Riva On-Premise now offers the ability to deploy a passive-passive fail-over Riva server to facilitate recovering from a crash of the Windows system that hosts a Riva On-Premise primary server. The Riva disaster recovery (DR) strategy requires enabling the DR features, implementing a backup schedule schedule for the primary Riva server, and using the Fail-over Riva server if the primary Riva server fails.
To implement active-active fail-over support, contact the Riva Success Team to schedule professional services to implement that kind of solution. Alternatively, Disaster Recovery training is available: for information, see Access and Request Riva On-Premise Training.
Important Note: The DR feature is not available for Riva for GroupWise. To learn how to recover a Riva for GroupWise server or a Riva server that was not enabled for DR, see Restoring Riva from a non-DR file backup.
The disaster recovery (DR) feature can be enabled on the primary Riva server. Once activated, the Primary Riva server is backed up to preserve the Riva server installation files and the data sync transaction records. If the primary Riva server suffers a non-recoverable crash of the Windows system that hosts the Riva server, Riva sync can be restored from the backup.
The DR feature includes custom options keys to enable CrmID tracking. After the options have been applied, the Riva sync service appends the Riva CrmID to each Exchange item that Riva syncs with corresponding data in the CRM.
When a DR-enabled Riva polls Exchange (or Notes) and the CRM for data, Riva checks all updates (new items, modified items, and deletions) for the existence of the Riva CrmID value:
Riva disaster recovery is supported for the following Riva On-Premise servers based on email platforms:
Riva On-Premise release 2.4.38 or higher is required.
Three scenarios are available:
These scenarios can be combined if corporate disaster restoration policy requires a minimum of two parallel recovery options:
Scenario 1 - Replicating a mirror snapshot of the primary Riva server virtual machine to a different data centre
In this option, a hot mirror of the primary Riva server virtual machine is replicated to an alternate data center. The advantage of this solution is that a backup of the server can be created without stopping the Riva service; and the Riva server in the alternate data center can be started with minimum effort or time lost.
To implement the hot mirror server scenario:
To enable DR options on the primary Riva On-Premise server:
To replicate a mirror snapshot VM:
To restore Riva on the mirror snapshot VM if the primary Riva server VM stops functioning:
To enable daily backups of Riva on the mirror snapshot VM:
Note: The passive fail-over Riva server can serve as the primary Riva server and is already DR-enabled.
To rebuild the primary Riva On-Premise server VM:
In this option, we recommend a timed script that
This activity would normally be scheduled during the Riva blackout period.
The advantages of this solution are that
To set up a daily backup of the primary Riva server to a disabled passive fail-over Riva server:
To enable DR options on the primary Riva server:
To build the passive fail-over Riva server in an alternate physical location or data center:
On the active primary Riva server, perform these steps.
Note: The changes will be copied to the passive DR Riva server.
On the passive DR Riva server, perform these steps.
Note: The objective is to create a mirror copy of the primary Riva server and test connections. Do not install the Riva services.
To schedule a daily backup of Riva data to the passive fail-over Riva server or to offsite storage:
If you have to recover from a non-recoverable crash (loss of) the Windows system that hosts Riva, update the DR Riva server from the latest backup, install the services (if not installed), and enable sync.
If you need assistance with activating your DR Riva server, contact the Riva Success Team.
To restore data sync on the passive fail-over Riva server:
Note: Steps 1 to 4 are not required if you used backup option 3.
To rebuild the primary Riva server, you use a copy of the operational fail-over Riva server.
Perform these steps on the operational fail-over Riva Server:
Note: From this point forward, this passive fail-over Riva server must remain disabled.
Perform these steps on the rebuilt primary Riva server:
Note: The objective is to re-create the primary Riva server by using the RIVA.ZIP from the recently disabled fail-over Riva server. Do not install the Riva services.
This option is recommended if resources are not available for Scenarios 1 or 2. This scenario can also be implemented as a secondary DR procedure to Scenario 1 or 2. In this option, we recommend a timed script that
This activity would normally be scheduled during the Riva blackout period. The advantages of this solution are that
This option would require restoring the backup file set to an alternate Windows host and then starting the Riva server.
DR procedures for this scenario include:
To enable the disaster recovery option on the primary Riva server:
To perform a daily backup of Riva data to offsite storage:
If you need assistance to activate your DR Riva server, contact the Riva Success Team.
To restore Riva data sync to a Windows host server in an alternate location:
These procedures and guidelines can be used in one or more of the DR scenarios identified above:
Standard backup utilities can be used to create and schedule the daily backup of the primary Riva server folder structure. We recommend the following backup strategy:
The backups must include these folders and optionally include these folders:
Backups should be checked daily to ensure that the backup files are valid and can be extracted.
If you are using Outlook Profile MAPI connections in Riva:
If you have created Riva Outlook Profile MAPI connections, consider exporting and storing the Outlook Profile that the Riva connection is using.
Document the Riva service settings:
After the disaster recovery option has been initiated, Riva admins should document and test procedures for restoring data sync. The typical test scenario would involve: