Re: IVM IBMi 7.1 Performance -- MIDRANGE-L

"Sub Optimal" yeah I would have to second that!

If our guesses are correct you configuration is also Sub Optimal for IBM i in general hosted or not. LackACache is generally painful no matter how it's stretched and even if you do have cache the four drives are enough for a single small LPAR but that's about it.

To Robs comment about VIOS hosted disk performance, that has been addressed with PTFs to VIOS and I believe IBM I as well. Too long for me to explain completely here (and Sue does it better anyway) but effectively it was excessive I/o to the spindles caused by the 512/520 byte sector handling on 4K drives.

Larry

Sent from my iPad

On Oct 21, 2014, at 6:57 AM, rob@xxxxxxxxx wrote:

I also tried having VIOS host internal disk to my IBM i lpars.
Disgustingly slow. Pretty much unusable.
So on the CEC disk I have 2 mirrored pairs. One pair for one vios server,
another pair for the other.
IBM i hosts all the expansion disk.

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: "Jim Oberholtzer" <midrangel@xxxxxxxxxxxxxxxxx>
To: "'Midrange Systems Technical Discussion'"
<midrange-l@xxxxxxxxxxxx>
Date: 10/20/2014 08:36 PM
Subject: RE: IVM IBMi 7.1 Performance
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>

Roberto's correct to ask about the configuration. If that DASD controller
is the one without any cache it was really intended to manage a couple of
pairs of mirrored disk for VIOS partitions. The storage for the IBM i,
and
AIX/Linux would ordinarily be on a SAN where the SAN does the cache and
such
via a Fibre switch.

If you have the controller with no or very little cache, you can expect
performance on all but very compute intensive tasks to be sub-optimal.

--
Jim Oberholtzer
Chief Technical Architect
Agile Technology Architects

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
Roberto José Etcheverry Romero
Sent: Monday, October 20, 2014 7:10 PM
To: Midrange Systems Technical Discussion
Subject: Re: IVM IBMi 7.1 Performance

Wait,

Are you using internal storage via IVM for the IBM i partition?
Do you have the complete configuration available? You might have made a
RAID5 with a no cache controller (not up to Power8 on my redbook reading
yet) and running VIOS on top of it. I assume you will not be using any
external SAN storage right?

On Mon, Oct 20, 2014 at 5:39 PM, Ted Breedon <tbreedon@xxxxxxxxxxx> wrote:

Hi,

I'm installing IBMi 7.1 on IVM or Integrated Virtualization Manager.
IVM is what you use when you don't have an HMC. Everything is run from
the server, you have a your typical web interface where you manage all
your lpars. I have a S814 8286-41A with four 139GB 6GB SAS Drive in
RAID 5. The lpar has 4GB, 2 dedicated processors. When I ssh into vios
I can see hdisk0 utilization at 100% which makes sense as I'm installing

the OS.

Installing the OS has been terribly slow. I have a Power i5 9406-520
that would install with-in 2 hours but my S814 has been going on for
more than two! I created the RAID Array using these AIX Instructions
and the Standalone Diagnostics Disc.

https://www.ibm.com/developerworks/aix/tutorials/au-aix-raid/
https://www-304.ibm.com/webapp/set2/sas/f/diags/home.html

What's a little perplexing to me is when I've watched videos of IVM
being installed and there's no standalone disc needed but this is what
IBM instructed me to use. As I am writing this I probably should have
used RAID
10 over 5 but should it really be this slow? After I apply CUMs,
should I expect a performance increase or did I do something wrong?

--
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.

--
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.