This content has been marked as final. Show 8 replies
The main advanages of active/passive nodes is that
----Node A is active
----Node B is passive until/unless Node A fails
----Only one Oracle license is required
Oracle RAC (2 nodes) supports active-passive mode enabling 1 node to "do the work" whilst node 2 remains comparatively idle,s so here we are not utizing the resources of node2.
Why dunt you use active/active
Active/active failover clusters achieve load balancing. Since at least one instance is active on each node at any given time (assuming no failover), workload can be spread across the servers. The efficiency of workload balance depends on the ability to partition workload across multiple, independent instances.
I am not an OS expert, but as per your current configuration.You are using Active/Passive for OS clustering.but as per the current mode one node will be sitting idle unless and until primary nodes gets failed.It means that oracle clients will be connecting to only primary OS and another node will be sitting idle,s o it means that here we are not utilising RAC features here. I may be wrong
To build a cluster in active/passive mode you don't even need RAC. We still have active/passive Tru64 clusters in 8i/9i for old applications and no RAC : the instance is only part of the Digital (then Compaq then HP...) CAA service.
Beginning with HP-UX on the new servers (Tru64 won't be supported for long) we used RAC but we didn't use McServiceguard. Oracle Clusterware is OK, it is very similar to Digital's even in the commands (crs_stat instead of caa_stat etc.)
McServiceGuard could have enhanced the configuration (port aggregation for network redundancy for example) but a lighter solution like APA can be used.