Re: SEQONLY(*YES nnn) recommendations on how big a block? -- MIDRANGE-L

Quick, easy, cheap....pick 2 :)

But since Dan specifically mentioned a process that writes 20M records,
it's possible that adding commitment control to just that process wouldn't
be to difficult.

Charles

On Fri, Jul 14, 2017 at 1:32 PM, Rob Berendt <rob@xxxxxxxxx> wrote:

True, perhaps. We found that buying the LPP was significantly less
expensive than modifying all of our vendor supplied programs to do
commitment control. Or even analyzing modifying which ones to get the
biggest bang for the buck.

Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1
Group Dekko
Dept 1600
Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com

From: Charles Wilt <charles.wilt@xxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Date: 07/14/2017 03:27 PM
Subject: Re: SEQONLY(*YES nnn) recommendations on how big a block?
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>

And just to clarify...

"journal caching" offers a big performance improvement for mass
writes/updates to a journaled table.

There's two ways to ensure it is occurring...
1) Use commitment control
2) Buy the LICPGM

Charles

On Fri, Jul 14, 2017 at 1:18 PM, Rob Berendt <rob@xxxxxxxxx> wrote:

Somewhere I had some time trials on
5770SS1 42 HA Journal Performance

Phenomenal difference.

Again, it's an easy install and you've got 70 days to get the key to try
it out. That's probably not the legal promotional way but I think it's
one of those "honest intentions" things.

Pretty much non disruptive unless you like to reinstall PTF's just to
ensure that if any PTF's came out for it you now have the latest and
greatest. Because they sure didn't get applied to it when it was not
installed.
It's one of those confusing things "Yes IBM I am at the latest cume,

hiper

and all groups. However, I've added an option since applying all

those."

Just try to find somewhere in the PTF cover letter where it says that

ptf

such and such was for option ## of 5770SS1. I don't think you'll have

any

luck.

Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1
Group Dekko
Dept 1600
Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com

From: Charles Wilt <charles.wilt@xxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>
Date: 07/14/2017 03:08 PM
Subject: Re: SEQONLY(*YES nnn) recommendations on how big a

block?

Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>

pretty sure those values are still valid...

On your application that writes 20M records to a journaled table, are

you

using commitment control?

If not, re-read the book. :) Note that the Journal Caching PPRQ

mentioned

is now a LICPGM..
5770SS1 42 HA Journal Performance

Rob has posted many times in the past that he got more improvement out

of

the journal caching than going to all SSDs.

Charles

On Fri, Jul 14, 2017 at 12:40 PM, Dan <dan27649@xxxxxxxxx> wrote:

In the "General performance considerations" chapter of the redbook

Striving

for Optimal Journal Performance on DB2 Universal Database for iSeries
(SG24-6286-00), it has the following recommendation:

"Specify Sequential Only on OVRDBF
Where possible, use the CL command OVRDBF to specify sequential-only
processing when adding rows to your database tables. Also, consider

using

separate open data paths (ODPs) and open these for output when adding

large

amounts of rows to your database tables.
Also, specify the NBRRCDS parameter with a value as close to
128KB/row-width as possible to maximize memory buffer utilization."

I have tested this, especially since we turned on journaling and
performance on this one application that writes 20M records to a

table,

and

found a signficant boost in performance. But I note that this redbook

was

written 15 years ago and mentions in the edition notes that "This

edition

applies to Version 5 Release 1 of OS/400". Can anyone advise whether

this

particular recommendation, specifically WRT the 128KB/row-width, still
applies today? (We're on Power 8, v7r1.) If I go with a bigger block

than

128KB, at what point do I need to be concerned with resource

contention?

- Dan
--
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.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD

--
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.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD

--
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.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD

--
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.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD

--
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.

Please contact support@xxxxxxxxxxxx for any subscription related
questions.

Help support midrange.com by shopping at amazon.com with our affiliate
link: http://amzn.to/2dEadiD