hi all,

I am using openser 0.9.5. I recently configured openser supporting gateway chaining with two failure routes. I tested it and it was working fine, and when I put it into actual production environment which processes close to 200 calls per second in peak hrs, openser crashed after an hours time. I think there was a memory leak.

The error message in syslog at that time showed:

ERROR: t_should_relay_response: status rewrite by UAS: stored: 487, received: 100

my snippet of cfg file is below:

 

modparam("tm", "fr_inv_timer", 27)
modparam("tm", "fr_timer", 3)
modparam("tm", "fr_inv_timer_avp", "3")

 

if(method=="INVITE") {
  setflag(1);
               
  if (uri=~"^sip:00[0-9]*@") { 
                        strip(2);
                        rewritehost("216.XXX.XXX.XX");
                        t_on_failure("1");
                        t_on_reply("2");

                        t_relay();
                        break; 
       };

 

failure_route[1] {
        rewritehost("217.XXX.XXX.XX");
        append_branch();
         t_on_failure("2");
         t_relay();
}

 

failure_route[2] {
         rewritehost("218.XXX.XXX.XX");
         append_branch();
         t_relay();
}

onreply_route[2] {
 if (status=~"18[0-9]") {
  t_on_failure("0");
 };
}

I had ser fr_timer to 3 secs so that if there is no response within 3 secs the other route is tried.

Please let me know what logical error was there in my cfg which made openser crash. Is it something related with the timer?

 

thanks in advance.

jayesh