Hi all,

I'm exploring the possibility to create a shared storage cluster using
software raid (mdadm).

The idea is to have two normal servers each equipped with a SAS HBA
(non-raid) with external SAS connections (e.g. LSI 9200-8e card)
Then a dual controller JBOD which connects to both servers so that
both servers can see all disks in the JBOD.

Now the interesting part. I would like to create a software raid6 set
(or multiple) with the disks in the JBOD and have the possibility to
use
the raid6 in an active/passive cluster.

For instance create a raid6 set (md0) of 10 disks with the first 10
disks in the JBOD on server1. Create a filesystem on md0. Export the
filesystem
as NFS export (using exportfs).

When doing a migrate of the resources it should stop exportfs, unmount
filesystem, stop md0 (mdadm -S /dev/md0) on server1
and start md0 on server2 (mdadm -A /dev/md0), mount filesystem, start exportfs.

Is this scenario possible and/or obsolete/superceded by another
mechanism because I couldn't find anything regarding mdadm in
combination with
pacemaker except a fairly old raid1 resource agent which only seems to
support raid1 sets.

Any other recommendations in the scenario?

Kind regards,

Caspar Smit
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to