ADSM-L

Re: [ADSM-L] NetApp for Primary Disk Pool

2012-08-30 08:10:49
Subject: Re: [ADSM-L] NetApp for Primary Disk Pool
From: Richard Rhodes <rrhodes AT FIRSTENERGYCORP DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Thu, 30 Aug 2012 08:03:27 -0400
>We are considering using a NetApp V6210 with some attached shelves as a
>block storage TSM primary disk pool.

Some thoughts . . .

I'm curious, just what is the disk read/write pattern of a
primary disk pool?

You have many, many backup
sessions writing to it concurrently.  The you turn around
are read a few nodes data at a time out if it onto sequential
pools (tape, file).  I know disk pools use a smaller block size (4k)
than file/tape (256k) (TechNote 1240946).  What I've never come
across is how TSM writes/distributes the blocks of a node onto the
volumes of a DISK pool.  I've wondered if it allocates some
kind of extent per session onto a single volume, across volumes, or
whatever.

Then there's the TSM database.  The blocks a node/session uses in
the DISK pool just be tracked/mapped.  So there must be a load on the TSM
database to read/write the blocks.  It would seem that a slow database
could effect the performance of DISK pool access.  This is why I
wonder if TSM uses some kind of extent within the disk pool - not as
much to track.

If the I/O pattern of a DISK pool is random reads/writes (iops intensive),

the the I/O design of the disk system must take this into account.
Throughput (MB/s) is the goal, but to get that you are going to need
many spindles sharing the work load of small block random reads/writes.
This is a very different design than a file pool with large block
sequential I/O.

A V6210 is a gateway filer.  Behind it is probably one or more disk arrays
from the
supported vendor list (quick look:  EMC, Fujitsu, HDS, HP, IBM, Sun, 3PAR
),
as well as NetApp standard disk shelves.  This level of box
is going to be used for many other applications/systems.  Getting
dedicated
spindles is probably not going to happen. There are lots of performance
related questions here!

Oh how a simple question can get complicated!


Rick


-----------------------------------------
The information contained in this message is intended only for the
personal and confidential use of the recipient(s) named above. If
the reader of this message is not the intended recipient or an
agent responsible for delivering it to the intended recipient, you
are hereby notified that you have received this document in error
and that any review, dissemination, distribution, or copying of
this message is strictly prohibited. If you have received this
communication in error, please notify us immediately, and delete
the original message.