RE: Power 6 vs Power 8 backup speeds

Rob,

Would like to add that DASD type can significantly impact backup speed.

From my notes your running 10K SAS, this is a performance hit.

I run production with SSD and R&D with 10K SAS, LTO5 HH FC on both.
There is a huge difference in not only backup speeds, but overall job times.
On production, the tape drives are constantly writing.
On r&D, the drives go idle, waiting on the disk to supply data.

I can gather details if anyone interested.

Paul

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Charles Wilt
Sent: Wednesday, May 14, 2014 4:41 PM
To: Midrange Systems Technical Discussion
Subject: Re: Power 6 vs Power 8 backup speeds

IBM had some pretty impressive numbers at COMMON for SAP on Power (IBM i?) with Power 8...


Charles


On Wed, May 14, 2014 at 3:58 PM, PaultinNZ <paultormey@xxxxxxxxx> wrote:

Hi Rob,
Sounds impressive...could you send me a copy of the spreadsheet.

We're in the throws of reevaluating our JDE application host
environment ie do we go with the Oracle stack or stay with the IBM i.

Any reasons why the Power8 would be that much better (or dare say
worse
than) any equivalent Oracle/Sun or Oracle Cloud engine?

Paul


On 15 May 2014 06:16, <rob@xxxxxxxxx> wrote:

You know, we tried our best to throttle the performance of the Power 8.

It just can't be done.


Instead of each lpar having their own fiber card to the SAN, like on

Power

6, we made them all share one.
We ran the saves on three lpars at once, through the same card.
Virtualized via VIOS.
The switch only supports 4GB at once, even though we tried an 8GB SFP.
We still managed to cut anywhere from 17% to 71.5% off of each
individual lpar's backup time.

The lowest lpar might have been even better but it switched from
being a 'Host' lpar down to being a 'Guest' lpar. That, and it is
still running
7.1TR8 while the other lpars are running 7.2. Overall, it still got
better performance.

Excel spreadsheet with details available upon request.

Tape library used was an IBM 3576, TS3310, with fiber LTO4 drives.
Backup program was BRMS.
Times on spreadsheet are gleaned from DSPLOGBRM *BKU.
All saves done in restricted state.

The boss closely monitored the port on the SAN switch. He saw it
peak

out

at 3.22gb.

You know, when we upgraded our tape drives from LTO3 to LTO4 I had
expected a bigger reduction in save times than we achieved. Who
knew

that

a Power 6 could not drive an LTO4 at rated speeds? Reminds me of
when I had upgraded a CPU a long time ago when we still had 3590's
and the save time rapidly improved.

Disclaimer: IBM did not supervise this test. No warranty that your
results may be anything like ours. Your Mileage May Vary (YMMV).


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

--
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.

--
Paul Tormey
3/11 Bramley Drive
Farm Cove
Auckland
2012
New Zealand
Home: 09 576 3415 or 0064 9 576 3415
Mobile: 021 810615 or 0064 21810615
--
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.