[Beowulf] distributing storage amongst compute nodes

[Loic Tortay]

According to Leif Nixon:
> Mark Hahn <hahn at mcmaster.ca> writes:
>
> >> Some people are running dCache pools on their cluster nodes.
> >
> > that's cool to know.  how do users like it?  performance comments?
>
> Sorry, I don't have much firsthand knowledge about this. I know
> Fermilab is consolidating storage from something like 650 worker
> nodes, and it seems to work for them.
>
On the other hand, as far as I know, BNL (Brookhaven) has decided to
move away from a similar setup (with more computing nodes providing
storage).

If I'm not mistaken, the major reason for this change is
operation/administration oversight/burden.


Loïc.
-- 
| Loïc Tortay <tortay at cc.in2p3.fr> -     IN2P3 Computing Centre     |