Re: Disk: Percent full a performance factor?

Ok, granted, I went from a mix of SSD's and
HDD's to pure SSD's, of a newer generation.

Right. My mileage got like four times better when I started buying gasoline in 2008 from a new station. Yes I began putting it into a Prius instead of a V8 4x4 Dodge Ram at the same time but I think it was just better gas :-)

Seriously you're likely behind better cards too and with faster SAS buses maybe even in faster PCI slots. Of course all that matters, but the one thing you gotta do is get each LPAR a minimum of 6 drives real or virtual. Want to prove it to yourself, create a single 1TB NWSSTG and load everything to that and see how it runs/walks/crawls....

Larry

Sent from my iPad

On Oct 21, 2014, at 7:05 AM, rob@xxxxxxxxx wrote:

<snip>
It depends on how many DASD devices there are. If there are say 40-50
devices then running up to 85% -- 88% might not be a problem, but with 6
devices, you're in for a hard time.
</snip>

Is this so true anymore, about the number of disk arms being so critical?
I hear people speak that mantra all the time yet I've found that the newer
disks, controllers, etc have always amazed me. Many years ago we dropped
a system down from 42 disk arms down to just 7 of the then newer drives
(still spinning disks) and my performance was significantly better.

Recently I've dropped from some lpars each having dozens of their own disk
drives down to 21 storage spaces on a system with only 32 drives (and that
was only 1 of the 5 lpars being guested from this host) and the
performance was much better. Ok, granted, I went from a mix of SSD's and
HDD's to pure SSD's, of a newer generation.


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 02:54 PM
Subject: RE: Disk: Percent full a performance factor?
Sent by: "MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx>



Rob,

It depends on how many DASD devices there are. If there are say 40-50
devices then running up to 85% -- 88% might not be a problem, but with 6
devices, you're in for a hard time.

Generally Databases tell you the knee in the response time curve starts at
about 80% and steepens quickly but with IBM i, as long as the DASD
device
busy percentages are OK and there is room for temporary indexes, then that
percent full you have to be concerned with goes up as well, so higher
percent full is OK.

--
Jim Oberholtzer
Chief Technical Architect
Agile Technology Architects


-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of
rob@xxxxxxxxx
Sent: Monday, October 20, 2014 1:29 PM
To: midrange-l@xxxxxxxxxxxx
Subject: Disk: Percent full a performance factor?

The old rule of thumb I had learned back on the S/36 was that you really
didn't want to go over 80% of disk space used. I had a manager (came from
the programmer pit and rose through the ranks) that wondered if that still
held true with the growth of disk. After all, 20% free of a whole gig of
disk space sure was a lot more than 20% of 300MB. I argued that object
size
was significantly larger and that reorgs and whatnot still needed
significant portions of disk space.
Now that we're measuring disk in TB and MB is nought but a rounding error
does 80% still hold true? If so, why?


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.


--
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.