[DRBD-user] Write order

Lars Ellenberg lars.ellenberg at linbit.com
Wed Jun 22 00:20:54 CEST 2011

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


On Wed, Jun 15, 2011 at 11:47:20PM +0900, Motoharu Kubo wrote:
> Hi Lars,
> 
> > If your file system or application is not crash safe,
> > DRBD cannot magically make it so, no matter what,
> > regardless of replication protocol.
> 
> I do understand the point.
> 
> For our good usage, could you give me some examples of crash-safe and
> non-cclash-safe filesesystems, or good pointer to learn more?

"crash-safe file system": hopefully all of them.

Of course they will need some sort of crash recovery procedure.

The two main categories are:
non-journaling file systems (e.g. ext2) need a full fsck run,
where the journaling file systems (e.g. ext3) only need a journal replay.

Anything that has a reliable crash recovery procedure to recover from
hard crash on a single node can safely run on a DRBD based failover cluster.

Anything that has serious problems to recover from a hard crash,
even on a single node, won't really get better by adding DRBD and failover.

-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com



More information about the drbd-user mailing list