[OpenSIPS-Users] One dispatcher mystery solved; Doesn't seem to be "remembering" a gateway is failed however

[Bogdan-Andrei Iancu]

Hi Jock,

Jock McKechnie wrote:
> Hey all;
>
> So I worked out what was going on with my previous test, where the 
> failover route wasn't being called - as Bogdan pointed out, it was a 
> TCP failure, because I was /sbin/route rejecting. If I actually 
> simulated a failure on the destination server OpenSIPS' fr_timer 
> caught it and it was pushed into the failure_block correctly. I was, 
> apparently, being too clever for my own good.
>
> My next problem is along the same vein - I continue to get dispatcher 
> failing over reliably (which it is now doing, hurrah!): I'm marking 
> the gateway as bad on a timeout, however it doesn't seem to want to 
> remember that the gateway is bad. I've got probing turned on, and I'm 
> marking it to be probed, but even an immediately subsequent call 
> (within seconds of the first call marking the gateway as bad) is 
> routed back to that first gateway. Now the failover works, it's 
> cleanly moving onto the next host, which is nice - but I'd still like 
> to utilise the probing scheme, if I can.
>
> I'm also getting a rather curious error in the logs:
> Found failover, working on set: 1101
> WARNING:dispatcher:ds_select_dst: algo 8 not implemented - using first 
> entry...
> Attempting to relay call to sip:+12125551212 at 192.168.0.99:5060 
> <http://sip:+12125551212@192.168.0.99:5060>
> FR: 204c758872fc3aa520599fd52a404783 at 192.168.0.2 
> <mailto:204c758872fc3aa520599fd52a404783 at 192.168.0.2> -- TIMEOUT for 
> Gateway 192.168.0.99 (marking as bad)
> ERROR:core:search_first_avp: 0 ID or NULL NAME AVP!
>
> I'm wondering if the above error is some kind of AVP storage thing I 
> haven't set up that is causing it not to properly "mark" the gateway, 
> or more likely, not remember the mark. But I can't spot anything in 
> the dispatcher documentation that says as such.
your guess is right - dispatcher module uses AVPs to keep the state 
(between sequential tries on the same dialog). 

I guess the ds_mark_dst() fails - can you check this?   (this function 
search for the dst and grp avps to identify the last tried destination.).

Regards,
Bogdan

>
> Relevant (hopefully) sections (cut-up config, this is not whole):
> Parameters initialisation:
> # Timers and failover flags
> modparam("tm", "wt_timer", 30)
> modparam("tm", "fr_timer", 3)
> modparam("tm", "fr_inv_timer", 120)
> modparam("tm", "disable_6xx_block", 1)
>
> # Set up the dispatcher
> modparam("dispatcher", "db_url", 
> "mysql://openser:password@192.168.0.20/company 
> <http://openser:password@192.168.0.20/company>")
> modparam("dispatcher", "table_name", "dispatcher")
> modparam("dispatcher", "flags", 2 )
> modparam("dispatcher", "dst_avp", "$avp(i:271)")
> modparam("dispatcher", "grp_avp", "$avp(i:272)")
> modparam("dispatcher", "cnt_avp", "$avp(i:273)")
> modparam("dispatcher", "ds_ping_method", "OPTIONS")
> modparam("dispatcher", "ds_ping_interval", 5)
> modparam("dispatcher", "ds_ping_from", "sip:+14109999351 at 192.168.0.2 
> <mailto:sip%3A%2B14109999351 at 192.168.0.2>")
> modparam("dispatcher", "ds_probing_threshhold", 3)
>
> Abbreviated main route block with route/failure_route blocks
> route{
>         xlog( "L_INFO", "[$Tf] Found failover, working on set: 1101\n");
>
>         if (!ds_select_domain("1101","8")) {
>                 t_reply("503","Unable to locate failover set requested");
>                 return;
>                 };
>
>         route(1);
>         }
>
> route[1] {
>         # Enable the failure block only on dispatcher lists which we 
> can fail over
>         t_on_failure("2");
>
>         xlog( "L_NOTICE", "[$Tf] Attempting to relay call to $ru\n");
>
>         if( !t_relay() ){
>                 return;
>                 }
>         return;
>         }
>
> failure_route[2] {
>         if (t_was_cancelled()) {
>                 exit;
>                 }
>
>         if( t_check_status("408") ){
>                 xlog( "L_NOTICE", "[$Tf] FR: $ci -- TIMEOUT for 
> Gateway $rd (marking as bad)\n" );
>                 ds_mark_dst("p");
>                 }
>
>         if( ds_next_domain() ){
>                 t_on_failure("2");
>
>                 if (!t_relay()) {
>                         return;
>                         }
>                 return;
>         } else {
>                 t_reply("503", "Service unavailable -- no more 
> gateways" );
>                 return;
>                 }
>         }
>
>
> As always, folks, I appreciate all your help. Enjoy your Thursday;
>
>  - Jock
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>   


-- 
Bogdan-Andrei Iancu
www.voice-system.ro