Anyone using ZFS and TSM Server on Solaris?

[mccleld]

Hi Team,

A kind-of follow up to my post from earlier this week on ADSM-L(http://adsm.org/lists/html/ADSM-L/2009-09/msg00115.html) around the use of TSM Server with ZFS volumes on Solaris 10.

I’m curious as to whether anyone else out there is using ZFS for DB, log or stgpool volumes (either raw or volumes in the fs) and in particular how performant, or otherwise, the perception of it is and whether any particular tuning parameters with it have been helpful.

I’ve seen one article on IBM.com which suggests some poor performance with DISK stgpool volumes (http://www-01.ibm.com/support/docview.wss?rs=663&tcss=Newsletter&uid=swg21390367) but it’s not clear to me whether Sun have fixed the bug on their side.

The underlying disk storage here is EMC DMX4 (mix of 146GB and 300GB 15k disks), the TSM Server platform is 5.5.3.0 on a Sun Fire T2000 with 32GB RAM on Solaris 10.

Any input or experiences shared is gratefully appreciated…

Cheers,

/David Mc
London, UK

 

[Bartdo]

ZFS and TSM Server on Solaris

I installed TSM server v5.5 in a Solaris cluster with a specialist from SUN. He choose to use ZFS. We did not have performance issues, but we used a VTL as the main repository for the primary StgPools.
Bart

 

[GregE]

Anyone know if Sun has fixed this?
http://www-01.ibm.com/support/docview.wss?rs=663&tcss=Newsletter&uid=swg21390367

That IBM doc is several months newer than mccleld's post, but still lists that particular Solaris bug.

 

[lnakata]

Anyone know if Sun has fixed this?
http://www-01.ibm.com/support/docview.wss?rs=663&tcss=Newsletter&uid=swg21390367

That IBM doc is several months newer than mccleld's post, but still lists that particular Solaris bug.

Solaris Bug ID 6596237 is addressed in Solaris 10 10/09 ( Update 8 ) or in kernel patch 141444-09 for SPARC and kernel patch 141445-09 for x86/x64. I left a comment on the TSM webpage above to include this info. If you're patched beyond this level or using S10U9, then you should be covered for that particular bug already.