ADSM-L

[ADSM-L] 6.2 database filesystem on Solaris

2010-04-29 05:18:21
Subject: [ADSM-L] 6.2 database filesystem on Solaris
From: Steven Harris <steve AT STEVENHARRIS DOT INFO>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Thu, 29 Apr 2010 21:17:06 +1200
Hi All

Regular readers will know that I'm now working in  a Solaris TSM
environment.  I'm planning a move of everything from 5.5 to 6.2, and
there's a move from old Sparc to new x86_64 hardware in there as well.

Today's question...

The 6.2 DB2 database must live on a filesystem.  UFS is traditional and
reported to be slow. ZFS is modern but does not write sectors in the
traditional way, but rather writes each changed block to a new location
on disk,  scavenging old blocks as necessary.  this enables it to be
snapshotted and cloned.  One downside is that you can't fill the
filesystem past 80% so there are spare blocks to write changes to.

So what filesystem should I put my database on?  If ZFS it will be one
zpool and one filesystem for each underlying LUN.  How can DB2 be
restricted to only use 80% of the apparently available space?  More
importantly will it perform?

I've looked for information on "TSM and ZFS" and "DB2 and ZFS" and have
found precisely nothing.  Has anyone done this?

Regards

Steve

Steven Harris
TSM Admin
Paraparaumu, New Zealand.

<Prev in Thread] Current Thread [Next in Thread>
  • [ADSM-L] 6.2 database filesystem on Solaris, Steven Harris <=