I have a single 10g instance on HP-Unix and i am looking for options of High availabilty for this database. I DO NOT want to use RAC or Dataguard.
What i have in mind is - a shared storage (SAN) and 2 instances in 2 seperate servers.
Please suggest other options/mechanisms to achieve this. Thanks.
Dataguard is for disaster recovery and not for HA.
By two instances on 2 separate servers you mean active/passive cluster?
This won't give you real HA as you will have some downtime when you failover.
The only way to achieve true HA is RAC....which is two instances on two separate servers with shared storage !!
Thanks for you reply.
Yes i am looking for HA. Minimum downtime is acceptable. Not looking to implement RAC.
Can you help me out move with active/passive cluster idea ? What options do i have here.
There is not much of an option to be honest. I've seen it implemented at in SME's where they don't want to have to pay for a RAC license, but it's very rare indeed.
Whatever clusterware you use will do the failover for you (Veritas cluster or HP ServiceGuard). The filesystems will be clustered and failed over in the event of a failure and work like the old fashioned cluster.
You will need to test to see how much downtime you have and if this is acceptable.
Bear in mind that there are licensing implications for this type of cluster. From memory you had the use of the second node in your cluster for xx days or so before you had to pay licenses for that server as well (I can't remember the exact details and this may well have changed anyway). However licensing costs should be discussed with your friendly Oracle sales rep !!
So which cluster mechanism will you suggest for this single-instance HP box.
Also, any other options apart from Veritas cluster or HP ServiceGuard. (Looking for the option which is best, cheap nd easy to implement).