[Linux-cluster] cluster did not fail over when san mount was lost

[Lon Hohberger]

On Fri, 2006-08-25 at 13:23 -0400, Neil Watson wrote:

> Aug 25 11:39:20 caesar clurgmgrd: [5719]: <crit> isAlive returnd: 1  
> 
> Is this a bug or, have I missed something?

isMounted does a mount/path check: is this device mounted where I want
it.  That will pass (even if the fs is not accessible).

isAlive does a very simple accessibility check and returns 1 if the
mount is not accessible (only does a read check if ro is specified).
Looks like it's doing the right thing in this case.

If isAlive returns 1, the file system status check should return an
error status, causing an attempt to restart the service (which probably
fail, and drop the service into the failed state if self_fence is not
specified) if that didn't happen, it's definitely a bug.

-- Lon