[Lustre-discuss] Oracle Linux instead of SUSE for 2.0?

Andreas Dilger adilger at sun.com
Thu Dec 3 21:55:22 PST 2009

On 2009-12-03, at 10:21, Lundgren, Andrew wrote:
> I noticed in the announcement that SUSE isn’t support as a server in  
> this release?  Oracle’s distro seems to have taken SUSE’s place.
> Is that going to be the case going forward?

The OEL "distro" hasn't really "replaced" SLES in the 2.0 release in  
any sense.  OEL was first added in the, and SLES10 and SLES11  
are still available there also.  Adding OEL support for Lustre is  
basically a no-op on our part, since it is RHEL5 with a few patches,  
and has no incremental support overhead.  This is similar to  
Scientific Linux and other RHEL-based distros - they can use the  
Lustre RHEL kernel and Lustre modules without any changes.

SLES on the other hand basically doubles the amount of effort needed  
for maintaining the server kernel patches, for a small fraction of the  
users who use it vs. RHEL.  Also, the ext4 code in SLES11 is not  
officially supported by SuSE, unlike the ext4 code in RHEL5.4 and  
eventually RHEL6, so the full burden of ext4 patch maintenance would  
fall upon us, and ext4 is still getting a lot of patches in the  
upstream kernel.

It is important to note that this is only affects the server, SLES  
will of course continue to be maintained and supported on the client.

That said, CVS HEAD already has the majority of patches for SLES11,  
but it will not be an officially supported distro from Sun for the 2.0  
See bug 18668 for details: https://bugzilla.lustre.org/show_bug.cgi?id=18668

We're also looking at reducing the number of kernel patches that  
Lustre needs, so that it is easier to allow Lustre servers to be run  
on less common kernels.  I'd been discussing that internally for some  
months already, and coincidentally had filed a bug yesterday  
describing details of how to remove the patches Lustre adds to the  
core kernel (though some patches would still be needed for ext4).

If anyone is interested to track (or better yet, help) this effort,  
see bug 21524 for details: https://bugzilla.lustre.org/show_bug.cgi?id=21524

Cheers, Andreas
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.

More information about the lustre-discuss mailing list