Re: Processing Time - Consecutive versus Simultaneous Jobs

[Charles Wilt <charles.wilt@xxxxxxxxx>]

Save / Restore time is a function of not only what you are backing up, but
how much unused memory and CPU you have plus disk I/O.

You're running out of one or more of them...probably I/O given that by
using save files you're reading and writing to the same set of disk arms.

WRKDSKSTS / WRKACTJOB / WRKSYSACT would let you figure out which...

Or simply try doing less at once...I'd start with two concurrent libraries.
 It wouldn't surprise me for each individual save to still take more time
than single threaded.  But you might end up with a shorter total time.
 Then again, maybe not.

You could try saving to virtual tape, I believe there's slightly less
overhead than to a SAVF.

However, if you really want to reduce the save window, switch to Save While
Active (SWA).  It sounds as if the "easy SWA" will work.

1) Kick everybody off
2) Start your SWA
3) Once a checkpoint is reached, (5-10min) the system is available.

Charles



On Fri, Jun 14, 2013 at 3:47 PM, <Cynthia.Rogers@xxxxxxxxxxx> wrote:

> We end QINTER and do a SAVLIB to *SAVF (in QBATCH) for our top seven high
> availability libraries before restarting QINTER and allowing users back on
> the system while the remaining libraries are backed up to tape.
>
> We were running the SAVLIB commands consecutively, one after another, in
> QBATCH.  This job started running over into our dayend processing, so we
> decided to submit these SAVLIB jobs separately to QBATCH so they could all
> run at the same time.
>
> We expected each of these submitted jobs to take about the same time as
> they were taking when they were run one after another.  That, however, is
> not what happened.  Our results were as follows:
>
>                             Run Date 6/13/13                        Run
> Date 6/14/13
>                         Running Consecutively                   Running
> Concurrently
> Library                 Start           End                     Start End
>
> PWRDTA          3:00 a.m.       3:36 a.m.               3:00 a.m. 4:23
> a.m.
> BLDDTA          3:36 a.m.       4:00 a.m.               3:00 a.m. 4:10
> a.m.
> PWRFOR          4:00 a.m.       4:02 a.m.               3:00 a.m. 3:09
> a.m.
> BLDFOR          4:02 a.m.       4:03 a.m.               3:00 a.m. 3:04
> a.m.
> IPWCSPRDF               4:03 a.m.       4:03 a.m.               3:00 a.m.
> 3:00 a.m.
> IPWBLPRDF               4:03 a.m.       4:03 a.m.               3:00 a.m.
> 3:00 a.m.
> IPWBSPRDF               4:03 a.m.       4:24 a.m.               3:00 a.m.
> 4:02 a.m.
>
> Our SAVLIB commands are:  SAVLIB LIB(library name) DEV(*SAVF)
> SAVF(HALIB/library name) PVTAUT(*YES)
>
> What am I missing?  Is there anything I can do to speed this process up?
>
> Thanks.
>
> Cindy Rogers
> Blue Line Foodservice Distribution
> 248/478-4213
> --
> 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.