How can i simulate Failover test between Primary RAC instance to single instance running on 11gr2 with both ASM running. Please someone can provide me the correct step to simulate. This is my first experience conduct test on Production server to DR server. I have configure dataguard broker in between.
816734 wrote:In this case you have to forget of your Primary RAC. :)
Thanks for the feedback. Actually i request by managment to test FAILOVER TO dr server when our primary RAC db was totally fail to recover . I need to test Failover to DR server using dgmgrl. Can you provide me some example
816734 wrote:These steps are depend on the availability of your RAC primary. Lets suppose whole cluster is not accessible, do you think you can execute "SRVCTL" commands?
I need some experience people verify my step as per below by using dgmgrl for failover test. My primary RAC db call nios1 and nios2 and my single DR server is niosdr. Hope someone can give me advise for simulation
STEP - connect from primary RA db
1. Disable automatic restart in Primary RAC db - srvctl disable database -d nios
2. ps -ef | grep pmon | grep -v grep - Kill pmon id on both nios1 and nios2
3. dgmgrl > show configuration
4. dgmgrl > failover to 'NIOSDR'
816734 wrote:If you would like to keep your primary as active, until you perform failover the primary database keep sending data to the standby database. Once you perform failover the new incarnation will be started and the standby role will be changed to primary role and no more redo will be received or applied. of course once you perform failover either you can stop the database or even you can disable from the configuration. That is what you looking for?
Thanks for the update. Can i just using srvctl stop database -d nios on primary RAC for shutdown database on primary db and straight go to DR server run dgmgrl > failover to 'NIOSDR' for simulate the test. There is not need to kill any backgroup processes. We asume that the primary db still accessible and all the log has been apply to standby db.
816734 wrote:No need to restore. You can use reinstate, ensure flashback is enabled. Refer link http://docs.oracle.com/cd/E11882_01/server.112/e17023/cli.htm#BGBBDEBC
Thank for the update. This is what i looking for if the primary still available for access without issue. If i fail over to my standby db (NIOSDR) and old primary will need to restore back right ?? I can use reinstate ??
What happen if i perform dgmgrl > switchover to niosdr -> dr server . After sucessfully switch role, then issue dgmgrl > failover to 'NIOS' ->primary db .1) If you perform switchover, then niosdr will become new primary database and nios becomes as new standby database
Do we still need to perform anything ?? This scenario is assume that both primary db and DR server is accessible without issue.After performing above role transition, Still you willl have two primary databases nios and niosdr , Then again you have to reinstate the database to covert it to standby role.
Thanks for your explanation. To perform dgmgrl > failover to nios, this need to login on primary rac and run dgmgrl or standby database ??Issue the "Failover to nios" from the target standby database you want to change to the primary role.
As you mention we have two primary database nios and niosdr, so we need to reinstate niosdr to become standby role. It is correct. what is the stepBroker performs failed primary database as standby database, Perform the below steps