SQLTeam.com | Weblogs | Forums

Disaster Recovery for Failover Cluster Instance SQL server database

Hi all,
I have 1 FCI SQL Server production contain 2 node (nodeA, NodeB), and data we keep in SAN (map to nodeA, NodeB).
Now i don't know, how to make high availability for this FCI like Always On Availability Group.
If all node (A, B) down or SAN problem. We can't using database for production. Please support me any ideas or comment for this case?

hi

please see the below link

hope it helps :slight_smile: :slight_smile:

i love feedback

https://docs.microsoft.com/en-us/sql/database-engine/sql-server-business-continuity-dr?view=sql-server-2017

Hi harishgg1,
Thank you very much for your information. When i try to install 1 node - NodeC (stand alone SQL), after that enabled Always on High Availability in FCI (NodeA, NodeB) to join NodeC in AAG (Always on Availability Group), however after add node C to Failover Cluster Manager, FCI can't running. Please support me, how to join exists FCI (NodeA, NodeB) and stand alone SQL (NodeC) to AAG?
I follow this instruction for add new SQL NodeC to exists FCI.

Note: I try to input link to this windows, but i don't know how to add link?

Google search for
Sql server
How to add node high availability

Please see the videos

Hi harishgg1,
i follow this instruction to add SQL to AAG, however can't running.

https://www.mssqltips.com/sqlservertip/3150/adding-sql-server-alwayson-availability-groups-to-existing-failover-clusters/

hi

sorry i dont have REAL TIME experience

i have never done this
:slight_smile: :slight_smile:

please post in SQL Server Administration Forum

To add AOAG to an existing FCI cluster:

  1. Add node to cluster with non-shared storage that matches the shared storage on the FCI nodes. That is - if you have a drive defined for user database files sized at 100GB you need that same drive letter with the same size on the new node.

  2. Do not change the cluster quorom - you should already be setup with quorom with some type of witness - either disk or file share.

  3. Modify the node weight on the new node to be 0 - this prevents this node from impacting quorum and causing a cluster outage.

  4. After the node has been added to the cluster and verified (without verifying storage - as that will fail due to non-shared storage) - install stand-alone instance of SQL Server and patch to same level as FCI instance(s). If the installation fails because it cannot find the storage:

4a) If the storage is not available because the cluster sees the non-shared storage - evict the node from the cluster. Install SQL Server as stand-alone instance and then add node back into the cluster.

  1. Once you have the new node installed and ready - you can create the availability group just the same as if you were creating it between 2 stand-alone instances. The secondary should be setup with manual failover - and asynchronous mode.

Note: if the requirement is no data loss for DR, then you would need synchronous mode, but that will require transactions to commit on the secondary prior to being committed on the primary. If there are any delays in network traffic between the FCI nodes and the DR node - you will impact performance on the primary nodes.

Hi jeffw8713,
Thank you very much for your instruction.
I already installed new stand-alone sql database to new node. After that, i enabled Always on High Availability Group by stick check box for both exists FCI and New Stand-Alone SQL.

After I added new node for stand-alone SQL to exists FCI by Failover Cluster Manager, FCI SQL services auto stopped and failover to new Stand-alone SQL just added to node.

Please support me, how to keep exists FCI SQL still running after added new node for stand-alone SQL?

Not sure I follow - adding a new node to a cluster would not have SQL Server as a possible owner of the SQL resources, therefore it could not failover to that node.

A stand-alone instance on the new node would still not be able to take ownership of the SQL resources in the cluster either...

Unless you installed SQL Server into the cluster on that new node - it would not be available for failover - so it appears that you did not install a stand-alone instance on the new node, rather you installed SQL Server into the cluster on the new node.