EV building blocks

[GTK]

hi just need some info or clarification

i have 3 EV servers in test lab. All 3 servers host EV services but only 1 has a journal task, the other 2 have no tasks ( i have inherited this, not my set up). Anyway.. when running USL, the services from EVSERVER 2 fail over to EVSERVER 3 as expected and the services on EVSERVER 2 say 'hosted' as I expect. However the journal task on EVSERVER 2 does not move over to EVSERVER 3. It should have moved shouldn't it?

My experience is mainly failing over to EV DR servers without any EV services (active /passive) so the tasks are created as required on the DR server.

Why is the journal task not being created on the target EV server after successful USL failover?

Thanks

[GTK]

just to add, all indexes for each server are on same letter drive

EG -

EVSERVER1  = T:\EVJT001\Index_Mt1

EVSERVER2 = T:\EVJT002\Index_Mt1

EVSERVER3 = T:\EVJT003\Index_Mt1

[AndrewB]

Update Service Locations (USL) is used for changing server name or DNS alias. what are you trying to do?

[GTK]

i am trying to simulate a server failing and one of the other servers taking over the resources of the failed server

[Merv]

The USL is used to update service locations and task based on a changed in your DNS alias. I.e. Evserver1 -> SYDEVMAIL01 Evserver2 -> MELEVMAIL01 Evserver3 -> SYDEVMAIL02 Did you ensure you did ipconfig /flushdns after making the dns changes and test with a ping to see if it is resolving to the updated IP? There is not much which can go wrong. Storage (for vault store and index) should be reassigned to the server you are failing over to before running the uodate service location . If services or task do not exist on the target they will get recreated. Any errors ?

[GTK]

yes everything works as expected but i was under the impresison that the existing journal task would move over to the new target server. However from what I have read today... as the target server already has EV services installed and running, then the existing journal task on the source server will remain there visibly and an amendment will be made in SQL to point the existing journal task at the new target server

this is correct?

[Andrew_G_ · Accepted solution]

A better way to think of this is to separate the Enterprise Vault Server (as listed in the admin console) from the machine (physical or virtual) it is running on. When you run USL, you effectively move (based on the DNS records) an Enterprise Vault Server from one machine to another. It is also possible to have more than one on a single machine, in which case one (or more) will become "hosted".

The main SQL update that USL makes is simply to associate the Enterprise Vault Server with a different machine. The services enumerate the Enterprise Vault Servers that are currently associated with the machine, and pick up all the tasks for all the Servers currently residing on that machine.

In the admin console, the tasks will remain with their original Enterprise Vault Server. In that way, you can then move them back once the machine has been repaired. You can see which machine each Server is currently associated with in the console tree: it's given in brackets after the Server name.

Does that help clarify things?

[GTK]

yes Andrew thanks, the confusion on my part was that I thought the tasks moved from source to target which i now know they dont

thanks all!