A couple of questions/comments regarding the 477 Send Failed condition:
1) When a 477 is sent, is there a way to detect it within a script so
that some custom handling can be applied (primarily logging)?
2) When the condition happens, and number of error logs are emitted. For
example:
Aug 27 10:10:33 devhost1 opensips[15884]:
ERROR:core:tcp_blocking_connect: poll error: flags 1c
Aug 27 10:10:33 devhost1 opensips[15884]:
ERROR:core:tcp_blocking_connect: failed to retrieve SO_ERROR (113) No
route to host
Aug 27 10:10:33 devhost1 opensips[15884]: ERROR:core:tcpconn_connect:
tcp_blocking_connect failed
Aug 27 10:10:33 devhost1 opensips[15884]: ERROR:core:tcp_send: connect
failed
Aug 27 10:10:33 devhost1 opensips[15884]: ERROR:tm:msg_send: tcp_send failed
Aug 27 10:10:33 devhost1 opensips[15884]: ERROR:tm:t_forward_nonack:
sending request failed
While one would obviously rather have a UA that is reachable, sometimes
things happens and it is not reachable. For example, a UA that dies
unexpectedly, or a mobile client that goes out of coverage area. Even
though this isn't a good thing to have happen, I question the
classification of it as an "error". I would think, especially given the
emission of a 477, that it is really an "application specific"
determination as to the severity of this case. Is it possible to get
these error logs reclassified as something less?
3) As a minor observation, it seems that in this case, the call to
t_relay() does in fact return a success indication. It's only after that
the error happens.
Thanks,
Gavin
_______________________________________________
Users mailing list
[email protected]
http://lists.opensips.org/cgi-bin/mailman/listinfo/users
