DSLB / ZCRNLOG - delay/ time for results to appear

[John Barrett]

​We're currently in the process of installing and verifying the ZCN product feature pack in CA-SYSVIEW V16.

We've managed to complete the configuration, create logstreams etc and have results coming thru' into ZCNRLOG.

However; there is a lag of >1 hr between requests being issued and them appearing in ZCNRLOG.

Is there a polling parameter which we need to change / create ?

[Sean McShane]

Hi John,

Are you collecting SMF type 123 subtype 2 records? If so, these records are only written when one of the following conditions are met (as documented here):

1. The server processed the required number of requests to fill a record. Currently this is 20.
2. The server is shutdown.

Is it possible these conditions are causing your problem? If you are only collecting type 123 subtype 1 records, something else might be at play.

On the ZCNLIST command, you will find a "FLUSH" line command that will give the z/OS Connect server a hint to cut the records early that you can give a try to see if it lessens the time you are waiting for the records to be cut.

[John Barrett]

Thanks Sean. Sorted thank you. ​

[Todd J. Gagle]

John,

One possible reason for the delay.

When z/OS Connect EE changed the SMF 123 records to version 2, they started blocking their records. z/OS Connect will only write these records when the buffer becomes full. The buffer typically holds about 20 records.  On a system where the activity is very low, it could take considerable time to generate enough records to fill the buffer.

I raises this concern with IBM. I have suggested that they need to include some form of timer event that would "flush" the buffers at a reasonable interval. They were not very receptive to the idea. They did not see this to be a problem.

SYSVIEW does include a "FLUSH" line command on the ZCNLIST command to flush the buffers on demand.

Please let us know if this matches your situtation.

------------------------------
Todd J. Gagle
Architect of Software Engineering, Team SYSVIEW
Broadcom
------------------------------

[John Barrett]

Ok thanks Todd. We're using SMF 123 V2 records but yes indeed the volumes were low. Let me put thru' a bank of requests. ​

[John Barrett]

​Magic - thanks Todd. Yeah either issuing a bank of requests or a FLUSH CURRENT did the trick.