I want to make the fastest migration plan to move a SQL Server database in simple recovery from a standalone server (SQL 2005) to a two node AlwaysOn cluster in full recovery.
Some databases are larger than 200gb so restore them in primary node, switch to full mode and add them to AlwaysOn by selecting full in initial data synchronization is not an option because that would make another backup/restore, taking a lot of time and resources (Time is critical!).
What I want is to restore the database in parallel in all the nodes at the same time (using norecovery I think? ) and then join them to the AlwaysOn group. I have already read the documentation about AlwaysOn and it's not clear about this.
Would it be ok if I follow these steps:
- In Standalone server: Set database to single user mode, switch database to full recovery
- Backup database from standalone server and restore it in primary node with recovery and restore it in secondary node with no recovery
- Add database to AvailabilityGroup by selecting "Join only" in Initial Data Synchronization
- The standalone server is 2005 standard, so it's not possible to join it as a new node.
- I want to make as less backup/restore as possible.
- What I meant with set the database to single user is that there will be no changes in the database during the migration. It's important because I could make a full backup and then I wont need a differential backup.
- After restoring database in Primary node and Secondary node: if I want to join the database to an AG Do I also need to make log backup/restore even though the database had just change to full recovery and no one but me has connected to it?
- Wich option should I choose in Select Initial Data Synchronization Page on the AG Wizard ? Join Only or Skip data synchronization?