[Bioclusters] Request for discussions-How to build a biocluster Part 3 (the OS)

Donald Becker bioclusters@bioinformatics.org
Sat, 18 May 2002 12:37:39 -0400 (EDT)


On 18 May 2002, Mike Coleman wrote:

> I agree that this is what Scyld is doing.  As a potential user, though, it
> concerns me.  If I run into a problem with one of the many hundreds of
> packages in the Scyld distribution, who do I talk to about it?  I wouldn't
> expect Scyld to have any special knowledge about most of these
> packages (since their expertise is in Beowulf and clustering).  I'd
> prefer to get clustering support from Scyld and distribution support
> from the distribution that I choose.

We have pretty broad expertise, and have created the key parts of the system.
Most distributions are packaging only software created elsewhere, and
have no special expertise with any of the software they ship.

> > Bproc moves complexity into the kernel (48k patch) in order to make things
> > simpler for administrators and users.
>
> This seems like a substantial minus (diverging significantly from the vanilla
> kernel) for what seems like a small plus (with the caveat that I've not
> actually tried Bproc).

BProc needs only a few, quite clean, hooks into the kernel.  The patch
to enable the BProc module to be loaded is quite small.  Compare that to
the many (sometime hundreds) of patches to fix known problems with the
"vanilla kernel".

You should worry much more about other changes: scheduler and VM
subsystem modifications frequently have anomalous behavior when
confronted with unexpected workloads.

-- 
Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Second Generation Beowulf Clusters
Annapolis MD 21403			410-990-9993