[OpenSIPS-Users] Opensips stalls on sending messages if resolving domain by dns

[Bogdan-Andrei Iancu]

Hi Bobby,

unfortunately the DNS query (done by opensips) is synchronous , this 
means it will block until it gets a response (which can take some time 
if the dns server is down).

you can minimize the impact by tuning the dns related params - see:
    http://www.opensips.org/index.php?n=Resources.DocsCoreFcn#toc39

Regards,
Bogdan

Bobby Smith wrote:
> The setup -- we're using an internal dns server to route stateless 
> transactions.  In the routing script, depending on who the request 
> user is, we replace the request-uri domain with a domain address and 
> forward the message statelessly.
>
> We've experienced an issue where on the stateless forward function, if 
> for some reason the dns server is down, opensips just kind of sits 
> there.  We see on an ngrep that the messages come to opensips, but 
> opensips doesn't really do anything and never gets to execute the 
> forward("mydomain") function.
>
> The way we know this is that, sometimes, when the DNS lookup is 
> delayed or finally comes back up, opensips seems to forward the 
> message finally.  If it's an invite, the message causes the sip device 
> to ring again after a call is established, an undesired behavior.
>
> Is there any simple way to request a maximum amount of time that it 
> will try to do something with the message, and if that time is 
> exceeded (because it can't resolve the domain), to simply drop the 
> message and exit processing for that particular message?
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users mailing list
> Users at lists.opensips.org
> http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>