Just implemented an new OPS (8.1.6) system on HP (3-node v2600 cluster).
Lots of headaches, many oracle patches and hp patches over the last 3 months
to keep things running. Cluster has only ever stayed up for 10 days (max).
Plus, can't do rolling upgrades of each node. The entire cluster has to
come down. Stay away from OPS if at all possible. It is easier to create a
standby database or implement some kind of HA failover setup, rather than
deal with the complexity of OPS.
Received on Thu Jan 11 2001 - 07:14:05 CST
