MATE_UNAVAIL in PGW

[denizpecel]

hi

We are using pgw in nailed mode and 5400s as gateways. When we define new channels and sometimes without no reason some of the cics going to state MATE_UNAVAILABLE and it is vert hard to correct the error. I try shut/no shut on controller, serial interface and rlm group, switch over in the pgw etc. Is there a a simple way to correct this?

Thanks in advance..

Deniz

[mohashim]

I'm also having same problem, after add Cic going to state MATE_UNAVAILABLE. how to resolve this problem.

[teodorgeorgiev]

This is a documented Cisco bug.

It affects those, who have applied the CSCO037 patch.

Apply patch CSCOgp038 and you should get rid of this issue.

[tominatornz]

I am seeing this problem also. But only in the following scenario...

I have IUA to some gateways and RLM to others.

If the back-haul network goes away briefly, then the cics on the PGW related to the ext nodes with RLM recover fine, but the ones with IUA do not. They remain in the MATE_UNAVAIL state, and a reset-cic has no effect. It is not until the nfas group dchannel has a shut/ no shut that the pgw sees the partner state correctly, or if this doesnt work, the Controller must be shut/no shut. If either of those do not cure the problem, as a last resort, a reload of the gateway does.

I am running 9.4(1) with system patch 40 and protocol patch 47

Our back haul is provided by a third party so if they have an outage (planned or unplanned), it affects us big time!!!

Does any one else experience similar problems? If so, how have you fixed them?

[tominatornz]

I just found this...

The following bug appears in the 12.4 Mainline release notes - section 5, Caveats

? CSCsc59584

Symptoms: Cisco IOS Release 12.4(3) and Release 12.4(5) GROUP SERVICE messages are not correct to the PGW.

Conditions: The following actions will result in all circuits on the PGW to remain in MATE_UNAVAIL state.

1. Set the IUA destination to out of service (OOS) and then set it to in- service.

2. Set the IUA association to out of service (OOS) and then set it to in- service.

3. Restart the PGW.

Workaround: Reload the IUA to clear the circuits.This is unacceptable to a live customer.

I am using 12.4(4)T and the same thing happens to me (see previous post). The 12.4T release notes make no mention of this issue.

It seems it may be in BOTH IOS trains...

[tominatornz]

This bug was first found in 12.4(5). 12.4(4)T was one of the affected versions. There is quite a list, you should check this to see if the IOS you are running is on the list.

It was first fixed in later releases of 12.4.(5), however these releases are deffered.

I loaded 12.4(6)T and the problem went away.

[mohashim]

In my case, one of my AS5400 have 8E1 divided into 2 and connected to 2 difference PRI switch with 2 RLM group. One site (4E1) always shown MATE_UNAVAILABLE but another 4E1 working fine. Any idea ?

[tominatornz]

Have the cic's associated with each signalling point been provisioned at the DPC end? and if so, have they ever been in service?

With newly provisioned circuits, the associated dpc (switch) must acknowledge the unblock or reset message before the pgw takes the affected cic's out of the MATE_UNAVAIL state.

If they remain in this state after you see either UBA or GRA messages from the associated switch, look at your gateway for a possible problem.

[anikas]

Deniz,

If you are just adding the configuration for the first time or adding new CICs it is always best to export your configuration first and make sure that the CICs are mapped to the correct NASPATH and the correct NFAS interface. The next thing you want to check is that your RLM connection is good. Make sure that the ports are open for RLM and connectivity is ok. Finally you will want to check the gateway.

Here look at the show isdn service and make sure that the CICs do not show a pending state. Also check out the cottrollers T1/E1 and make sure all is well there.

This should cover 95% of these type of problems.

Best of luck.