Server Manager DR Options ?

[ice_cube210]

With Server Manager being an integral part of the E1 System Architecture I was curious to see how others are handling the high availability / DR aspect of server manager.

With tools 9.1 I know you can deploy SM console to a regular WLS or WAS server , not sure how you would "cluster" the SM console. Would you use a virtual name for the SM console ?

If the SM console was running on a VM server , restoring a snapshot to a DR site should be fairly easy if a good backup strategy is in place .

Most customers have the SM console running on the deployment server and so if you had a DR strategy in place for your deployment server you would be covered , especially important if your media object attachments reside on the deployment server.

Hope to hear how others are handling this.

Thanks in advance!!

 

[Tom_Davidson]

Ice,

We don't consider SM to be 'Mission Critical' You can still start and stop services and JAS manually (from the appropriate machines). The only thing we really lose is the monitoring and upgrade tasks. And if we are in a DR situation, we don't think we will be doing much of either.

We run SM on our deployment server, which would be in the second wave of recovery for us.

Tom

 

[ice_cube210]

Hi Tom,

Thank you for your response. Yes you are right , the SM functions should not be mission critical if you can start and stop JAS services using the native WLS or WAS commands / consoles. Just things like clearing cache etc become a pain.

 

[Tom_Davidson]

Ice,

That is our thought. While a pain, it can be done, inconvenient but doable.

I'm watching to see if anyone finds a reason it is mission critical. If one comes up I will have to reconsider.

Tom

 

[peterbruce]

Tom,

I can't say we view the SM as mission critical either, but I believe the way we do our DR it will be there. I don't know the details, but we use our test/dev machines (completely separate JDE system) and somehow bring up the production system on them. I did not test the SM last time we did our DR testing, but considering the SM lives on the DS and the agents are on the other test/dev machines, I don't see why it would not work.