Setting failover rankings for servers and operations
The AR System Server Group Operation Ranking form contains entries that define the failover ranking for servers that handle certain operations in the server group, and the delinquent threshold that helps determine when another server takes over an operation. Data in this form is automatically created when you specify the first server as a member of the server group and restart the server.
When the form is created, it is populated with default entries and the first server added to the server group is assigned the primary ranking for all operations. The remaining server group members have null (empty ranking) entries, serving as placeholders. Entries for operations that require a license (for example, DSO) are not pre-populated unless a valid license is detected. You can add these operations at any time.
Remove the default entries for operations that do not run in your environment.
AR System Server Group Operation Ranking form
(Click the image to expand it.)
The fields named Operation, Server, and Rank work together to define which server is the primary server for the operation, which server takes over if the primary server fails, and so on. The installer populates the fields named AR Server Version, CMDB Version, and Atrium Integrator (AI) Version during the Zero Downtime Upgrade process. In other cases, for example, when the ranking entries are recreated, these fields may remain blank.
When the server is in operating mode, the operation ranks are backed up in to the Operation Backup field. When the operating mode is restored, the backup is restored in the Rank field.
For more information on operating mode, see https://communities.bmc.com/community/bmcdn/bmc_remedy_ar_system/blog/2018/02/27/upgrade-mode
The E-Mail Engine entry in the AR System Server Group Operation Ranking form is no longer used. If this entry exists, it is ignored by AR System Server. Email Engine failover is handled by the service failover operation. For more information, see service failover.
Guidelines to set operation rankings for the server group
Use the following guidelines to determine how to set operation rankings for the server group:
- The servers for any one operation are ranked lowest to highest. A value of 1 indicates the server chosen first to perform the operation. The next highest value indicates the server that takes over the operation if the first server fails, and so on.
- Ranking numbers do not need to be consecutive.
- If a value is null, the server ignores the entry.
- If an operation has no server designated with a valid rank, it is not run on any of the servers in the group.
- Avoid assigning two servers the same ranking for the same operation. (For ease of configuration, the form enables you to do this temporarily.)
- Operations can be spread freely across different servers, with the exception of operations involving BMC Atrium CMDB, and the BMC Service Level Management engine (labeled Business Rules Engine in the form). These operations must reside on the same server as the administration operation; therefore, the operations must have the same ranking as the administration operation so that they move as a unit.
- If you are implementing full text search (FTS), an additional restriction for multi-platform server groups exists. The Administration and Full Text Indexing operations must be restricted to server group nodes that can have a compatible directory structure for the Search Server configuration files.
- Define ranking only for the indexer server. Do not rank non-indexing servers on the AR System Server Group Operation Ranking form.
To establish operation rankings in the server group
- In a browser, log on as a user with Administrator privileges to any server in the group.
- Open the AR System Server Group Operation Ranking form in search mode.
- Perform an unqualified search to see the entries in the form.
- Modify entries as required to construct a fail-over hierarchy for ownership of operations.
- Save the AR System Server Group Operation Ranking form.
- Restart all the AR System servers in the group.
Refresh Ranking allows you to save your changes without restarting the AR System server.
The Delinquent Threshold field determines the number of times the specified server can miss reporting its status before the next server in the ranking takes responsibility for the operation. This setting works together with the Check Interval to determine the total time to failover for any operation.