Data Center Redundancy

[Kris Dunn]

We are currently running Service Desk 14.1 in Advanced Availability mode with Process Automation 4.2 in a Hyper V clustered environment with a single Oracle database.

 

We are in the planning stages of migration to a new Hyper V cluster and what to include data center redundancy over 2 data centers.

 

As I understand it, Service Desk is cluster tolerant, but not cluster aware. I assume that means that I could spread our servers over multiple data centers without issue, putting the background server, an application server, and a PAM server at one data center and the standby server, another application server, and another PAM server at a second data center. Any reason why this would be a bad idea?

 

Thanks

--Kris

[Jon_Israel]

Hi Kris, this should be OK as long there is not a ton of network latency.  As you mentioned, Service Desk IS cluster tolerant, meaning it can be installed in a cluster environment, but it is not aware of the cluster and will not failover gracefully within a cluster environment.    The best thing to do in this situation is really to test it all out before doing this with production because every environment is different.

Thanks,

Jon I.

[buian01]

The one issue I see in your plan is to have a PAM node in each data center. PAM/JBOSS use multicast messaging for managing the members of the cluster. If the members of the PAM cluster are going to be on separate subnets you will need to bridge this traffic across the subnets to allow the nodes in the PAM cluster to be able to recognize which nodes are currently active. Otherwise you need to keep the node in one of the sites down but this raises issues around repository replication between the nodes that you would need to take into consideration.

 

Andy

[Kris Dunn]

Thank you this is exactly what I needed. Our IT guys were concerned about multicast and I was pretty sure it was not an issue with Service Desk. Thanks for pointing out this issue with PAM.