We are having setup like openser+radius+mediaproxy. Openser version is 1.2.3.
When we started to send high volume calls , we get below mentioned ERRORs & WARNINGs. 1. /usr/local/sbin/openser[29869]: ERROR:tm:t_forward_nonack: discarding fwd for a cancelled transaction /usr/local/sbin/openser[29869]: ERROR:tm:w_t_relay: t_forward_nonack failed WARNING: script writer didn't release transaction
2. WARNING:tm:run_failure_handlers: no UAC or CANCEL support (1, 0)
3. /usr/local/sbin/openser[29870]: ERROR:nathelper:nh_timer: out of memory
4. /usr/local/sbin/openser[29870]: convert_rows: No memory left (openser crashed)
As we were using default shared memory for openser, we increased shared memory to maximum(2GB) to fix out of bound issue. We checked openser script for memory leacks using valgrind, but did not find any memory leack.
During doing relay of CANCEL,in relay route t_on_failure was added. So we removed t_on_failure in relay route for CANCEL. So WARNING:tm:run_failure_handlers: no UAC or CANCEL support (1, 0) was fixed.
Still we are facing, 1. /usr/local/sbin/openser[29869]: ERROR:tm:t_forward_nonack: discarding fwd for a cancelled transaction /usr/local/sbin/openser[29869]: ERROR:tm:w_t_relay: t_forward_nonack failed WARNING: script writer didn't release transaction
2. /usr/local/sbin/openser[29870]: ERROR:nathelper:nh_timer: out of memory
3. ERROR:usrloc:get_all_db_ucontacts: raw_query failed
4. rc_avpair_gen: received unknown attribute 80 of length 18: 0xA40AA6A78F04D67B03C75F8CA233F008
Will increasing private memory will solve all issues?
Thanks in advance!!
-- Krunal Patel
On Tuesday 09 December 2008, Krunal Patel wrote:
We are having setup like openser+radius+mediaproxy. Openser version is 1.2.3.
Hi Krunal,
i'd suggest that you update to a more recent version of the server, if you use this in a production environment, e.g. 1.4.3 that will be released next week.
When we started to send high volume calls , we get below mentioned ERRORs & WARNINGs.
- /usr/local/sbin/openser[29869]: ERROR:tm:t_forward_nonack: discarding
fwd for a cancelled transaction
I think this is harmless. it says you are trying to forward a request which was already cancelled (this may happen due parallel processing of the INVITE and CANCEL request). (quoted from some earlier discussion)
/usr/local/sbin/openser[29869]: ERROR:tm:w_t_relay: t_forward_nonack
failed WARNING: script writer didn't release transaction
- WARNING:tm:run_failure_handlers: no UAC or CANCEL support (1, 0)
This are just warnings, you can ignore them. For the first one you need to release the transaction in the script, google for this error to get more informations.
/usr/local/sbin/openser[29870]: ERROR:nathelper:nh_timer: out of memory
/usr/local/sbin/openser[29870]: convert_rows: No memory left (openser
crashed)
As we were using default shared memory for openser, we increased shared memory to maximum(2GB) to fix out of bound issue.
This is not caused from a lack of SHM_MEM, its in the private memory pool. You need perhaps increase the PKG_MEM_POOL size in the config.h file. Also possible that you've a memory leak, see here for further troubleshouting hints: http://www.kamailio.org/dokuwiki/doku.php/troubleshooting:memory
We checked openser script for memory leacks using valgrind, but did not find any memory leack.
During doing relay of CANCEL,in relay route t_on_failure was added. So we removed t_on_failure in relay route for CANCEL. So WARNING:tm:run_failure_handlers: no UAC or CANCEL support (1, 0) was fixed.
Still we are facing,
- /usr/local/sbin/openser[29869]: ERROR:tm:t_forward_nonack: discarding
fwd for a cancelled transaction /usr/local/sbin/openser[29869]: ERROR:tm:w_t_relay: t_forward_nonack failed WARNING: script writer didn't release transaction
/usr/local/sbin/openser[29870]: ERROR:nathelper:nh_timer: out of memory
ERROR:usrloc:get_all_db_ucontacts: raw_query failed
rc_avpair_gen: received unknown attribute 80 of length 18:
0xA40AA6A78F04D67B03C75F8CA233F008
Will increasing private memory will solve all issues?
Not sure. :-) This are a lot of different errors. Before debugging this further, i'd suggest that you update to a newer version. Probably at least some of this problems are already fixed.
Cheers,
Henning