > _______________________________________________ > > Users mailing list > > [email protected] <mailto:[email protected]> > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Users mailing list > > [email protected] > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users >
_______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
