Yet another operational enhancement that came to light in NSX 6.2.3 is a quick way to validate ESXi host health with a focus on NSX requirements but also overall host health. It's a simple command available from the NSX Manager Central CLI.
dc1-nsxmgr01> show host <host-id> health
This command will go out and collect data on the hosts and respond back with different status. As always, a heavily over-subscribed home lab doesn't disappoint in showing interesting messages. Here are some I captured:
dc1-nsxmgr01> show host host-52 health
UNHEALTHY, Host host-52 free memory percentage less than 15%: "3088"/"24575".
dc1-nsxmgr01> show host host-43 health
UNHEALTHY, Host host-43 memory is over commited: "0.05".
dc1-nsxmgr01> show host host-119 health
dc1-nsxmgr01> show host host-47 health-status
UNHEALTHY, Standard Switch "vmservice-vswitch" has no uplinks.
CRITICAL, There is no vxlan vds configured on host host-47.
One of out four isn't that bad, is it?
If you are wondering where I got the <host-id> here's the sequence I used.
dc1-nsxmgr01> show cluster all
No. Cluster Name Cluster Id Datacenter Name Firewall Status
1 DC1-MgmtEdge domain-c31 DC1 Enabled
2 DC1-Compute domain-c33 DC1 Enabled
This gives us the cluster-id which we can then use to expand the domain and in turn, see the host-id.
dc1-nsxmgr01> show cluster domain-c31
No. Host Name Host Id Installation Status
1 dc1-edge02.fuller.net host-43 Enabled
2 dc1-edge01.fuller.net host-119 Enabled
3 dc1-edge03.fuller.net host-52 Enabled
I love finding these improvements we're making!