Summary:ASTERISK-04751: SIP: Maximum retries exceeded on call 00120138-3ef9019f-27a0903e-7062e8d4
Reporter:lters (lters)Labels:
Date Opened:2005-08-01 15:03:56Date Closed:2011-06-07 14:10:00
Versions:Frequency of
Environment:Attachments:( 0) sip_errors.txt
Description:Cisco phones running 7.3 firmware, when calls load gets semi busy we see tons of these errors.

WARNING[25526]: Maximum retries exceeded on call 00120138-3ef9019f-27a0903e-7062e8d4


See attatched file.

Comments:By: lters (lters) 2005-08-01 15:25:14

Users are not able to pick calls.
Users try to make calls and they will not complete.

It seems to be directly sip registration related.

By: Michael Jerris (mikej) 2005-08-01 15:29:36

we need the debug with verbose as well for the call.  w/ verbose and debug set to 4.

By: Tilghman Lesher (tilghman) 2005-08-01 16:33:43

It probably means that you have insufficient bandwidth on your switch to handle the call volume.  How many calls, what codec are you using, what's the rated bandwidth on your switch?

By: Olle Johansson (oej) 2005-08-02 02:45:27

Please file in the SIP category if it's about SIP, thank you.

By: Olle Johansson (oej) 2005-08-02 05:13:04

I can't find "Maximum retries exceeded" in the debug file you uploaded. Please try again.

By: Olle Johansson (oej) 2005-08-02 05:45:47

Major bug: "MAJOR: A bug which completely prevents Asterisk from operating in a method that it normally is expected to operate -- and particularly if it cannot be reasonably worked around -- is MAJOR. Significant protocol violations that are not simply policy decisions are MAJOR."

I can't see that this is a MAJOR bug, so I change it to Minor.

By: lters (lters) 2005-08-02 06:49:15

Thanks for all the comments.

I have more debug info that I can send. Would there be an email I can send it to rather than a public post?

Bandwith? Well, monitoring shows us to use a max of 2 to 3 meg of traffic.
2 interfaces with the majority of the voice on a 1000FD intel card going to a cisco 10/100/1000 switch. There are 3 10/100 switches directly connected to the 10/100/1000 switch. All the sip devices are then connected to the 10/100 switches.  

The real issue though the fact that the problem seems directly connected to call setup. Calls in progress never have been a problem. But, new calls coming in suddenly seem to not be answerable or there is a 5 second pause before the audio starts. The other weird thing is: The max retries appears to also happen to the * boxes own ip address.

I tend to rule out bandwidth issues, because call quality is not an issue.
And, we have the tT in our dial statements, so, calls likely are never bridged directly together.

It does seem MAJOR for us, but it is fine with me to change it :)

By: twisted (twisted) 2005-08-03 13:45:39

It will show the * ip, as the sip URI it's trying to reach is registered to itself.  

SIP URI's are something like:

username@   where is the server IP.  Basically, it's saying it is having trouble reaching the device registered with that URI.  

This does sound like a network problem, however.  I have only encountered this in the following scenarios:

1) the asterisk box is told that it's externip is different than what it actually is
2) the network load bottlenecks at the asterisk side
3) there are other issues involving packet loss, or routing.

also, the sip_errors.txt you uploaded indicates no problems - that is a one sided call flow from asterisk to the phone.  Everything there looks normal.

By: Mark Spencer (markster) 2005-08-09 16:14:28

Can you test with CVS head and provide a new debug?

By: lters (lters) 2005-08-12 15:19:50

At the moment I don't have a way to test the Max retries the busiest * box, which is where the problem was happening. I can't run cvs-head due the locking problem with agents/queues. (This may be fixed, I don't know.)
On another one of our * boxes, cvs-head seems to run fine.
If you want to close this, you can.
Thanks everyone for your good comments.