Re: [Jfs-discussion] JFS in SuSE Enterprise 9

[Sonny R. <so...@bu...>]

On Wed, Mar 16, 2005 at 07:50:26PM -0500, Chris Penney wrote:
> I'm considering using JFS for an NFS server and I'm curious if anyone
> has any feedback on any of the following questions.  The hardware is a
> dual cpu pentium 4 with 2gb ram, gigabit ethernet, and dual qlogic
> 23xx hbas running SuSE Enterprise 9 w/ SP1.  I'm using lvm2 to turn
> four 1TB luns into a 4TB volume.

Ok, since Shaggy hasn't piped in I'll try and answer your questions. 
 
> Questions:
> 
> 1) With service pack 1 applied to SLES 9 is there anything I'd be
> missing in terms if important JFS patches?

SLES 9 SP1 should most if not all of the recent bugfixes.

> 2) Are the default mkfs.jfs options ok?  There doesn't seem to be any
> options to specific stripe widths, etc.  I assume it autodetects or
> doesn't need that?

Yeah JFS doesn't preallocate groups or aggregates at mkfs time so
there's really not much to specify unless you need to set up an
external journal or set a volume label.  

> 3) Is there any reason not to use jfs for an NFS server?

There shouldn't be, we've used it fairly extensively on NFS servers.

> 4) Are there issues with SMP systems + JFS + NFS?

Again, there shouldn't be, we've run JFS on many different
servers/archs without problems.

> 5) Are there any tuning options I should consider?

If you're typical workload is very meta-data intensive with lots of
threads etc. you might want to increase the nTxLock or nTxBlock module
parameters, but in general it shouldn't be necessary.  

Sonny