CINIT failure

[fusahinkaya]

Our branch offices got a problem to work with a CICS.

The terminal sent initself request and got +rsp ( there should be bind from CICS) andthen we got NS procure error for CINIT ( CINIT error in reaching the PLU) Is anybody face this problem before?

Best Rgds,

Funda

[dixho]

Hi,

I am a TAC engineer. I worked on a similar issue before. I am searching for information and want to find out the resolution. Please post the network topology and IOS version.

[dixho]

After further researching, I find out that CSCdy89290 has a similar sympton. The problem only occurs in a very specific topology. You need to run SNASw and use DLUR for the downstream PU 2.0. Also, you need have one NNS. The problem occurs after a network outage. The CP-CP session between the router and NNS times out, but the HPR/IP link does not. Thus, the SNASw router runs without a NNS. When you run a PU2.0 via DLUR, VTAM cannot send a bind because there is no CP-CP session between VTAM and SNASw router.

Not sure if you hit the problem or not. Please provide topology and IOS version.

[fusahinkaya]

Hi,

thanks for reply. First of all our ios ver 12.1.15 on snasw router. We doen't implemant HPR on them yet. During the problem CP_CP between NNS and DLUR router was up at least it seemed as up may be it was not working properly.

Topology like that:

VTAM( Backup NNS,CICS)

/

/

VTAM(NNS)-CIP-SNASW& DLSW router- Backbone router--Frame Relay- Branch router- TR- Branch server and clients.

Rgds

Funda

[fusahinkaya]

Hi,

thanks for reply. First of all our ios ver 12.1.15 on snasw router. We doen't implemant HPR on them yet. During the problem CP_CP between NNS and DLUR router was up at least it seemed as up may be it was not working properly.

Topology like that:

VTAM( Backup NNS,CICS)

/

/

VTAM(NNS)-CIP-SNASW& DLSW router- Backbone router--Frame Relay- Branch router- TR- Branch server and clients.

Rgds

Funda

[dixho]

I think that you have to find out why VTAM/PLU does not send out the bind. As the brand server and clients are PU2.0, they are not capable of sending bind. The bug I mentioned in my previous reply occurs when the CP-CP session is lost. Thus, DLUS/VTAM does not know where the PU2.0s are.

[bkelley]

Funda, this type of problem is relatively common. SInce you didn't mention your LU type that is sending the "INITIATE-SELF", I assume it's either a LU type 2 or LU6.2. Regardless of the type, one thing is for certain - it's a dependent LU.

This means that the LU is dependednt upon VTAM to assist in session setup.

Once VTAM receives the INIT-SELF from the secondary LU (SLU), it will compile data including BIND Image (from the MODE TABLE), LU names, and other control vectors. This data is then placed into a Control-Initiate (CINIT) RU and passed the the PLU (CICS in your case). If all goes well, then the PLU with respond to the CINIT with a +rsp and then send out the BIND request.

However, there are many things that can go wrong here to inlude:

* CICS not having the model type to match on

* Authorization

* VTAM route issue (ER/VR)

* CICS defintion problems

* Incorrect logmode

I would suggest looking at your systems logs (VTAM,CICS messages) and running a trace (VTAM buffer) to look at the CINIT. You will usually see the error in a log message with an associated sense code. The sense code will give you more information to work with.

Good Luck!

B.K.