DeepSee Implementation Guide
DeepSee and Disaster Recovery
[Home] [Back] 
InterSystems: The power behind what matters   
Class Reference   
Search:    

This appendix describes the recommended procedure for write-protecting copied source data on an async mirror member using DeepSee. It discusses the following topics:

Configuration
This section describes the necessary initial configuration tasks.
  1. Set up the async mirror as a disaster recovery (DR) async with all source data databases and the newly-mapped database for ^OBJ.DSTIME. This will perform more validation of the system and push any issues with the ISCAgent and so on to configuration time instead of recovery time. Note that this mode does not allow for a read-write database.
  2. Once configured, switch the DR to a read-only async member.
  3. On a read-only async, each specific database has a ReadOnly flag that can be cleared, allowing writes. Do this for the database containing ^OBJ.DSTIME.
The source data is now write-protected and the cubes can be synchronized properly.
Disaster Recovery
This section describes the steps to take during disaster recovery.
  1. Remove the database containing ^OBJ.DSTIME from the mirror configuration. Note that the database is still available.
  2. Switch the async member back to a DR member.
  3. Promote the member to primary.
  4. Synchronize cubes.
The ^OBJ.DSTIME buffer needs to be treated as out-of-date on any other systems that may now be relying on this one, as there will be no attempt to synchronize that data with other async members. The database containing ^OBJ.DSTIME needs to be added back into the mirror set as part of the recovery procedure.