- Add license and assign the new license key to both hosts. As per answer here...had no issues doing this live.
- You’ll need to create a cluster in vCenter to enable HA and DRS features. Reading min of 2 hosts is fine but 3 is best. Make sure all VM’s can reside on 1 hosts and that hosts can as enough memory and CPU to accomodate the extra Vm's.
- Once cluster object is created. Drag and drop the hosts into the Cluster. The quick steps wizard error’s out on 7.0.2 when adding the hosts.
- Once both hosts in cluster you may get a message about
- message can be safely ignored if you choose not to.
- The number of shared datastores between all hosts is less than 2 so add another shared storage or disable the message :
- From vCenter. Right click the Cluster / from vSphere HA tab Click EDIT settings > From Advanced options: add the following option.
- das.ignoreInsufficientHbDatastore. Value True - Under Cluster Features - Turn off HA and turn on again.
- "This host currently has no management network redundancy" :
As per KB.vmware.com/s/article/1004700: Recommends to a add a second vmnic to the service console or management network vSwitch configured iwth NIC teaming and that you keep both as active adapters or place one in standby" - "The number of vSphere HA heartbeat datastores for this host is 1 which is less than required: 2":
As per KB.vmware.com/s/article/2004739
Prior to configuring new networking settings, DRS and HA make a plan on how you’ll use the new features before enabling. See below.
Consideration in HA
When high availability is required: Think how much redundancy you have, want, can afford or need.
Do you have dual path to the storage, 2 fabric switches, 2 Nic’s for storage ?
Do you have redundant storage (offsite or onsite) ?
Do you have dual path to the network switches for VMO and VMK traffic ?
Do you have redundancy in your hardware (dual teamed Nic's assigned to the VMK traffic etc...)
Do you have a plan or requirement to have VM's prioritized to reboot or move to another host if one fails.
How do you want to deal HA with VM's.
If ESXi host is still up and has lost access to the network or datastore
Host isolation: Host is still up.
Permanent Device Loss (PDL): Some kind of physical storage failed.
All path downs (APD): -All path downs no storage and comms from host,
VM MONITORING : If only the VM stops responding to heart beats using VMware tools
Consideration for DRS configuration
Do you want to assign your vCenter vm to override and not automatically migrate to another host so that you always know where it is. Say if you want to access vCenter when doing a direct login to the host via web browser.
Do you have a need a specific order in how VM's should reboot ie: DB before the app, LDAP auth before the APP etc...
VM/Host rules:
You create VM and HOST Groups for say PROD and DEV and group those Vm's in the groups if you want to specify what hosts have what Vm's. and then you create rules to specify which groups or vm's run on which hosts ie: which VM's should run together, which VM's are redundant ie: clustered.
Affinity rules:
Affinity and AntiAffinity rules: Make sure vm's are always on same or separate hosts. App and web server on same host while 2 domain controllers are on separate hosts.
All Dev Vm's on same old host or specific assigned host for Developers. Less or more hardware memory etc...
Vm Overrides: Some licenses required you to keep a VM on same hosts, dongle to a host/vm combo, You want to pin a VM to a specific host to prevent unnecessary movement.
VM Overrides: How to treat certain VM's differently on a per vm basis or say Disable HA for certain VM's that you don't want to get move should there be a host issue. ie: Your own software enabled clustered app
