Hello,
sqlops is using pkg and tm shm, so they should not be directly related, but maybe in the way config file works.
Can you run it again with memlog lower than debug and see where the allocated (not-freed) chunks were done? It should appear soon, not waiting for out of mem message.
If you haven't restarted, there is a way to attach with gdb and walk through shm allocated chunks to spot the occurences.
Cheers, Daniel
On 10/4/11 12:08 PM, Klaus Darilion wrote:
Hi!
I recently had a problem with Kamailio 3.1.4 (provided Debian packages):
I had some DB problems (missing tables). Thus, the timer module failed to insert the statistics (for siremis):
ERROR: db_mysql [km_dbase.c:120]: driver error on query: Table 'kamailio.statistics_tmx' doesn't exist ERROR: <core> [db_query.c:130]: error while submitting query ERROR: sqlops [sql_api.c:217]: cannot do the query
This happened for some time (weeks?), other DB queries were unaffected.
Then, suddenly Kamailio ran out of memory:
ERROR: <core> [sip_msg_clone.c:506]: ERROR: sip_msg_cloner: cannot allocate memory ERROR: tm [t_lookup.c:1338]: ERROR: new_t: out of mem: ERROR: tm [t_lookup.c:1478]: ERROR: t_newtran: new_t failed ERROR: sl [sl_funcs.c:282]: ERROR: sl_reply_error used: I'm terribly sorry, server error occurred (1/SL)
It do not think it is a load problem as the server is more or less idle. May it be a memory leak due to wrong error handling?
Where there any fixes recently?
Thanks Klaus
SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list sr-users@lists.sip-router.org http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users