Bill,
DB2 will hold the thread open for a short time waiting to see if it can be
reused. It's expensive to create/destroy threads.
Do you have Omegamon, or CICS/Mon, or some other monitoring tool, mostly
for CiCS? If so, take a look at those transactions
and I think you'll see the DB2 time is very small (DB2 is unbelievably
efficient these days), but CICS will have some latency
like transaction creation/termination, if cross-memory services are being
used, there's a cost for that. You should probably
see the vast majority of that 30-50 seconds spent in CICS internally,
though CICS' actual CPU time will be very small also.
CICS is quite efficient, too.
The last CICS/DB2 shop I was in had extremely quick internal times for CICS
and DB2, but the entire transaction would
easily take several seconds due to transaction routing, cross-memory
services, etc.
30-50 seconds sounds pretty extreme, though. There may be some tuning
opportunities on the CICS side.
Ramsey
On Thu, Sep 23, 2021 at 1:16 PM Bill Giannelli
wrote:
> This is a question about at what point does CICS signal Db2 that the
> thread is to be closed.
> Db2 will close the thread and write the accounting record.
>
>
>
> Situation:
>
>
>
> DB2 SMF accounting records are showing subsecond Db2 time, and elapsed
> time of 30-50 seconds.
>
> The caller is CICS, on the same LPAR.
>
>
>
> Db2 is waiting for something before the thread is closed.
>
>
>
> The possible candidates for the latency are:
>
> VSAM
>
> I/O subsystem
>
> Network
>
>
>
> DB2 systems programmer is trying to diagnose with the assistance of CICS
> Systems programmer.
>
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN