Re: NFS & Clustering

["Ragnar Kjørstad" <nfs@ragnark.vestdata.no>]

On Wed, Sep 04, 2002 at 09:40:08AM -0700, Bryan O'Sullivan wrote:
> On Wed, 2002-09-04 at 09:24, Emanuel Quass wrote:
>=20
> > but doesn't it work if you share /var/lib/nfs/* (containing etab, rmt=
ab, sm, =20
> > sm.bak, state, xtab)?
>=20
> Not reliably.  You're left with a single point of failure, for one
> thing.  You also have races between one server permitting a mount and
> others knowing that the mount has been allowed, which can confuse
> clients.

I suppose you're suggesting that the disk with /var/lib/nfs/ is a single
point of failure? Well yes, unless you use a redundant storage-system.


The race-condition is only present when the nfs-service is enabled on
both server at the same time. For active/passive failover you have
software that monitor the health of the the primary-system and the
secondary system is only started up if the primary fails. So, there is
no race.



--=20
Ragnar Kj=F8rstad


-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone?  Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs