I do not have VMware enterprise plus license only foundation at DR site, so I can’t use a VMware Distributed vSwitch. While I could storage vmotion the vm from the lab I wouldn’t have the notification in Veeam of which ones are still waiting like I do with instant recovery. Some VMs may take a day or two to move, as we will be busy with the DR.
With a change in Veeam to allow selecting of virtual switch, I could place VMs in application group and assign to the lab attached to the esx host I want the VMs to run, Fire up each set and I am done.
With a way to chain the job that turns on of the labs I would have a one click failover to DR site.
The only part Veeam is missing is a way to allow me to pick the virtual switch the application group will run on, the lab picks the esx host.
Why I am doing this
I have been tasked with doing a whole site DR test to another production site. I have the instant recover from remote SAN snap shot functioning. Will be using the lab appliance just as a router between vlans, static mapping and internet proxy appliance. I can now do my tests and everything works as planned, I just have not added the 30 users testing all at once. The user load and recovery will take place during production hours and I need to meet my RTO/ RPO. I will also not know ahead of time when the test is.
I now have one lab per ESX host all talking with about 8 VMs per lab. I could add a couple more labs for internet proxy and host mapping if needed for throughput. I will also be connecting up to 30 user PCs to my test network, which allows a complete test and DR capacity verifying.
I cannot use the labs out of the box for the whole test because if it were a real DR I would be using instant recovery from current SAN snapshot, then quick migration or storage motion. All servers will have a 15 minutes RTO and 1hour– 24hour RPO. All servers but Exchange can meet a RTO of 5 minutes when I am sitting at the keyboard and ready with instant recovery.
If this was a real whole site DR I would follow the same procedures plus, I just need to power off the labs (not the VMs), move my IPs to DR site, assign IPs to valns, allow routing between vlans. Most of these changes are one configuration line.