[sr-dev] Crash bug

[Alex Balashov]

I have two core dumps from the last few days with the same crash, and I 
took a look at the call flows for both Call-IDs.

Both seem to play out the same way as the crash I sent you:

Mar 19 13:22:57 Proxy1 /usr/local/sbin/kamailio[21580]: INFO: 
[R-OUTBOUND-VENDOR-ROUTE-TRY:e5e48a99-48dd-1233-96b7-782bcb13da6a] -> 
Final request URI: sip:yyy at xxx:5060

Mar 19 13:22:57 Proxy1 /usr/local/sbin/kamailio[21580]: INFO: 
[R-INVITE-OUTBOUND-HANDLER-POSTROUTE:e5e48a99-48dd-1233-96b7-782bcb13da6a] 
-> Request processing delay: 27 ms

Mar 19 13:22:57 Proxy1 /usr/local/sbin/kamailio[21580]: INFO: 
[R-OUTBOUND-VENDOR-ROUTE-TRY-BRANCH:e5e48a99-48dd-

1233-96b7-782bcb13da6a] Sending media through rtpproxy
Mar 19 13:22:57 Proxy1 /usr/local/sbin/kamailio[21572]: INFO: 
[R-MAIN:e5e48a99-48dd-1233-96b7-782bcb13da6a] CANCEL received from xxx 
(RURI=sip:yyy at xxx:5060)

Mar 19 13:22:58 Proxy1 /usr/local/sbin/kamailio[21569]: INFO: 
[R-MAIN:e5e48a99-48dd-1233-96b7-782bcb13da6a] CANCEL received from xxx 
(RURI=sip:yyy at xxx:5060)

Mar 19 13:23:00 Proxy1 /usr/local/sbin/kamailio[21566]: INFO: 
[R-MAIN:e5e48a99-48dd-1233-96b7-782bcb13da6a] CANCEL received from xxx 
(RURI=sip:yyy at xxx:5060)

They also have another common characteristic:

The normal call flow is to relay the INVITE on branch .0 to an LNP 
gateway, then to capture the LRN from a redirect and relay the INVITE to 
its actual destination via branch .1. In all cases, however, there was a 
Redis cache hit on the LRN that did not require branch .0.

-- Alex

-- 
Alex Balashov - Principal
Evariste Systems LLC
303 Perimeter Center North
Suite 300
Atlanta, GA 30346
United States

Tel: +1-678-954-0670
Web: http://www.evaristesys.com/