upgrade to 0.4.9f log-stonith issues

[Rob]

--- Alan Robertson <alanr@unix.sh> wrote:
> Rob wrote:
> > I installed 0.4.9f(heartbeat -stonith -pils),
> > When I start heartbeat on CLUS1 it sees CLUS2 is
> not
> > alive...  Then it reset's the machine!  If
> heartbeat
> > is not started on CLUS2 should CLUS1 really be
> > resetting it?  (by the way, I don't have CLUS2's
> up
> > attached to CLUS1 at the moment (just in case....)
> so
> > it didn't really reset the machine...  But it
> wants
> > too!)
> 

> The only safe way to take over resources from a node
> you haven't heard from 
> since you booted is to STONITH it.
> 
> In the case described here, each node will wake up,
> STONITH the other, and 
> then get STONITHed by the other node when it comes
> back up.  This may sound 
> dumb, but it's not nearly as dumb as allowing both
> to mount the disk at the 
> same time.

OK, but wouldn't it be more appropriate (remember, I
have a Sun Cluster background here...) to fail to
__start__ if the other node's not alive? (but allowing
an administrator to manually force a start?)  -- 
Second, wouldn't it be advisable to have a forked
daemon which (like Sun Cluster here...) maintains
certain information (if cluster is running) about the
member nodes and their states (i.e.
UP/Down/Maintenance(?)) and maybe machine information
such as cpuinfo, meminfo, uptime, version?

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com