Uniqueness of aggregate id

[gshiman]

Can somebody please explain whether aggregate id is unique per TSM server instance or it's unique per TSM server?

TSM server may have multiple TSM server instances and I would like to know whether it is possible for two aggregates on two different instances of the same TSM server to have the same aggregate id.

Each TSM server instances has its own catalog database. On that basis alone aggregate id should be unique per instance.

Depending on the answer to the above question either (server hostname, TSM server instance name, aggrid) or (server hostname, aggrid) will be unique.

Thanks in advance,
George

 

[moon-buddy]

Everything is unique within a TSM instance. For example, you cannot just move DB and LOG volumes on the OS layer to another TSM instance and run it. You need to restore the DB so TSM will 'reseed' all of its uniqueness on this 'new' setup.

 

[rgg]

I'm sort of curious as to why you're asking this question, but as far as I'm concerned, two instances on the same machine aren't necessarily aware of each other, so everything should be unique per TSM instance, not per machine.

 

[Mita201]

I am also curious, why is this important at all? In what scenario?

 

[gshiman]

Everything is unique within a TSM instance. For example, you cannot just move DB and LOG volumes on the OS layer to another TSM instance and run it. You need to restore the DB so TSM will 'reseed' all of its uniqueness on this 'new' setup.

Yes, that is what I was hoping for. That there is reseeding to prevent collisions of restored aggrid's from another instance and pre-existing aggrid's.