Minimizing Downtime for OS patching with Data Guard
x94qviJan 18 2012 — edited Jan 19 2012Hi there,
I'm trying to come up with a solution for a situation that will occur often, and I'm not entirely sure how to go about it. We don't have DG installed, but we will be implementing it very soon.
We will have a primary and standby DB (both 11gR2 Enterprise Edition) setup with an observer for automatic server-side failover.
Our team needs to perform operating system patches on both primary & standby systems every few months, but the 2 systems will not be patched at the same time. So they will patch one system on one day, then patch the other system on another day.
When it comes time to patch the systems we would like to minimize the outage... In fact, I would like to have a situation where clients would be connected to the primary DB, while the standby is being patched, and then perform a switchover and then do it again on the other system - all without too much of an outage for the client (i.e. only the delay during the switchover)
Would the following approach be valid for this type of scenario?
1. stop application of redo to standby
2. shut the standby database down
3. team performs Operating System patching, etc.
4. bring the standby system back up
5. restart redo apply to standby - let it catch up to primary
6. perform a switchover from primary to standby
7. then we repeat steps 1-6, but on the new standby (i.e. the old primary DB)
Is this approach valid? Or is it incorrect/impossible?
If not, how do any of you handle scheduled outages for OS-level patching?