[cisco-voip] RTMT: MTP resources Active

[Wes Sisk]

During periods of low call volume that number should decrease.   
Depending on your traffic modeling you should see significant dip in  
late night/early morning.  You can get more correlation by comparing  
to number of active calls.  Be sure to compare against number of  
active calls across all cm nodes as an MTP can be allocated from  
anywhere within the cluster.

if you truly find 0 active calls but MTP remaining high you very  
likely have a leak.  Then there are a couple of ways to go:
1. provide your complete CM version.  we can look to see if there are  
known MTP leaks in that version.
2. truly diagnose the problem.  collect all CM SDI and SDL traces from  
all nodes in the cluster.  the traces should cover from a low use  
time, through peak usage, and back to low use;  example midnight day1  
to midnight day2.  Submit a TAC case and provide these traces.  TAC  
has a tool to parse and look for leaks.  Note your traces must be  
complete, must be from all nodes, and must cover the whole time.   
Partial traces will not work.

3. workaround the problem.  periodically restart the ip voice media  
streaming application service.

/Wes

On Jan 29, 2009, at 10:00 AM, Hennigh, Elexis wrote:

I have a question…

I have been monitoring our MTP resources. We use software MTP with H. 
323. gateways. We originally had trouble with running out of MTP  
resources with the threshold set to the default 48 per server. We have  
since increased the number of available resources to 64 per server and  
having 2 servers means 128 available MTP resources. Using RTMT I have  
been monitoring the number MTP Resources active. It seems to  
constantly stay around 70 this week. The trend though seems to be  
increasing as time progresses. I can see the MTP resources decline but  
not significantly. It gradually increases. RTMT reports over 70 MTP  
resources active even through the night and through the past few snow  
days. I am not sure what is going on. I assumed the number of active  
MTP resources would eventually return to zero if no one was using the  
phones. Isn’t this strange? Doesn’t it seem that we have a problem  
with how the call manager releases MTP resources?

If anyone can explain what is happening here I would appreciate it. J
_______________________________________________
cisco-voip mailing list
cisco-voip at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-voip

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/cisco-voip/attachments/20090129/995ca7ec/attachment.html>