Lustre file system set for spit 'n' polish

Whamcloud, the startup created in July 2010 to continue development of the open source Lustre supercomputer file system, has secured a $2.1m contract from OpenSFS to spruce it up with new features and functions. Lustre – used on about 60 per cent of the largest supercomputers in the world – is a parallel clustered file system …

[Anomalous Cowturd]

Gives new meaning to the phrase...

Brings a lustre to your cluster.

Fnaar fnaar. ;o)

[Anonymous Coward]

re:Gives new meaning to the phrase...

A rather lack lustre joke.

[John Goodwin 4]

Luster or lustre?

I'm confused, I'll get my coat.

[Alan Brown]

You can't polish a turd

Lustre was dumped by several $LARGE UK universities during 2010/2011 for sitewide filesystems access because quite simply, it doesn't scale - and sustained outages aren't good in anyone's book.

In one case they'd been trying to make it work properly for 3 years. There are several million pounds involved and it was at the core of IT services plans.

Real world beats lab testing every time - and there's not enough real world in Lustre.

[Ru]

Yes?

Seems like it is being used here as a filestore for a compute cluster, not user home directories. The use cases (and ease of administration) are quite different.

[Alan Brown]

Usage

Yes, in this case it _is_ being used as a store for a compute cluster (limited size, predictable traffic, etc etc)

It was sold to the universities as being suitable for wider deployment and lustre's own documentation tries to push it in this kind of direction. It doesn't work for that use and it was a pretty epic fail.