03-15-2011 09:09 AM
I have one mailbox and one journal Vault server in each of our two datacentres, both are on physical hardware rather than VMware. We run an active passive configuration; so currently the mbx and journal vault servers are live in datcentre 1. We'll use USL to failover the services for both Vault servers to the other datacentre; this has been tested and is working as we expect it to. However this failover assumes the source servers, so the servers currently in datacentre 1 are online. what is the recovery procedure if we lose the two servers in datacentre 1 ?
All our indexes and partitions are hosted on netapps that are replicated and the SQL dbs are mirrored so we can recover all the user data.
Thanks
03-15-2011 11:53 AM
Hi Phin
Do you use DNS-aliasses for the EV-servers?
Seeing you use USL you should.
Switch the aliasses, then do USL
example:
EV1 = server1
EV2 = server2
EV3 = server3
EV4 = server4
EV1 and 2 in DC1 which failes
Set alias EV1 to Server3
Set alias EV2 to Server4
make sure data is replicated etc.
USL on EV1 (then running on server3)
I think....
03-16-2011 02:22 AM
Hi,
Thansk for this; is this failover step still valid if the live nodes become unavailable? So in a DR scenario where you lost connectivity to a datacentre would the USL procedure of flipping the DNS names then running the USL process work?