|Summary:||ASTERISK-17722: SIP SRV lookups for registration discard the port when dnsmgr disabled (the default)|
|Reporter:||Terry Wilson (twilson)||Labels:|
|Date Opened:||2011-04-19 14:50:23||Date Closed:|
|Description:||When doing the following with srvlookups=yes in sip.conf and dnsmgr=no in dnsmgr.conf (the default):|
register => tls : // foo:firstname.lastname@example.org
the registration will fail if example.com specifies a port other than the default port of 5061. Asterisk looks up the correct port, but overrides it. This can be worked around by enabling dnsmgr, but it should work in either case. Is there ever a reason to not have dnsmgr enabled?
|Comments:||By: Luke H (luckman212) 2011-09-22 13:55:05.566-0500|
1) I would also like to know whether there is any downside to dnsmgr=yes -- very little documentation exists about this!
2) I am running Ast 1.8.7-rc2 and I believe this is still an issue, even with non-TLS connections. e.g. I have a SIP trunk with Callcentric and it appears that it wants to be registered at UDP port 5080 (not 5060). e.g. the output of 'host -t SRV _sip._udp.callcentric.com' yields this list of 9 proxies:
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha5.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha6.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha7.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha8.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha9.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha1.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha2.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha3.callcentric.com.
_sip._udp.callcentric.com has SRV record 20 0 5080 alpha4.callcentric.com.
however when I check via CLI 'sip show registry' I see this (note port 5060 not 5080):
callcentric.com:5060 N 1777xxxxxxx 45 Registered Thu, 22 Sep 2011 14:52:24
In addition, my console is littered with timeouts every few minutes for that trunk (not the 4 other SIP trunks, just Callcentric) e.g.
[2011-09-22 14:11:28] NOTICE: chan_sip.c:12681 sip_reg_timeout: -- Registration for 'email@example.com' timed out, trying again (Attempt #3)
I think those errors are related.
By: Luke H (luckman212) 2011-09-24 09:42:46.709-0500
I have more details to add to this issue:
1) I've updated to 1.8.7 final and no change - this seems to still be a bug.
2) dnsmgr=yes|no makes no difference for me, Asterisk always registers at port 5060 disregarding what gets returned by the DNS SRV lookup (in this case port 5080)
3) manually forcing a port in the register= string for the peer does work but that essentially negates the point of using DNS SRV in the first place.
By: Luke H (luckman212) 2011-09-25 23:10:34.854-0500
Also I question the classification of this bug as "Minor" - I would have thought it "Major" if not "Critical".
By: Paul Flores (spm.pflores) 2012-07-25 22:57:02.415-0500
Any update on this issue? I'm also experiencing this kind of problem. Thanks