Skip to content

Installation.md

Latest commit

 

History

History
79 lines (55 loc) · 3.66 KB

Installation.md

File metadata and controls

79 lines (55 loc) · 3.66 KB

Jump To Uninstallation

Migrating from v1? See migration

Deployment

Deploying to an AG follows the below footprint.

Examples herein assume we have a two node cluster (primaryNode, secondaryNode) which hosts an AG that is primary/secondary on the respectively named nodes, containing a database called agDatabase.

For a two-replica AG, the scheduler solution is deployed 3 times (number of replicas plus 1), and two auto-upsert jobs are created.

agDatabase
  • Contains the scheduler schema
  • Contains no tasks by default
  • Contains a stored procedure that creates an agent job on the current instance based on tasks in the task table
primaryNode, secondaryNode
  • Contains the scheduler schema
  • Contains a single task (and linked agent job) which will periodically call the upsert procedure in agDatabase

Key here is that the two standalone deployments will both periodically call into agDatabase and invoke the upsert stored procedure, creating agent jobs for that AG on both nodes. Note that this requires the AG to have readable secondaries.

Availability Group Mode

  • Clone the repository
  • Open a powershell session and change to the src folder
  • Import the powershell module that contains the install scripts
  • Deploy the solution against a database in the AG
  • Deploy the solution against every instance which can host the primary replica (not in an AG database)
  • Deploy the AutoUpsert task on all nodes which can host the AG
    • The notify operator must exist on the instance or the job will not be created
  • Deploy the UpdateReplicaStatus job against the AG solution
Import-Module .\tsqlscheduler -Force

# Deploy once in AG mode
Install-SchedulerSolution -Server primaryNode -Database agDatabase

# Deploy once potential primary per-node (needed once per node only, not once per AG)
Install-SchedulerSolution -Server primaryNode -Database Utility
Install-SchedulerSolution -Server secondaryNode -Database Utility

# Create the job on every node that can host the primary
Install-AutoUpsertJob -Server primaryNode -Database Utility -TargetDatabase agDatabase -NotifyOperator "Test Operator"
Install-AutoUpsertJob -Server secondaryNode -Database Utility -TargetDatabase agDatabase -NotifyOperator "Test Operator"

In an instance hosting multiple availability groups, the Utility database needs to contain one AutoUpsert task for every AG.

If the Utility database on each instance is going to be used to schedule instance tasks (rather than just the tasks in agDatabase) an additional pair of tasks would be required:

Install-AutoUpsertJob -Server primaryNode -Database Utility -TargetDatabase Utility -NotifyOperator "Test Operator"
Install-AutoUpsertJob -Server secondaryNode -Database Utility -TargetDatabase Utility -NotifyOperator "Test Operator"

Standalone (Instance) Mode

  • Clone the repository
  • Open a powershell session and change to the src folder
  • Import the powershell module that contains the install scripts
  • Deploy the solution mode against the target instance
  • Deploy the AutoUpsert task on the instance
    • The notify operator must exist on the instance or the job will not be created.
Import-Module .\tsqlscheduler -Force

Install-SchedulerSolution -Server primaryNode -Database Utility
Install-AutoUpsertJob -Server primaryNode -Database Utility -TargetDatabase Utility -NotifyOperator "Test Operator"

Uninstallation

Server objects can be removed with the RemoveAllObjects SQL script. Note that this will remove the objects but not any SQL Agent jobs which have been created.