ESXi hosts may intermittently become unresponsive, with host management operations impacted. The hostd service may crash immediately or within a few minutes after a restart attempt, and in some scenarios, esxcli commands may also become unresponsive.
The issue can be identified through log entries similar to the following:
/var/log/vmkernel.log
yyyy-mm-dd In(182) vmkernel: cpu72:2110465)uw.2110464 (44739) requires 1024 KB, asked 1024 KB from likewise (792) which has 93112 KB occupied and 72 KB available.
yyyy-mm-dd In(182) vmkernel: cpu72:2110465)Admission failure in path: host/vim/vmvisor/likewise:lwsmd.2110464:uw.2110464
/var/log/hostd-probe.log
hostd detected to be non-responsive
VMware vSphere 8.0
The observed behavior correlates with a known issue in ESXi 8.0 U3g involving memory leaks in Likewise services during Active Directory operations or when smart card authentication is enabled on ESXi hosts.
Memory leaks in Likewise can cause the associated processes to exhaust available system memory. When this condition occurs, ESXi host management services may become unstable, resulting in:
Intermittent disconnections from Active Directory or vCenter
hostd service crashes
Unresponsive esxcli operation
Upgrade the ESXi hosts to ESXi 8.0 U3h to remediate the known issue and stabilize the environment.