[Linux-cluster] GFS Based File Server Advice

[Brown, Rodrick R]

If I recall correctly, all nodes in your GFS cluster need to be on the
same kernel rev, so mixing and matching FC4/5 might work, but I don't
think this configuration is supported from Redhat or even something you
might want to do later down the line.

What exactly are you accomplishing by spreading the storage across
multiple systems and using them in a GFS shared file system? It really
doesn't sound like you need GFS here? Are you just trying to keep a copy
of each code repository on each system and provide r/w access from each
node? Unless you're using some kind of distributed source code revision
system that will handle you're locking and consistency checking I
wouldn't advise going down this route you are really just increasing
your chances of file/data corruption. 

If bandwidth isn't a concern I would go with either NFS, or dedicated
iscsi LUNS to each host not under a shared fs. 



-----Original Message-----
From: linux-cluster-bounces at redhat.com
[mailto:linux-cluster-bounces at redhat.com] On Behalf Of HAWKER, Dan
Sent: Wednesday, August 30, 2006 7:39 AM
To: (linux-cluster at redhat.com)
Subject: [Linux-cluster] GFS Based File Server Advice



Hi All,

I have a new iSCSI array arriving soon and would like to carve it up and
use
GFS to provide a central storage pool for a group of servers we use for
building/compiling software.

Each team of developers only has access to their own specific build
server
and as such I don't need the servers to *act as one* however as they all
use
a common data snapshot, a cluster FS would seem a sensible direction to
take.

My developers are coding an unholy mix of Java, C & C++ in the main, and
although they access the data snapshot simultaneously (presently spread
across a couple of servers and dragged around using NFS) they rarely, if
ever are reading/writing the exact same part of the data store at the
same
time. They use a view of the snapshot (in essence a duplicate),
edit/write/etc the code, build and then submit it back to our versioning
software (Rational ClearCase) which is stored elsewhere. 

After reading through the docs and other resources, it seems GFS would
be a
good fit, what with us running a mix of FC4/5 and RHEL4 boxes.

Having read through the docs and stuff it all looks pretty
straightforward,
however thought I'd ask some advice to check for any gotchas and general
*that didn't quite work as anticipated* experiences from ppl that I
presume
use GFS on a daily basis, and hopefully anecdotes from ppl who use GFS
in a
software development role.

TIA

Dan
--

Dan Hawker
Linux System Administrator
EADS Astrium

-- 

This email is for the intended addressee only.
If you have received it in error then you must not use, retain,
disseminate or otherwise deal with it.
Please notify the sender by return email.
The views of the author may not necessarily constitute the views of
Astrium Limited.
Nothing in this email shall bind Astrium Limited in any contract or
obligation.

Astrium Limited, Registered in England and Wales No. 2449259
Registered Office: Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2AS,
England

--
Linux-cluster mailing list
Linux-cluster at redhat.com
https://www.redhat.com/mailman/listinfo/linux-cluster