[Samba] CTDB with 3 nodes and local disks without a dedicated SAN

[Michael Wood]

Hi

On 8 October 2010 13:04, Max <mxx2 at mailspot.at> wrote:
[...]
>> you should try iscsi drbd gfs if you think and mix wisely it should
>> start. THis is, you have on central point of storage where all three nodes
>> write to and this storage backed up.
>
> wouldn't I have one single point of failure then (in terms of HA) or did I
> get something wrong?

Yes.

Maybe you could use a distributed, replicated filesystem like Ceph.
Not sure though.  I've never used ctdb or Ceph.

-- 
Michael Wood <esiotrot at gmail.com>