An ESXi host fails to enter maintenance mode, and the corresponding task appears stuck or hangs.
This issue is caused by the embedded vCLS VMs (introduced in ESXi 8.0U3 or later) failing to automatically evacuate from the host.
Note: This article applies to embedded vCLS VMs. For similar symptoms related to external vCLS VMs (pre-ESXi 8.0U3 versions), please refer to: ESXi hosts fail to enter maintenance mode in VCSA
[timestamp] warning vpxd[18668] [Originator@6876 sub=MoCluster opID=PodCrxMgr-domain-cXX-1882] Failed to disable vCLS Pod Crx; [vim.HostSystem:host-XX,<ESXi host name>], [vim.ClusterComputeResource:domain-c35,<Cluster Name>], N4Vpxd7Clients24UnauthenticatedExceptionE(Error:--> com.vmware.vapi.std.errors.unauthenticated--> No messages!--> )[timestamp] Er(11) settingsd[132451]: Auth_AuthenticateUserPAM:424: PAM failure - Authentication failure (7)[timestamp] Er(11) settingsd[132451]: error [ConfigStore:a5a1a7a700] UserPass authentication failed for user: vpxuser[timestamp] Er(11) settingsd[132451]: error [ConfigStore:a5a1a7a700] Method failed
ESXi 8.0U3 or later
This issue is caused by a password mismatch for the vpxuser account between vCenter Server and the ESXi host.
To resolve this issue, resolve the vpxuser password inconsistency between vCenter Server and the ESXi host.
The underlying cause for this password mismatch can vary.
One potential cause and its corresponding solution are outlined in the following article:
vpxuser password change fails repeatedly