MVSFORUMS.com

[Shankarganesh_gopal]

Hi,

I have a question for you all. We have an MQ which will be sending some message into our MQ and our CICS transaction gets triggered.

But sometimes we loose the MQ connection and couldn't able to proceed further. Then we are initiating the CICS transaction again to get the connection back.

My question is - what would be the error message when there is a loss in MQ connection. Here our application has so many error message and I couldn't able to find the exact one. Is there a standard error message for connection issue?, if so what?

Thanks,
Shankar.

[bob_buxton]

CICS automatically reconnects to MQ following a connection failure and starts the trigger monitor so your transactions should be automatically restarted.

Applications using MQ would receive a reason code when connection is broken, one of:
MQRC_CONNECTION_BROKEN
MQRC_CONNECTION_QUIESCING
MQRC_CONNECTION_STOPPING
MQRC_Q_MGR_STOPPING
MQRC_Q_MGR_QUIESCING
_________________
Bob Buxton
Ex Websphere MQ for zOS development

[Shankarganesh_gopal]

Very thanks for the information.

But truly speeking, CICS is not reconnected to MQ automatically. I got the concept of CKTI and it is been implemented in our system. As per your statement, it seems CICS should reconnect. Could you please explain me how this is achived, so that I can have a check.

Also, I have another question. I have a transaction say AA, which will be triggered automatically every 15 minutes. Recently we got a abend in CICS with AEIP during CICS shutdown. We are assuming that the transaction started during the Shutdown process and hence got abended. Is it really possible?. Is there any other way by which we can predict the CICS shutdown and stop the transaction AA from Getting triggered.

Many Thanks,
Shankar

[bob_buxton]

If CICS is not not reconnecting to MQ when MQ restarts you should raise a problem with the IBM support centre. At the least they will want to see the joblogs of the CICS and MQ address spaces. There should be a long running transaction, CKAM, in your CICS region, it is CKAM's job to monitor the connection and to restart it when needed.

AEIP results from an unhandled INVREQ condition, without the EIBFN, EIBRESP, EIBRESP2 values it is not possible to say exactly what caused it but a transaction running during CICS shutdown may well experience such abends.
It is good programming practice to specify MQxxx-FAIL-IF-QUIESCING option on MQ API calls and to terminate your transcation if you receive one of the QUIESCING/STOPED/BROKEN reason codes.
_________________
Bob Buxton
Ex Websphere MQ for zOS development

[Shankarganesh_gopal]

Hi Bob,

Thanks for your valuable information.

I want to explain the whole scenario and my questions.

CICS CKTI should automatically trigger CICS transaction AA10 when there is a message in MQ. Since we are loosing the connection between MQ and CICS, we are not able to process the data. Hence we defined that a transaction AA20 will trigger another transaction AA10 immediately to read the data from MQ and will inform the CICS to retrigger the same AA20 after 15 minutes. (In case AA10 is not triggered automatically).

Now we want to remove the AA20 and have some logic in AA10 itself to be retriggered automatically even when the CICS-MQ connection lost. What we have done now is, we changed the program such that when AA10 fails due to connection lost, it will be triggered after 2 minutes by introducing START TRAN with Interval. Is this the proper way of handling this situation or is there any build in option available in CICS?

Thanks in advance,
Shankar.

[Shankarganesh_gopal]

One more question,

Assume there are 20 records in MQ which is not processed and the MQ conection is lost. What will happen to those records. Will it be there or we will loose those records?

Thanks in advance,
Shankar