Shared, redundant/failover fs (gfs)

I'm trying to build a redundant / failover backend for my cluster.

I have two machines connected using scsi to a promise vtrak15100 cabinet.
The logical drives of the vtrak looks like this:

----------------------------
| LD#1   | LD#2   | LD#3   |
| 200G   | 200G   | 1988G  |
----------------------------

My idea looks like this.

LD#1 holds the OS for backend-server #1 (be1) and LD#2 holds the OS for
backend-server #2 (be2).
LD#3 will be used for the filesystem of my diskless nodes. Planning to use
GFS (if possible).

I use multiple LUNs on the logical drives so that be1 only have access to
LD#1 and LD#3 and be2 has
access to LD#2 and LD#3.

I've started to set up GFS on Fedora Core 4 using only one of my
backend-server. Since this give me no
redundancy/failover I want to get things up n' running on both be1 and be2,
but this is where I have
stumbled into problems. I can't find any good documentation on how to do it
and since this I my first
time working with clusters I really need help.

Hope you guys can help me out a bit here.


Best Regards
Martin


-- 
Ultra Monkey - http://www.ultramonkey.org/
To UNSUBSCRIBE, email to [email protected], with a body:
unsubscribe ultramonkey-users [email protected]
where "[email protected]" is YOUR email address.