[Linux-cluster] Advice on Storage Hardware

[Michael Will]

I don't see why that should be a problem, thats a common solution we 
recommend to
customers even without GFS, doing just ordinary NFS and using heartbeat, 
active/passive
is uncritical.You can even have two NFS mounts of separate partitions 
and have an active/active
that fails over the missing one if one of the two machines goes down. 
This means you mount one
from one IP address and the other from the other, and the IP address 
gets migrated over.

This can even be done with SCSI attached storage (4TB per enclosure, up 
to two connected to
a 1U server), but of course the fibre attached storage (direct attached 
as well as a complete san)
is considered more reliable and performant.

Michael Will
Sr. Sales Engineer
www.penguincomputing.com

David Brieck Jr. wrote:
> We're planning on building a cluster to handle web operations but
> we've run into a dilemma on storage. From all the reading I've done it
> appears that the best case would a fibre channel SAN with GFS; however
> we'd like to keep the cost down but still retain very reliable
> storage.
>
> In our cost cutting efforts I've come up with an idea that I think
> would float, but that our hardware vendor (whose trying to sell us the
> SAN) doesn't think is possible.
>
> I'd like to replace a SAN device with 2 servers clustered together and
> connected to an external SCSI array. Only one of the servers would be
> accessing the array, the other would just be a standby in case of a
> failover. The accessing server would use GNBD and GFS to export the
> block device to all clients which of course would run the clients of
> GNBD and GFS.
>
> The documentation seems to imply this is a viable alternative to a SAN
> but I'd like to know if anyone is using this type of setup or if there
> would be a reason why it wouldn't work properly. I would appreciate
> any help since I'm not an expert on this yet. (but should be soon due
> to RedHats courses).
>
> All servers in the cluster will have dual ethernet, etc so hopefully
> bandwidth shouldn't be a problem.
>
> Thanks in advance.
> David Brieck
>
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster
>   


-- 
Michael Will
Penguin Computing Corp.
Sales Engineer
415-954-2822
415-954-2899 fx
mwill at penguincomputing.com