[Gllug] Oracle Shared-Nothing

[Bruce Richardson]

On Mon, May 12, 2008 at 02:43:24PM +0100, Martin wrote:
> Stephen Nelson-Smith wrote:
> > My own personal bias is that I don't trust drdb - what happens if it
> > stops working?
> > Yes, of course you would monitor it.
> 
> When stuff stops working, it stop working, that's why you have HA.
> 
> I worked at a company that used drdb for about 18 months, i don't recall 
> a single drdb failure.  Hardware failure, yes, but that was handled by 
> HA infrastructure.

John is not somewhere he can reply, so here is his reply by proxy:

"Yes, you are correct. Anything is a possible source of failure.  My
skepticism is just prejudice, as you say. I was involved in a project
several years ago but where i recommended shared-RAID as the common disk
between an HA pair.  But the system had already been specced up, and I
had to use drbd.  As I say, my fear is that there is a silent failure,
and the two systems get out of sync - and when failover happens...  Yes
of course this should be monitored, and restarted if the process fails.
Its when this stuff is being packaged up for, shall we say
unsophisticatd users and as a turn-key system then it concerns me."


-- 
Bruce

Those who cast the votes decide nothing.  Those who count the
votes decide everything. -- Joseph Stalin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 196 bytes
Desc: Digital signature
URL: <http://mailman.lug.org.uk/pipermail/gllug/attachments/20080512/ca473c8a/attachment.pgp>
-------------- next part --------------
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug