Re: Temporary space

Dawn is of course correct on tracking it down. However we have successfully worked through issues though on earlier releases. I remember on in V5R4 time where the customer was shutting down *ALL HTTP servers twice a day and IPLing every night just to keep the box up and it had about 50% of 600GB free after an IPL. IIRC it was a Java problem in this case and PTFs resulted that corrected the problem. IPLs reverted to quarterly after the fix.

Definitely one thing to do is get to a point where you plan an IPL but before you do, end jobs and subsystems tracking any reduction in the temporary storage. While not a complete solution this may at least get you to a point where you know of jobs that if restarted can clear significant temp storage.

As has been noted a complete reclamation of temp storage happens only at IPL.


- Larry "DrFranken" Bolhuis

www.frankeni.com
www.iDevCloud.com
www.iInTheCloud.com

On 2/18/2015 11:04 AM, Dawn May wrote:

Tracking down temporary storage usage can be difficult. on 7.1 and earlier
releases.


This is an area where IBM made significant enhancements in the 7.2.
release.

Rather that write the details here, please see these blogs:

System infrastructure changes for improved tracking -
http://ibmsystemsmag.blogs.com/i_can/2014/06/72-improved-temporary-storage-tracking-part-1-.html

System-level interfaces for understanding temporary storage consumption -
http://ibmsystemsmag.blogs.com/i_can/2014/06/72-improved-temporary-storage-tracking-part-2.html

Job-level interfaces for understanding temporary storage consumption -
http://ibmsystemsmag.blogs.com/i_can/2014/07/ibm-i-72-improved-temporary-storage-tracking-part-3.html

Collection services and the performance data investigator support that was
added -
http://www.ibmsystemsmag.com/Blogs/i-Can/January-2015/IBM-i-7-2-Improved-Temporary-Storage-Tracking-(Par/

Part 5 will be on automated notification mechanisms added to monitor growth
of temporary storage use.

Dawn May
www.ibmsystemsmag.com/Blogs/i-Can/



From: John Yeung <gallium.arsenide@xxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Date: 02/18/2015 10:40 AM
Subject: Re: Temporary space
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>



On Wed, Feb 18, 2015 at 9:59 AM, Luis Rodriguez <luisro58@xxxxxxxxx> wrote:

Maybe I'm wrong but couldn't (as the article states) tracking your jobs's
behaviour be a help of how much temp space your system consumes? If so,
*maybe* you could avoid running your most "expensive" (in terms of temp
space) jobs simultaneously...

Luis, it's not that we're close to capacity. There's no point trying
to rearrange "expensive" jobs. None of the jobs is very expensive,
relative to the capabilities of the system. We could easily double or
triple our normal workload, and our machine would not break a sweat.

When looking at individual jobs, nothing looks out of the ordinary.
There are no obvious runaway or hung jobs, and no *obvious* disk
leaks. Yet every week, *TMPSPACE marches upward. The longer the
system is up, the bigger *TMPSPACE becomes. It's not like it ebbs and
flows. It just goes up.

Also, on our system, the "Current unprotect used" is always very close
(within a few percent) to "Maximum unprotect". *TMPSPACE, current
unprotect, and max unprotect basically go up in lockstep.

John Y.
--
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: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.