[Linux-cluster] architecture discussion -- NFS clustering with iSCSI

[Terry]

On Jan 28, 2008 2:11 AM, Gordan Bobic <gordan at bobich.net> wrote:
> Rainer Duffner wrote:
> >
> > Am 28.01.2008 um 02:36 schrieb Terry:
> >
> >>
> >>
> >> Good questions:
> >> 1) Do you have concurrent writes to the same file from different nodes?
> >>   1a) No
> >
> >
> > Well, that's one of the things GFS is good at ;-)
> >
> >
> >> 2) How many nodes do you have?
> >>   2a) 3 to start, probably won't go beyond 12
> >>
> > OK, that's still in the range GFS can handle (AFAIK).
>
> With an order of magnitude room for growth left.
>
> >> I appreciate alternative ideas to NFS.  NFS could possibly introduce
> >> performance issues (comments here appreciated).
> >
> >
> > One problem might be that NFS was never supposed to run on  GBit-networks.
> > Thus there is overhead.
> > But, OTOH, the vendors I mentioned have managed to squeeze a lot of
> > performance out of NFS.
> > It's also a question of optimizing/matching NFS clients and servers.
>
> I've found that NFS v3 over UDP with large rsize,wsize and jumbo frames
> works pretty well.
>
> >>   The majority of the
> >> system is write.  I would say 80%.
> >>
> >
> > Do you have a lot of small files?
> > Small files are usually what degrades GFS-performance.
>
> I don't think small files are what kills it, it's lots of files that
> slow things down.
>
> Gordan

In an active-standby NFS cluster scenario, is GFS still required as
the format for the data drives?  I would be using a 2 node cluster so
I would need a quorum disk too.  Since only 1 node will have access at
any one time, could I get away with ext3 (or whatever) formatted
volumes?

Along the same note, I am going to ask an obvious question and answer
it myself.  :)   In an active-active, this would definitely need to be
GFS formatted, correct?

Thanks!