CS1000 keeps losing connection to CCMS for about 25 sec.

[Mooselite04]

Hi,

We have 2 PBX in 2 diferrent location and both PBX are doing this. They both have CCMS on a Virtual server Microsoft 2003 and every 1 - 3 days they will give out this alarm.

nortel-ippbx : sev=1 ErrorCode=CSA002 almtype=8 cause=202 desc=16 7:55:19 12/12/2012 6:IP Address: 10.187.35.74:Original IP: 10.131.97.21:Circuit Name: :Remote Dial Up Number:

Then if you print the history in the PBX you get this. I hide the IP by putting an X for the IP but it actually display the IP of my CCMS.

% CSA002 16 5:34:24 4/01/2013 6
%
% ELAN007 ELAN 16 host IP=x.x.x.x disabled, read from socket
% fail due to far end disconnect or Ethernet problems
%
% CSA003 16 5:34:26 4/01/2013
%
% CSA105 16, 5:34:26 4/01/2013
%
% ELAN014 ELAN 16 host IP=x.x.x.x is enabled
%
% SRPT091 HB: Remote side AML health change from 4 to 2.
%
% SRPT092 AML: remote side AML connection 16 to x.x.x.x health change: 2.
%
% SRPT091 HB: Remote side AML health change from 2 to 4.
%
% CSA104 16, 5:34:56 4/01/2013

 

[tgolds]

You should have an ELAN and TLAN connection to a Switch/router/hub. I had a bad port on a cisco switch one time for my ELAN connection. ALso have th IT folks check that the port is configured correctly.

 

[Mooselite04]

Port configuration do you mean 100MBPS and fullduplex?

 

[KCFLHRC]

The ports should be set to auto negotiate

 

[DFKSydney]

The error states that the cause is far end disconnection.

Given that, I would be looking outside the PABX. Heres a few suggestions:

1: Check the CCMS for anti virus activity. This could be a full system scan, or a periodic update. Symantec Endpoint protection is bad for this.

2: Have your elan checked. Ensure that the vlan configs do not allow for any other vlans traffic.

3: Check the elan connections (individual) to ensure that they are both on the same vlan.

4: Depending on the vintage of the system... older works on 10Mb half duplex, newer is 100Mb full. Auto negotiate is unreliable on the 100Mb. If you verify that you are truely 100M then have the switch ports set that way.

5: Have you had a look in the PABX to see if the failure time is around the time that the midnight routines run?

 

[sl1input]

Make sure Windows Time Service on the CCMS is stopped and set to disabled.
An AML CCMS will get its time from the CS1K.

 

[MikeInHouston]

Another possibility is excessive Spanning Tree reconvergence time. If spanning tree is triggered the port blocking may exceed your link keep alive timer.

 

[Mooselite04]

Ok we did some test and we are getting closer to our issue. We are able to prove it's the CCMS adjusting it's time. We set up another box to ping the CCMS and CCMS to ping the ohter box. The other box miss like 13 ping in about 2 sec and CCMS never miss a ping. Then if you check where CCMS is not responding it's adjusting it's time. Why some times it will adjust it's time one time and everything is still fine and some times it has to do it 3 times before it gets back to normal and at the same times messes up the PBX for a respond time and the PBX will generate alarms for this issue. Also our network card is setup to 1gbps for network speed which is coming off a virtual machine. Any help. Thanks