[cisco-voip] MGCP - Fallback to SRST very often even though connectivity to CUCM is fine

Tue Nov 3 09:29:24 EST 2009

timely question.

MGCP gateway can be viewed as:

MGCP Gateway
    mgcp/udp based registration and keepalives
    analog endpoints
       mgcp/udp based registration and transactions
    digital endpoints
       backhaul/tcp based

on CM if you see the alarm:
MGCPGatewayLostComm then the top level mgcp process stopped 
communicating with CM.  Usually the GW sends keepalives to CM similar to:
12/27/2005 10:16:40.173 CCM|MGCPHandler received msg from: 10.10.33.250
NTFY 333382 *@HQ-VG224-3rdFlr MGCP 0.1
X: 0
O:
|<CLID::MFCU-CM-1-Cluster><NID::10.10.200.11><CT::2,100,66,1.23017474><IP::10.10.33.250><DEV::> 

If CM does not receive keepalive from gateway CM will attempt to query 
the gw with this message:
12/27/2005 10:17:07.002 CCM|MGCPHandler send msg SUCCESSFULLY to: 
10.10.31.250
AUEP 13561613 AALN/S2/0 at HQ-VG224-1stFlr MGCP 0.1
F: X
|<CLID::MFCU-CM-1-Cluster><NID::10.10.200.11><CT::2,100,66,1.23017448><IP::10.10.31.250><DEV::> 

This AUEP is not 'normal'.
F = RequestedInfo
X = RequestIdentifier
Normal AUEP requests much more information. This is a special "hello, 
are you there" type exchange.

the gateway should respond:
12/27/2005 10:17:07.002 CCM|MGCPHandler received msg from: 10.10.31.250
200 13561613
X: 2
|<CLID::MFCU-CM-1-Cluster><NID::10.10.200.11><CT::2,100,66,1.23017648><IP::10.10.31.250><DEV::> 

This is getting very close to unregistration.  Another way to look at 
this is to look for indicates of lost messages to the gateway.  Each 
MGCP transaction is retransmitted up to 3 times if not ack'd.  You can 
see retries in the CM SDI traces:
01/13/2005 10:34:33.603 CCM|MGCPHandler TransId: 1097943 Timeout. Retry#1

If you see frequent retries then you are intermittently dropping or 
excessively delaying the UDP packets carrying the MGCP payload.

There is also an issue where endpoints may stop responding to CM.  CM 
will retry the transaction 3 times and then unregister the gateway.  
This looks similar to the retries tracked above.  The main difference is 
that you will see valid exchanges with other endpoints on the gateway or 
you will see successful keepalives with the top level gateway MGCP 
process.  This was historically caused by CSCsf26617 and similar.  The 
signature of this failure is repeated retransmits of the DLCX, RQNT, or 
CRCX messages from CM to the gateway while other endpoints are 
responding.  If this is happening then the gateway is having an internal 
error such as resource allocation or dsp hang.

HTH.

/Wes

On Tuesday, November 03, 2009 3:39:24 AM, Wilson Hew 
<wilsonhew at gmail.com> wrote:
> Bob/Ryan, appreciate your feedback. Thanks.
>
> Guess I need to look at the connection between my MGCP gateway and 
> CUCM. Any idea what else I may need to check? I am looking at the SDI 
> traces, but have no idea what to look at.
>
> Thanks,
> Wil
>
> On Tue, Nov 3, 2009 at 2:35 AM, Bob Fronk <bob at btrfronk.com 
> <mailto:bob at btrfronk.com>> wrote:
>
>     I had this happening and found out it was an MPLS circuit going
>     down.  Due to location of this particular site, our 12mbps MPLS
>     circuit is supplied by multiple T1s bonded with MLPPP.
>
>      
>
>     One of the T1s was going up/down several times a day (telco
>     problem) and each time, the MLPPP would reset for a couple
>     seconds.   The MGCP gateway responded by going into SRST and the
>     PRI would go down for a moment.
>
>      
>
>     Just something to check
>
>      
>
>     *From:* cisco-voip-bounces at puck.nether.net
>     <mailto:cisco-voip-bounces at puck.nether.net>
>     [mailto:cisco-voip-bounces at puck.nether.net
>     <mailto:cisco-voip-bounces at puck.nether.net>] *On Behalf Of *Wilson Hew
>     *Sent:* Monday, November 02, 2009 11:47 AM
>     *To:* cisco-voip at puck.nether.net <mailto:cisco-voip at puck.nether.net>
>     *Subject:* [cisco-voip] MGCP - Fallback to SRST very often even
>     though connectivity to CUCM is fine
>
>      
>
>     Hello there,
>
>     Greetings. I am having problem with my MGCP gateway, and I need
>     little help and advice. My MGCP gateway is running as SRST, and it
>     will fallback to SRST very often (twice a day). And it will go
>     back to normal operation from fallback just after that. The
>     connectivity from my MGCP gateway (remote site) to CUCM is fine.
>
>     I noticed my E1 is going down everytime when it falls back to SRST
>     - is it considered normal?
>
>     My gateway is running 12.4(24)T1 and CUCM version 7.0.2.
>
>     In 'sh ccm-manager', I have the below:
>     --------------------------------------------------------------
>     TFTP retry count to shut Ports: 2
>
>     Statistics:
>         Packets recvd:   857
>         Recv failures:   1
>         Packets xmitted: 852
>         Xmit failures:   0
>     --------------------------------------------------------------
>     In 'sh mgcp stats':
>
>      UDP pkts rx 557379, tx 558783
>      Unrecognized rx pkts 0, MGCP message parsing errors 0
>      Duplicate MGCP ack tx 9, Invalid versions count 0
>      CreateConn rx 36256, successful 36249, failed 7
>      DeleteConn rx 36274, successful 36101, failed 173
>      ModifyConn rx 66178, successful 66126, failed 52
>      DeleteConn tx 154, successful 154, failed 0
>      NotifyRequest rx 54652, successful 54516, failed 136
>      AuditConnection rx 3, successful 3, failed 0
>      AuditEndpoint rx 14887, successful 8080, failed 6807
>      RestartInProgress tx 6248, successful 6248, failed 0
>      Notify tx 342779, successful 342779, failed 0
>      ACK tx 201075, NACK tx 7191
>      ACK rx 349100, NACK rx 0
>      Collisions: Passive 0, Active 0
>     --------------------------------------------------------------
>
>     Can I tell what is wrong with the above? Apart from that, I see
>     numbers of slips in controllers e1 increasing, and I have
>     network-clock-participate configured.
>
>     Would appreciate if you could give me your feedback about this.
>     Any feedback is very much appreciated.
>
>     Thanks,
>     Wil
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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/20091103/ee377f2b/attachment.html>