Intermittent Call issues With SIP Trunk

[Shawnlo]

Hi,

We are all of a sudden having issues with SIP calls either dropping on incoming calls or when making a call out sometimes there is a long pause with no ring out, occasionally users say they just get silence and the call does not go through.  

We have two CUBE routers round robin on the sip trunk connected to 3 call manager servers.  I see calls active in both routers all the time.  In wire-sharking the traffic we see all calls are using EF. We have no other traffic passing as EF. 

I am lost on where to look next to figure out what is going on and why all of a sudden we are dropping calls.  Our SIP provider had very little information and said it looked at one point like they could not reach our routers.  As far as I can tell our routers have been available 100%.  I am hoping someone can help me out with a troubleshooting process.

[Georgios Fotiadis]

At first, verify network connectivity between all parties (CUCM, CUBE, ITSP). Have you made any changes in your network topology?

Can you find a pattern in "faulty" calls; i.e. calls from specif branches,users,phones?

Then, collect logs from CUCM (CallManager SDL) and CUBE (debug ccsip messages) for a faulty call (incoming or outgoing) and attach them (along with calling/called number and timestamps).

[Dennis Mink]

I would agree with George, get your user to note times of troubled calls, and running detailed traces all the time so you can analyze the signalling of a failed call, on your cube, write the SIP debug info to flash.

also traffic being marked as ef does not really mean a lot if it is not being honored of if there is contention with other traffic that is incorrectly marked as ef. so talke to your provider to get insight on what happens with your traffic classes

[Shawnlo]

Thank you,

I have a trace however I have to now figure out the best way to sanitize the trace from call manager before I attach it.  As far as EF, our fiber provider did note we were dropping some packets on EF, so I have increased the bandwidth to 17 on our uplinks with them.  I have also applied a setting at our head end routers to move CS3 from call managers to Af31, as our provider only gives us 3 classes we can use.  DSCP 0, 26, and 46.  I am wondering if some of these issues started when we started dropping traffic on 0.  CS3 was getting dumped to 0 across the WAN, I am assuming this is call control traffic.

Writing the debug to flash I have not done as of yet, however I will try and get that done today,