Veeam Replication and multiple restore points affecting perf

[garymansell]

Hi,

I am looking to replicate a high performance Enterprise wide Java Web App and its associated backend MS SQL DB Server using Veeam Replication.

I would like the option of having a reasonable number of restore points on the replica site so I can be sure that at least one of them will come up OK, and also that I can chose to go back to a particular point in time if I need to.

It seems having lots of restore points (vmware snapshots) absolutely kills the performance of the Web Application and DB when I run it from the Replica site with multiple snapshots.

It seems to me that once I have decided which restore point snapshot to come up on I need to be able to delete all the others so that I am running at the replica site without any snapshots.

I understand that it is not possible to delete these restore points as this would break that Veeam Replication configuration, so the best solution would be to perform a Permanent Failover to the Replica site which will do this and delete all the restore point snapshots.

My question is, once I have done a Permanent Failover to the replica site how do I go about switching the whole lot back to the primary site (ie what would normally be accomplished by a "Fail Back" if I had not done a permanent Failover) - specifically the new master vm's at the Replica site will all be named "_replica" whereas the original vm's at the source site will not.

Does anyone have any advice they could offer on this?

Rgds

Gary

[DGrinev]

Hi Gary,

Once you have done a permanent failover your replica became original VM on remote site, so you could perform replication job once again followed by another permanent failover to primary site (to be sure you have latest state of the VM).
However, it’s highly recommended to use Replica mapping as you have your ex-original machine on primary site (it will reduce amount of data to be transferred back).


Regarding your question: “ie what would normally be accomplished by a "Fail Back"”
Please see detailed process explanation on Replica Failback

Thanks!

[garymansell]

Hi Dmitry,

Thanks for getting back to me - yes it seems to me that this (https://www.veeam.com/kb2018) is the process that I need to follow to the switch back manually to the original source site and VM's (by using replica mapping).

Do you agree?

Rgds

Gary

[DGrinev]

Agree.
I'm glad that you found the right process in KB, it will help you to solve the issue.

Thanks!

[skrause]

Gary,

What edition of SQL are you running on? If you are using Enterprise, have you looked at using SQL Always On to replicate to the failover location? Your performance issues are likely to be caused by running SQL on a series of snapshots. It also lowers the RPO on transactional data substantially as you don't have to roll back to wherever your last restore point was, it is where the last write was committed to the database. It also cuts down on the number of potential snapshot commit stuns.

[garymansell]

SQL Standard - as Enterprise is 4x the cost.