Server Maintenance & Veeam

[penfold]

I've finally sorted my head round some of the features Veeam has, and it looks like we need to develop how we are using it to get the full benefits of it. One thing that I'm not sure about the best way to go is how to perform Server maintenance. I have read that it is possible to failover to a replica to allow users continued access but perform maintenance on the server and allow you to test things. I am now looking at upgrading some software on one of the servers we have and thought it would be a good time to test the procedure.

 

Is it possible to allow users conitinued access to the live server and perform a software upgrade & testing which can then be rolled into the live server with limited disruption to the end user? If so what is the best way of doing this?

[Ephelyon]

You can replicate it away and then make changes with zero disruption, but in terms of integrating those back in, you might need to make use of its Failover and Failback features.

[penfold]

If I use Failover in order to integrate changes, don't these need to be done on the current live server(replica)? So any updates are intergrated back into the oringianl VM during the Failback procedure.

 

My understanding is that you can create a replica before you install any software, and if there are any problems it is possible to failover to the most recent (working replica). But this means you are performing updates on the live server. If you install software on the replicated server is this saved into the replica so once testing was complete you can failover to the updated replica with the new software installed without problems?

 

Not sure if I have made that clear or am I just confusing myself?

 

Basically I just want to be able to test software being installed on a server without doing it on the live server and to be able to roll it into the live server if everything works.

[Ephelyon]

Not without some disruption on the failback I don't think. So I wouldn't use it to actually complete the work, but rather just to run through the procedure to make sure everything is okay. It does mean doing it twice, but better safe than sorry. If everything goes smoothly on the replica but you click the wrong thing on the live server and FUBAR it (disruption inevitable), you can always just copy the replica files back over to the host and fix it that way. Edited March 14, 2013 by Ephelyon

[penfold]

Basically confirms what I was thinking....test on the replica and if all goes well repeat on the live server? So long as I know this thats fine, I just wanted to be sure before I start to play. I was just hoping there was a nice easy way of incorporating changes if everything worked

 

Cheers