[SR-Users] out of shm without any visible reason

[Juha Heinanen]

K reported during about 90 sec period that it is out of shared memory:

Feb 28 09:47:28 rox1 /usr/bin/sip-proxy[19725]: ERROR: tm 
[t_hooks.c:136]: insert_tmcb(): out of shm. mem
Feb 28 09:47:28 rox1 /usr/bin/sip-proxy[19725]: ERROR: acc 
[acc_logic.c:394]: acc_onreq(): cannot register additional callbacks
Feb 28 09:47:28 rox1 /usr/bin/sip-proxy[19725]: ERROR: <core> 
[core/sip_msg_clone.c:499]: sip_msg_shm_clone(): could not allocate 
shared memory from shm pool
Feb 28 09:47:28 rox1 /usr/bin/sip-proxy[19725]: ERROR: tm 
[t_lookup.c:1293]: new_t(): out of mem:
Feb 28 09:47:28 rox1 /usr/bin/sip-proxy[19725]: ERROR: tm 
[t_lookup.c:1439]: t_newtran(): new_t failed
...
Feb 28 09:47:29 rox1 /usr/bin/sip-proxy[19725]: ERROR: <core> 
[core/mem/q_malloc.c:297]: qm_find_free(): qm_find_free(0x7f1f2a506000, 
1232); Free fragment not found!
Feb 28 09:47:29 rox1 /usr/bin/sip-proxy[19725]: ERROR: <core> 
[core/mem/q_malloc.c:434]: qm_malloc(): qm_malloc(0x7f1f2a506000, 1232); 
Free fragment not found!
Feb 28 09:47:29 rox1 /usr/bin/sip-proxy[19725]: ERROR: tm 
[t_reply.c:1957]: relay_reply(): cannot alloc reply shmem
Feb 28 09:47:29 rox1 /usr/bin/sip-proxy[19725]: ERROR: <core> 
[core/sip_msg_clone.c:499]: sip_msg_shm_clone(): could not allocate 
shared memory from shm pool
Feb 28 09:47:29 rox1 /usr/bin/sip-proxy[19725]: ERROR: acc 
[acc_logic.c:562]: acc_onreply(): failed to clone the request - acc aborted
...
Feb 28 09:48:51 rox1 /usr/bin/sip-proxy[19724]: ERROR: <core> 
[core/mem/q_malloc.c:297]: qm_find_free(): qm_find_free(0x7f1f2a506000, 
5728); Free fragment not found!
Feb 28 09:48:51 rox1 /usr/bin/sip-proxy[19724]: ERROR: <core> 
[core/mem/q_malloc.c:434]: qm_malloc(): qm_malloc(0x7f1f2a506000, 5728); 
Free fragment not found!
Feb 28 09:48:51 rox1 /usr/bin/sip-proxy[19724]: ERROR: tm 
[t_lookup.c:1293]: new_t(): out of mem:
Feb 28 09:48:51 rox1 /usr/bin/sip-proxy[19724]: ERROR: tm 
[t_lookup.c:1439]: t_newtran(): new_t failed

And after that period it started working normally again and 
core.shmmem showed:

{
     total: 134217728
     free: 124173808
     used: 9530960
     real_used: 10043920
     max_used: 134217728
     fragments: 349
}

As can be seen, shm had been full, but normally only about 10% of it
is in use.

Syslog does not show any traffic spikes or other unusual activity before
the memory got full.

Any ideas what could cause such a high memory usage or could there be
a bug in shm management?

-- Juha