Re: Two node storage failover with lvm and ISCSI

[Stan Hoeppner <stan@hardwarefreak.com>]

Denny Schierz put forth on 2/20/2011 11:56 AM:
> hi,
> 
> Am Freitag, den 18.02.2011, 20:37 +0800 schrieb Justin Jereza:
> 
>> I'd consider running clvm + gfs2 instead. That way, both nodes can
>> stay up and connected to the same filesystem at the same time. The
>> only decision left would be which node to use. OTOH, you can have an
>> HA configuration as well.
> 
> two points against gfs:
> 
> 1. I red very often, that GFS is very slow
> 1.1 So we have to create files and export them too as ISCSI targets.
> 2. We want to export ISCSI devises for other OS, like Windows, Solaris,
> OSX etc. So we stick on plain ISCSI exports:
> 
> @Stan?
> 
> For what we need drbd? Cloning? We don't need, cause all hosts can see
> the disks.

Maybe I didn't understand your application need correctly.  Can you
describe, in some detail, exactly what you need to accomplish, from a
high level overview perspective (i.e. machine and OS independent--think
flow chart)?  It's not clear to me at this point if you need real time
file/filesystem sharing or simply manual fail over from a dead host to a
backup server.

-- 
Stan