I have reviewed this article:
Do you have documented best practices for using Octopus on a SQL Server cluster? I can see the need in our environment to run things on the SQL Instance and also run things on each SQL node like mentioned at the bottom of the discussion.
Is there a standard way of doing this that is documented?
Thanks for getting in touch. It turns out there was another discussion that goes into quite some depth with some of the same participants: http://help.octopusdeploy.com/discussions/questions/4434-proper-way-to-deploy-to-windows-cluster-with-multiple-resources
We don’t have any documentations or walkthrough on running Tentacle on clusters because it really depends on each scenario as discussed in these threads.
In your particular case speaking about SQL Server:
- anything that can be done with T-SQL could/should be done remotely via the active node (usually by DNS name to the clustered resource)
- anything that needs to be done on each server node in the cluster (regardless of whether those nodes are hosting SQL Server or anything else) you could try installing the Tentacle on each node just as normal Windows Services (not a Clustered Service), one per node. This way each node has a Tentacle, and therefore each node will show up in Octopus Server as individual Machines. In that case you could add a Role like “SQL Server Node” to each Machine/Node in the Cluster and execute PowerShell scripts on them etc.
Hope that helps!