I am currently looking into the DR solutions available for an EV 10 Cluster (Win 2K8 R2). I would like to have a single server in DR.
I've looked into the following options:
- Have a cluster node in DR (EV does not support clusters with more than one IP address resource so I guess that is a no?)
- Have a server in DR with a different name and add it to the EV site. In the event of a disaster, present the storage to the DR server and failover using Building Blocks. (According to the documentation, this is not supported)
- Have a server in DR. In a DR situation, present the storage to the DR server and do a data-only restore (We have an RTO of 1 hour and this might be fairly time consuming. I take that is effectively a hardware migration?)
Are there any other ways to accomplish this which would be relatively quick to perform and be supported?
Thanks in advance!
Please have a look into option available for DR & high availablity.
High Availability Options for Enterprise Vault
Disaster Recovery for Enterprise Vault
Having a server in DR with a different alias & host name and add it to the EV site looks good. In the event of a disaster, present the storage (indexing & partition) to the DR server and perform failover by changing production alias to DR Server.
I don't think this is unsupported in case of windows cluster.
Thanks for your reply. I'm a little confused about the steps required for failing over to a DR server with a different name. I have been told before that it could be done by simply repointing the the production DNS but would that be all that is needed? No changes in SQL? I wonder since Building blocks requires running Update Service Lications on top of that.
You just need to change DNS alias pointing from production to DR EV server and then run USL (Update service location).
USL does rest of the work such as updating entries in SQL database.
There are example giving in tech note below for building block configuration.
How to setup building blocks and configure Enterprise Vault for fail-over http://www.symantec.com/docs/TECH38701
Hi EV Counselor,
I'm also thinking to do the DR site EV server as well. So far at the moment I can use the EV Virtual machine that is replicated to the DR storage array, including the physical RDM LUNs presented to it using EMC Replication manager utilizing EMC SAN copy from one VM datastore to another.
Many thanks for your clarifications EV-Counselor.
My ultimate aim was to get a confirmation as to whether the Clustering + Building blocks combinations actually works since http://www.symantec.com/docs/HOWTO37984 indicates it is not a supported setup.
Whilst looking through the forum threads/discussing the solution with consultants over the past days I have been told of various sucessful implementations. I even stumbled upon a post from a Symantec employee confirming that it not only works but is also supported:
"There has been recent discussions on an internal forum about the exact subject of USL with cluster which prompted discussions between Product Management, Engineering and certification teams and the result of this was that this configuration is supported.
Also a technical architect was in the discussion and attested to its functionality as a few clients do have this configuration.
The thing is that it has not been officially tested and certified which is why you do not see this information in the compatilbility list.
Since USL is DNS alias based if you failover the alias for one EV server (clustered or otherwise) to another EV server and run USL the process will work. "
It would be good if Symantec could officially test and certify the solution at some point to end the confusion surrounding it
Thanks once again for your help!
I would be agreed with Paul as he said in another fourm thread that this configuration works, even I don't see any reason of not working such setup. But still you can reach out support team in order to get confirmation via tech note.
You guys may want to have a look on QUADROtech's EVnearSync.
It's an HA & DR Solution that integrates deeply with EnterpriseVault.
It replicates (like a Centera) archived content between Servers/Datacenters and has a RPO of Zero! and an RTO of some seconds.
In the upcomming Version the toolset allows you also to switch not only storage - it also supports an EV Server failover/failback either manually (one click) or automatically.
Let me know if you need more info on that topic.