Hello!
On Sun, Aug 12, 2007 at 08:23:34AM -0600, Robert LeBlanc wrote:
> We need to have MySQL fail over and since we already have a low traffic LUN
> set-up with Lustre that is shared and fault-tolerant we want to put the
> database files on it instead of carving out another LUN for the low traffic
> database. We are only going to access the files from one node at a time.
Ah, I see.
So you do not really need full cluster-wide flock support, it seems.
> Any pitfalls I should know about if I enable locking on two nodes like large
> performance degradation? I take it that the localflock is a client option
No, the only pitfall I am aware of are some bugs, if you need cluster-wide
locking support, you need to apply patches from bug 13103 and from bug
5135 (to 1.6.1 lustre release).
> and the flock is an OST mount option?
Both localflock and flock are client mount options.
Bye,
Oleg
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
