Oracle Database Migration to Any Cloud (AWS,Azure etc )
Pre-communication (No-Downtime)
- Ensure and agreed with plan & cutover window
- Ensure all stakeholder should be on same page
- Ensure everyone copid in mail and meetings
- Ensure all issue should be documented database & application wise during iteration/testing
- Dicussed and aggred who is from available during cutover winodw (application owner,database lead,infa lead etc)
Pre-Migration Pahse (No-Downtime)
On Source Database
- Delete/Remove all backup tables /unused objects from database with help of application team
- Do username & password backup
- Get read only access to database server
- Ensure db_unique_name is the same as the db_name
- Set and crosscheck all parameter related to dataguard configuration
- check remote connectivity is working fine and port should be open at firewall level
- check and validate database corruption
- prepare & test shared mout point (NFS/EFS/S3FS)
- Take the database backup including archivelogs and standby control file to create standby database
- check and backup networkfiles (tnsnames.ora,listiner.ora)
- check if wallet is configured
- check if any encryption (TDE )is enables
Prepare Target envirnoment (No-Downtime)
- Install oracle binrary software & Build the target database server
- Install Grid software if required
- Install Oracle database binary
- Configure & enable SSH & SCP (with help of root account - Infra team)
- Configure and validate all mount points (database/aplication)
- Configure and validate same shared mount points (NFS/EFS/S3FS) in the target server/s.
- Restore the database backup and configure standby database
- Incase ASM add spfile to grid software using srvctl
- Validate remote database connctivity
- Setup dataguard replication from primary (source) to standby (target) database
Migration pahse (Downtime start)
- check and validate primary and standby database in sync and no GAP
- Stop Application associated database
- create test user/test tablespace and test table on primary database
- Do 4-5 archive log switch on primary (source) database .
- Again check and validate primary and standby database in sync and no GAP
- Check active database sessions in primary (source) database. if still acitve after stopping the application , kill those users
- Create a guaranteed flashback restore point - for this enough FRA space should be available
- Multiple way to do
- Using manual swithover
- Using DGMGRL ( need to configure DGMGRL )
- Using manual Failover (No rollback in this case)
- Validate the migated database
- Setup & validate connection with application
- Configure & update on premise DNS with new IP address (target) and create load balancer rules (if using)
- Handover to application team for validation
Post Migration phase (No-downtime)
- Take golden database backup after successful migration
- Configure & schedule database jobs
- Configure & setup database monitoring
Rollback Phase
- Stop the application on target database
- Rollback the database using restore point (source database)
- Update the DNS with old entry for rollback (if required)
- Start and validate the database
- Connect with database & start the application
