That Metalink note documents a RAC problem, not an ASM problem. Most
service-group oriented clusters (Veritas VCS, HP/UX Service Guard) have
the same behavior, which is by design, not a bug. (The behavior being,
if a resource goes off-line, the other resources which have been defined
to depend on that resource also go off-line).
In the situation documented in the note, only the one node is affected,
not the other nodes.
Anyone who stores their only backup inside of the production ASM
instance is violating the golden rule of backups and needs to go back to
DBA school. Sure, put the Flash Recovery Area on ASM, but copy it
somewhere else, too.
-Mark Bole
Kevin Closson wrote:
Metalink Note:277274.1. Period.
------------------------------------------------------------------------ From: John Smiley [mailto:jrsmiley@xxxxxxxxx] Sent: Wednesday, June 22, 2005 1:39 PM To: kevinc@xxxxxxxxxxxxx Cc: oracle-l@xxxxxxxxxxxxx Subject: Re: 64 node Oracle RAC Cluster (The reality of...)
While we're on the topic of ASM, how many are using it in production
today? After hearing the horror stories about how bugs in ASM
caused it to basically lose its mind and forget where everything was
left me wondering if I'll ever trust this product with my data. Some poor souls followed Oracle's advice and stored their backups in
ASM along with their database and lost everything.
John Smiley
Technical Management Consultant
TUSC, Inc.
-- //www.freelists.org/webpage/oracle-l
