Oppolzer - Informatik / Blog

This is a totally different szenario, but we had a problem recently,
where one misbehaving SQL started by DSNTEP2 used up all the
DB2 work datasets doing a large sort, so that at a certain point in time
(after 45 minutes) all other SQL activity on the same subsystem which
required DB2 work datasets - doing sorts - abended with error.

The problem was that the misbehaving SQL was totally unimportant, but the
abending SQL activity was mission critical.

So we looked for a parameter which allowed us to stop a SQL statement
when it has used up a certain percentage of the overall work dataset storage,
and IIRC, there is such a parameter (DSNZPARM), but I don't recall its name
(I'm not the DB2 admin).

Kind regards

Bernd



Am 18.04.2012 12:20, schrieb R.B.:
>
> Last time I heard of this it was caused by an "unrealistic" MEMLIMIT
> setting in the DB2 address space and in my case a batch job with
> REGION=0M. Basically no limit was set and so it could grab all the AUX
> storage it needed and when it dumped it took the LPAR with it. Quickly
> added a MEMLIMIT and all was well...
>
> HTH!
>
> R.B.
>
>
> *From:*J.A.
> *Sent:* Wednesday, April 18, 2012 12:08 PM
> *To:* DB2-L@lists.idug.org
> *Subject:* [DB2-L] - Regressive Testing Impact - query
>
> Hi,
>
> One of our DB Testing is performing Regressive testing on table
> loading(Using DB2CONNECT). Today Our ASM went on shortage due to which
> the Page dataset was totally exhausted and our server totally came
> down. I would like to know if someone has experienced a similar
> situation if yes what diagnosis was done to avoid. Since our
> environment server is capped at 13 MSU and we just have only one DB2
> subsystem.
>
> Regards,
>
> J.
>
> -----End Original Message-----
>
> -----End Original Message-----