Hi Due to some network issue from ESXi hosts cant ping the gateway due to this i think the hosts get disconnected after some time in cluster and gives the following error:
vsphere ha agent on this host could not reach isolation address gateway
Can you please confirm is the above error due to gateway issue....I have also tried to change the hostisolation address for a testing purpose i give another ESXi server IP in hostisolation. Disabled and re-enabled HA again but even now servers get disconnected from vCenter and goes into not responding state. PFA image for host isolation settings.
If the hosts cannot ping the default gateway, the hosts would get disconnected from Virtual Center!
Yes, the error message you see is due to an issue with the isolation address. However, the reason your host shows as disconnected could also be due to address resolution issues. Please check whether the host can resolve the vCenter Server's short name a well as its FQDN and vice versa.
Hi Thanks for your help.
I am able to resolve the FQDN of hosts from vCenter as we have edited the hostfile of vCenter. But we cant ping the fqdn of vcenter from ESXi hosts....
We have other environments where only vcenter can ping ESXi FQDN but HA works perfect.
Secondly due to any restriction if the gateway is not reachable the different host isolation address should work? but in our case its behaving same even after specifying the different host isolation.
I can't test this at the moment, but you may try to use das.isolationaddress0 instead of das.isolationaddress. After changing the setting make sure you disable/enable HA. Please confirm the isolation address you enter can is pingable from the ESXi host.
Well there seems to be two things going on here
1. Connectivity between the VC and ESXi hosts
2. Isolation address issues
Let's take this one beast at a time.
In order to have the hosts "connected" to VC there seriously are just two requirements
1. You should have port 443 open and that would enable the VC to push the HA/vpx etc agents to the ESX hosts
2. UDP traffic has to be allowed between the VC and ESX hosts else the hosts would connect and disconnect from the VC after a minute
The KB above should help with the hosts connection issue
Now, if the hosts are getting disconnected, the hosts are likely not able to heartbeat with VC and each other as well - hence this is likely a by-product of the other issue.
So to get to a resolution, get the hosts connected first - as that would also help in the HA issues you are seeing!