Hello everyone,
We are looking to leverage Classic Data Capture for VSAM (IIDR for VSAM) to
replicate our VSAM file changes to a distributed tier for caching servicing
data.

My first question is if anyone has gone down this path already, and if so
were there any surprises along the way?

My understanding so far, from the materials I have read, is that we will
turn on CF Logstream logging for the specified files. Then any real time
updates made to the files in our CICS environments will trigger capture
agents running in the CDC Address Space.
We also need to install/configure CICS VR to capture the same Logstream
events for any Batch JOB updates to the files.

Is it possible for the VSAM files to become out of sync with the replicated
data?
I would expect that from a CICS perspective, once the update is committed
to the VSAM file, the log images are marked as committed, and these would
then pass through replication.
I can't see how CICS work could push things out of sync, but I want to ask.

When looking at things from the Batch side of the equation, well a Batch
JOB can ABEND.
Then, normally, we would restore the VSAM file environment (correct
whatever the issue was) and re-run the Batch JOB.
I read something about DWWBACK, the CICS VR Batch Backout Program.
Should I be introducing DWWBACK wherever/whenever we restore our VSAM files
and re-run a Batch JOB?

Last question for now, and I really do understand this is much like asking
how long is a piece of string.
When we turn on log replication for the VSAM files, should we be expecting
to see a small bump in CPU and wall clock times? I would expect that since
we do more stuff, it should take more time and resources. Any real world
experience here? Say five percent increase in MIPS and ten percent more
wall clock elapsed times? Any examples?

I have been chatting with our IBM folks, and I have been reading materials.
There is just so much information out there, it is overwhelming.

Thanks in advance for any information

.......Cameron

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to