[Linux-cluster] Oracle failover cluster, whole /u01 on shared storage?

[Cristian Mammoli]

I'm not talking about rac and/or gfs, just a plain old active/passive 
cluster. What's the right way to implement it?
As far as I understand the easiest way is to put the whole $ORACLE_HOME 
on shared storage (SAN) and mount it on the active node.
But in this way I loose the chance to do "rolling updates":
Update Oracle software on node2, migrate instance to node2, update 
software on node1 ecc

If install Oracle software on each node and put database files on shared 
storage (eg. /u02) there are some files which reside inside $ORACLE_HOME 
anyway (spfile, listener, audit trail ecc). How do I keep them in sync? 
Symlinks?

Thanks
-- 
Cristian Mammoli
APRA SISTEMI srl
Via Brodolini,6 Jesi (AN)
tel dir. +390731719822

Web www.apra.it
e-mail c.mammoli at apra.it