I will need to power off a "Metro Standby Site", which has VMs running on it.
I would like to manually migrate those VMs to the primary (active) site and Shut down / Power Off the nodes in the Standby site approx. 24-36 hours.
All VMs need to be running on the active side and no VMs are to be restarted (RTO + RPO =0).
The steps I propose to take to achieve the goal are as follows:
I have included as much environment info as possible.
Any insight or corrections to this process are welcome.
Thanks in Advance
Phase I (Configure Standby Site)
- Disable/suspend the (x1) asynch PD (from Primary)
- Update VMware DRS Rules (Manual) (from Primary)
- Disable HA
- Manually perform vMotion of all VMs to active site.
- Configured VM affinities to the hosts of the primary site and set the DRS rules to Fully Automated.
- Create new VMGroup and Host Group if needed
- I do not want to disable DRS and lose my DRS Affinity Rules
- Disable Metro per PD (x3) (from Primary)
Phase II (Power Off)
- Shut down CVM(s) on (Standby site only)
- Put VMware Host(s) in "Standby Mode" (Standby site only)
- Power Off Nutanix Node(s) (IPMI) (Standby site only)
Phase III (Power Up) (Standby site only)
1. Power on Nutanix Node(s) (IPMI)
3. Disable "Standby Mode"
4. Ensure CVM(s) Powered Up
a. Status Checks
Phase IV (Re-enable PD)
- Migrate VMs back to Standby Site (previous VMs via affinity rules?)
- From the Primary Prism console establish the metro availability configuration by clicking Re-enable on each PD.
- Reestablish the DRS affinity Rules to VMs to original Groups etc.
- Reenable the async PD
Power Off DEB Hosts for Extended period not to exceed 36 hours
DEA (Primary) (5 Node)
DEB (Standby) (5 Node)
DEB to be Powered Off
VMs are running on the Standby Hosts
Ntx-XX-XXX / Ntx-XX-XXX
AOS 5.10.5 (LTS)
Blocks 1-2 NX-8035-G4 x2
Block 3 NX-8035-G5 x1
Blocks 1-2 (x4 Hosts)
Block 3 (x1 Host)
DRS: Enabled (Fully Automated) (Default settings)
x1 Remote site
Hci-XXX-adv-01 : hci-XXX-adv-01-mirror
Metro Availability PDs: both sides have identical resources, Disk, CPU, Memory
x3 active Containers
x3 standby Containers
Async: (active on primary site)
Failover capacity is defined by reserving a percentage of the cluster resources.
Reserved failover CPU capacity:
Reserved failover Memory capacity:
Name HAEnabled HAFailover DrsEnabled DrsAutomationLevel
---- --------- ---------- ---------- ------------------
Cluster Name True 1 True FullyAutomated