At 11:55 AM 10/4/2000 -0400, Connie Sellitto wrote:
>Hello all,
>
> From time to time over the last few weeks, my 937 has been experiencing
>periods of extremely slow response.  We're running about 22 sessions, 1
>job, with 192 MB memory (additional memory has been ordered). 9.6 GB total
>disk, less than 55% full.   Other than some DB capacity increases lately,
>nothing (to my knowledge) has been changed from a month ago, when it was
>fine.
>
>I've checked TurboImage dataset capacities - at first it seemed to go slow
>during DDX - disabled that; still have a problem.  No one is doing serial
>reads in QUERYNM, or other programs.  Response time may be as long as 10
>seconds.  Then after 10 minutes or so of this, it speeds back up and is OK
>for a few hours.
>
>We don't have HP Glance or other performance tools, and I'm waiting on HP
>response center for feedback - any ideas?

<plug>
You could go to the Millware website at http://www.millware.com and
download a free copy of TheDash.
This will provide you with a tool that will show you the "Top 5 Busiest
Processes".  If you have it running when the problem re-occurs, the cause
of the problem will likely be one of those "Top 5" processes.
</plug>

Obviously, tools such as Glance, SOS, or Probe will be helpful in these
situations.

Howmessy (from Robelle or via Adager) or DBLOADNG (from CSL) will help
diagnose TurboIMAGE problems.

Sudden increases in response time into the 10sec range are often caused by
something within TurboIMAGE.
- If you changed the capacity of a master dataset with an integer key, that
might do it.
- If you have integer keys and the (MaximumKey - MinimumKey) suddenly
exceeds the capacity, that will do it.


Brian Duncombe  [log in to unmask]  http://www.triolet.com
voice: 1-877-TRIOLET (874-6538) (905)632-2773 fax: (905) 632-8704
"Inside every large program is a small one struggling to get out"
C.A.R. Hoare