[Pacemaker] DRBD and fencing

[Lars Ellenberg]

On Wed, Mar 10, 2010 at 07:13:28PM +1100, Matthew Palmer wrote:
> On Wed, Mar 10, 2010 at 02:32:05PM +0800, Martin Aspeli wrote:
> > Florian Haas wrote:
> >> On 03/09/2010 06:07 AM, Martin Aspeli wrote:
> >>> Hi folks,
> >>>
> >>> Let's say have a two-node cluster with DRBD and OCFS2, with a database
> >>> server that's supposed to be active on one node at a time, using the
> >>> OCFS2 partition for its data store.
> >> *cringe* Which database is this?
> >
> > Postgres.
> >
> > Why are you cringing? From my reading, I had gathered this was a pretty
> > common setup to support failover of Postgres without the luxury of a
> > SAN. Are you saying it's a bad idea?
> 
> PgSQL on top of DRBD is OK.  PgSQL on top of OCFS2 is a disaster waiting to
> gnaw your leg off.
> 
> > Mmm, you're not:  
> > http://fghaas.wordpress.com/2007/06/26/when-not-to-use-drbd :-)
> >
> > Or is it OCFS2 you're objecting to? We're using this because there are a  
> > few shared files ("blobs" in our CMS) that get written by processes on  
> > both nodes. This is very infrequent, though.
> 
> Split them -- put PostgreSQL on a regular filesystem and mount it before
> starting the database, and run a separate dual-primary for your blobs.

Or, if this is as infrequent as you say it is, have those blobs in a
regular file system on a regular partition or LV, and replace every
"echo > blob" with "echo > blob && csync2 -x blob" (you get the idea).

	Lars

> > Also note that this database will see relatively few write transactions  
> > compared to read transactions, if that makes a difference.
> 
> Cluster filesystems suck at high IO request rates, regardless of whether
> they're reads or writes.
> 
> - Matt