[Kolab-devel] KDE/Kontact unsuitable for NFS? (was: Debbuging help for unreliable Kontact needed)

[Martin Konold]

Am Montag 04 Juni 2007 schrieb ITSEF Admin:

Hi,

> I'm a bit surprised, to be honest. Given that almost any company I've been
> working at had $HOME mounted via NFS, be it for back-up reasons or because
> people need to be able to access $HOME transparently from different
> workstations, $HOME on NFS seemed to be quite standard to me as setups go.

Yes, and they regularily cause problems.

http://www.exim.org/pipermail/exim-users/Week-of-Mon-19961230/001113.html

"Basically, you can't reliably lock a file on an NFS mounted system.  :-("

http://www.redhat.com/magazine/008jun05/features/gfs_nfs/

"POSIX semantics ... 
NFS:		No, reads may or may not get last data written
GFS/iSCSI:	Yes, reads always get data that was last written, does not break
                applications"

http://www.opengroup.org/austin/mailarchives/ag/msg10343.html

"> > That is well known.  For example, "In general NFS is very problematic
 > > as NFS does not provide POSIX filesystem semantics."
 > > <http://intevation.de/pipermail/kolab-users/2004-June/000505.html>.
 > 
 > Thank you for proving that this statement is just FUD!
 > 
 > Someone who claims that NFS is not POSIX compliant should be able to 
 > point to the deviations from POSIX semantics. Let us safely asume that 
 > NFS is POSIX compliant until someone describes real deviations.

 Are you aware that there is an Open Group specification for NFS,
 called XNFS?  It describes the ways in which NFS filesystem semantics
 differ from POSIX/SUS in detail."

> these environments? What would you recommend instead of NFS for
> environments where $HOME needs to be accessible on all workstations?

e.g. AFS or GFS.

Regards,
-- martin konold

-- 
e r f r a k o n
Erlewein, Frank, Konold & Partner - Beratende Ingenieure und Physiker
Sitz: Adolfstraße 23 Stuttgart - Partnerschaftsregister Stuttgart PR 126
http://www.erfrakon.com/