[asterisk-users] Asterisk uses 3 seconds to send ACK after OK

[Pan B. Christensen]

Taking a look at the DEBUG statements that are associated with the
thread processing the SIP response:

[Mar 15 13:16:05] DEBUG[27947] netsock2.c: Splitting 'FQDNz:5060' into...
[Mar 15 13:16:05] DEBUG[27947] netsock2.c: ...host 'FQDNz' and port '5060'.
[Mar 15 13:16:08] DEBUG[27947] netsock2.c: Splitting 'yyy.yyy.yyy.yyy'
into...
[Mar 15 13:16:08] DEBUG[27947] netsock2.c: ...host 'yyy.yyy.yyy.yyy' and
port ''.
[Mar 15 13:16:08] VERBOSE[27947] chan_sip.c: [Mar 15 13:16:08]
Transmitting (no NAT) to yyy.yyy.yyy.yyy:5060:
ACK sip:b at FQDNz:5060 SIP/2.0

If I had to guess, the DNS resolution of 'FQDNz' probably took 3
seconds. You may want to consider a local DNS cache to help speed up
results.

Matt


Thanks for your reply, Matt.

I thought about that as well after sending the original mail, and indeed the
first server in resolv.conf was not replying. I fixed that and DNS lookup is
now very fast, but the problem still remains. Could Asterisk still have the
old resolv.conf cached or something? I tried a reload but that didn't help.
As this is a busy server, I'm unable to stop and start Asterisk now during
business hours.

With kind regards,
Pan


Update: I think the problem is solved now, but I don't understand the 
solution. Since last time I've tried playing around with the dnsmgr to no 
avail. Today I enabled qualify on the trunks in question and the problem 
seemed to go away. Does anyone have any idea why?

With kind regards,
Pan