cancel
Showing results for 
Search instead for 
Did you mean: 

EV9 - Move Archive and Vault Cache sync

Mark_Clardy-USA
Level 2
Partner

All,

We have just concluded a large PST migratio project in which we added a number of additional EV servers to help with the ingestion of PST files.  We now want to move the archives back to the original EV servers.  We are using EV 9.0.2 and determined that there is no vault store migration available.  If we use 'Move Archive' then we have to go through numerous steps listed in the forum posts to get the vault cache to sync again.

Would these 2 options below be the best that we can do as of now?

1.  Move the vault store partition to the original server by modifying the storage service entry, partition root path entries and index volume locations?

2.  Use USL to decommission the extra servers and change the DNS and perform a recovery on the original servers to now host vault stores, storage and index services.

 

Thanks

1 ACCEPTED SOLUTION

Accepted Solutions

Jeff_Shotton
Level 6
Partner Accredited Certified

Running in building blocks failover to aggregate services together would be easier in terms of the fact that the database changes are made for you, but you are left with extra EV server instances in the vault admin console. i.e It's a bit messy.

Also, you will be left with one set of services as 'hosted'....although thats kind of cosmetic.

On the other hand, while it is a more 'clean' solution, you better be sure of what you are doing if you want to edit the databases to merge the services together. While there isn't necessarily much to change, getting it wrong means service outage. Test in a lab first - you've missed a couple of things off your list of changes to make.

One thing to think about here though is EV10. If you are going to migrate over to 64 bit indexing and you want to convert the indexes, you might want to leave the extra servers in place to process that workload - which means either doing nothing for a few months, or running the building blocks config in failover mode till you add extra servers back in again.

Regards,

Jeff

 

View solution in original post

1 REPLY 1

Jeff_Shotton
Level 6
Partner Accredited Certified

Running in building blocks failover to aggregate services together would be easier in terms of the fact that the database changes are made for you, but you are left with extra EV server instances in the vault admin console. i.e It's a bit messy.

Also, you will be left with one set of services as 'hosted'....although thats kind of cosmetic.

On the other hand, while it is a more 'clean' solution, you better be sure of what you are doing if you want to edit the databases to merge the services together. While there isn't necessarily much to change, getting it wrong means service outage. Test in a lab first - you've missed a couple of things off your list of changes to make.

One thing to think about here though is EV10. If you are going to migrate over to 64 bit indexing and you want to convert the indexes, you might want to leave the extra servers in place to process that workload - which means either doing nothing for a few months, or running the building blocks config in failover mode till you add extra servers back in again.

Regards,

Jeff