Worker Server GRs
Worker Server GRs contain a series of rules that control which work is permitted to run on which servers.
Worker Server GRs are defined in these terms:
- Position
- Has no effect in a worker server GR (Position is used by Authentication GRs only)
- Server ID
- Indicates which server this rule applies to.
- Delay Seconds
- Time in seconds that a server will wait before picking up available work
- Exclude
- Prevents this server from executing work
- Ignore
- Ignore this rule (used to 'comment out' rules)
Worker Server Algorithm
Each Netcool Configuration Manager worker server builds a cache of what worker servers can execute against each realm. This is calculated every 30 seconds by default, but can be changed in rseries.properties (Core/workerServerUpdatePeriod=30). It is not recommended setting the workerServerUpdatePeriod to less than 30 seconds. If a new worker server is added to a realm it may be up to 30 seconds before it becomes active.
When the Netcool Configuration
Manager worker
server grabs a task it firstly checks if it exists in the worker server
cache for the realm that the device belongs to. If it does, it then
checks which of these worker server GR's VTMOS is a match for the
device. The process does not attempt to identify the best VTMOS match;
every GR is either a match or is not, and all that are a match are
equally valid. The process then inspects the worker server rules defined
in the matching GRs that relate to the executing worker server based
on the Server ID.
- If the executing worker server has been 'excluded', then the task will not be executed by this worker.
- If the executing worker server has been included, then once the appropriate delay (which could also be zero) has elapsed, the work is executed.
Note: If there are more than one worker server in the deployment
and more than one worker is applicable to pick up the work then its
the first worker to pick up the work that executes it even if the
worker is in a parent realm in the hierarchy. There is no preference
given to worker server in the same realm as the device.