Re: Temporary storage movement over a day -- MIDRANGE-L

Hi Don

On a few systems I look after I make a note of the highwater mark as
sometimes if they are running their storage high (like 90% +) then
the difference between the current temp stg and peak temp stg can sometimes
be enough to actually overflow the system if that high water mark is
reached again. Sometimes the highwater mark is the result of a runaway
query, but I don't generally take chances on that.

I noticed that generally temp stg is more volatile on system running loads
of SQL, particularly where the SQL is being driven out of a java job (that
might be co-incidental), but certainly lots of SQL/QZDASOINIT/QZRCSRVS etc
jobs would have me watching Temporary Storage more closely.

This document
http://www-01.ibm.com/support/docview.wss?uid=nas8N1019501
provides a calculation formula and a guideline that "for most systems, 5%
or less is normal" which I find is kind of true for older more "RPG-based"
systems, but as I said if the system is running lots of ODBC/SQL
particularly applicated generated like out of Hibernate or WebSphere apps I
have seen 10-15% consumed regularly.

The SQL Plan Cache snapshot tool and some of the other SQL tools are useful
for getting a handle on what queries are using a lot of temps stg; I factor
this into my index creation requests. Reviewing an SQL Plan cache snapshot
and then doing a descending sort by the "Average Temporary Storage" Column
is generally a useful guide to me of jobs/queries that deserve some
attention in the temp stg department.

It also pays to monitor the highwater mark and current temp storage as
gradual and constant growth can indicate a memory leak; in this respect it
pays to keep on top of the temporary storage PTFs which are not always just
on the cume.

Hope this helps a bit.

On Thu, Nov 9, 2017 at 4:34 PM, Don Brown <DBrown@xxxxxxxxxx> wrote:

Hello All,

I was wondering if I could get some feedback in relation to the temporary
storage values on busy systems from the WRKSYSSTS display. Both current
and peak.

I have check about 10 systems and all are under 10000 but none of these
systems are very large or very heavily utilised.

One system though has a figure of 130,640 and looking at a performance
graph this grows through the day and decreases overnight - just seemed
rather high to me but may be normal on a busy system. The system has 3450
active jobs currently.

Appreciate any comments.

Thank you

Don Brown

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD