Re: Finding reason for high temp storage maximum

[Vern Hamberg <vhamberg@xxxxxxxxxxxxxxxxxxxxxxxxx>]

Well, that's the point of disk/HUNTER - it runs real-time as a 
never-ending-program, waking up at configurable intervals and checking the 
free space. If that free space has gone down by more than a configured 
amount, it does a number of things, including which jobs are using the most 
temporary storage at the time.

E.g., if you have 70% of a TB used, you'd have 30% free space, or 300GB. A 
1% trigger would notify you when the free space went down by 3GB - far from 
the 90% threshold you want to avoid.

The command I mentioned is used by the product when the trigger is 
exceeded. It can also be run independently as needed.

Sorry for what seems a long commercial, but I believe it addresses very 
well the situation you describe.

Thanks

Vernon M. Hamberg (Vern)
Senior Software Engineer
Centerfield Technology, Inc.
507.287.8119 x111

At 06:56 PM 12/4/2003 +0100, you wrote:
> On 4 Dec 2003 at 10:14, Vern Hamberg wrote:
>
> > Oliver
> >
> > Our disk/HUNTER product has a command with which you can get a snapshot of
> > the top 20 jobs by temporary storage.
>
> It is not so much a question of finding the culprit jobs (mostly, index 
> builds are also
> consuming lots of CPU), but catching the temp storage growth before it 
> hits our ASP1
> threshold (at 90%).
>
> Beyond SNADS, what else will be shut off when the threshold fires?
>
> Regards,
>
> Oliver