Symptoms:
An ESXi host is disconnected periodically in vCenter.
Logs:
/var/log/vmware/vpxd/vpxd.log
YYYY-MM-DDThh:mm:ss:uuu warning vpxd[63037] [Originator@6876 sub=MoHost opID=HostSync-host-37931-7f75b4b7] host [vim.HostSystem:host-37931,] connection state changed to NO_RESPONSE
YYYY-MM-DDThh:mm:ss:uuu warning vpxd[64705] [Originator@6876 sub=MoHost opID=HostSync-host-37931-73b9cd07] host [vim.HostSystem:host-37931,] connection state changed to NO_RESPONSE
<snip>
YYYY-MM-DDThh:mm:ss:uuu warning vpxd[64784] [Originator@6876 sub=IO.Connection opID=HB-host-37931@1686-52d66fc6-SWI-4b70ddca] Failed to connect; <io_obj p:0x00007f4dd49726c8, h:152, <TCP '10.221.113.17 : 44276'>, <TCP '10.163.243.46 : 443'>>, e: 110(Connection timed out), duration: 31436msec
Two IP addresses of the ESXi host is returned to vCenter but one of them is non-working IP address:
Working IP Address : xx.xxx.xxx.xx
Non-working IP Address : xx.xxx.xxx.xx
The issue arises when vCenter Server receives multiple A and PTR records for a specific ESXi host. While attempting to connect to one of these records, vCenter Server encounters an unreachable IP address.
The problem stems from non-working A and PTR records in the customer's DNS configuration. Resolving these entries will address the issue.
Impact/Risks:
An ESXi host can be disconnected and connected repeatedly.