RE: ASYNCBRING performance -- MIDRANGE-L

Mike,

Currently, 230034 objects saved. 89 not saved.
Not sure of the count back in 2012.
Our IFS doesn't change drastically, so around the same number.

I found some other notes from May 2013 that impact IFS save times, SWA and its various options.

IFS full test saves from a BRMS control, with and without SWA.
SWA on simply lengthens the IFS save time, does not improve the IFS objects in use (with locks).
Also, we have confirmed that most of these locks are not exclusive, shared read and/or shared read/write.

IFS full save with SWA *YES - 242902 objects saved. 192 not saved. 36 minutes
IFS full save with SWA *NO - 242902 objects saved. 192 not saved. 17 minutes
IFS full save with SWA *SYNC - 242902 objects saved. 192 not saved. 25 minutes

Paul

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Mike Krebs
Sent: Friday, February 14, 2014 7:00 PM
To: Midrange Systems Technical Discussion
Subject: Re: ASYNCBRING performance

Paul,
How many objects in the SAV portion?
Mike

On Friday, February 14, 2014 3:27 PM, "Steinmetz, Paul" <PSteinmetz@xxxxxxxxxx> wrote:

Mike,

Here are my notes from when we implemented ASYNCBRING.
Saved 1 hour to 1 1/2 hours
740-E6C
2 processor active
256 gb memory
SSD with 5913 controller
3573 with LTO5 HH fiber drives

BRMFULL05 *NONE 0 4/24/12 21:00 4:25 7:25 *NORMAL (NO ASYNCBRING)
BRMFULL05 *NONE 0 4/28/12 21:00 3:47 6:47 *NORMAL
BRMFULL05 *NONE 0 4/29/12 21:00 2:52 5:53 *NORMAL (ASYNCBRING)

Paul

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Mike Krebs
Sent: Friday, February 14, 2014 4:06 PM
To: 'Midrange Technical Discussion'
Subject: ASYNCBRING performance

Executive summary: ASYNCBRING saved 20% of SAV time - about 6 minutes on our box.

A couple of earlier threads mentioned using ASYNCBRING on SAV commands. I don't remember seeing actual performance numbers but the concept sounded interesting. I ran a test using our normal nightly backup procedure.

Testing was carried out on a 720 with one processor active and 8GB main memory to a Ultrium 5 tape drive in a dedicated backup procedure that uses ENDSBS(*ALL) with a BCHTIMLMT to achieve dedicated status. The SAV part of the system did not change much over the time period although there was slight increase in number of objects saved.

14 Backup Averages
Obj       Saved SAV time
Base
473418
1954
Asyncbring 490053
1578
Change 16635
-376
Percent 3.51%
-19.24%
The SAV time is in seconds. The base backups were taking 33 minutes. We saved just over six minutes per backup using ASYNCBRING while saving 16,000+ more objects.

Procedure to measure...
Using TAATOOL CVTQHST to save the QHST data from the backup job. Match the SAVDLO (CPC9410) message and the SAV (CPC370C) message to create the SAV interval.

with start as
(SELECT * FROM backuphist where msgid = 'CPC9410'),
     end as
(SELECT * FROM backuphist where msgid = 'CPC370C') SELECT start.yymmdd,
       start.TIME,
       end.TIME,
       TIME(Substr(end.TIME, 1, 2)
            || ':'
            || Substr(end.TIME, 3, 2)
            || ':'
            || Substr(end.TIME, 5, 2)) - TIME(Substr(start.TIME, 1, 2)
                                              || ':'
                                              || Substr(start.TIME, 3, 2)
                                              || ':'
                                              || Substr(start.TIME, 5, 2)),
       Substr(end.msgtxt, 1, 9)
FROM   start
       JOIN end
         ON start.jobnam = end.jobnam
            AND start.usrnam = end.usrnam
            AND start.jobnbr = end.jobnbr ORDER  BY 1,2

This gave results like this:

Date as Time    Time    Numeric Expression SUBSTR YYMMDD
131229   033433 040638          3,205       474624 ob
131231   034221 041300          3,039       461412 ob
140101   115042 122126          3,044       461782 ob

Which in turn were deciphered in excel to turn 3205 into 32 minutes and 5 seconds = 1925 seconds. Last column is number of objects saved.

Conclusion:
ASYNCBRING is worth considering adding to your SAV commands.
--
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.

--
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.
--
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.