When performing a cluster remediation or host upgrade to ESXi 8.0 Update 3k (build 25595708), the host may enter maintenance mode but fail to complete the remediation process. Upon manual reboot or power cycle, the system stalls, and the following error is observed on the physical console or virtual console during the ESXi load process:
Bank5: Invalid Boot State Bank5: Invalid Boot State
The host remains unresponsive in vCenter and may require multiple cold boots before successfully loading the hypervisor and exiting maintenance mode.
The Invalid Boot State error is generated by the ESXi safeboot process when the newly flashed alternate bootbank partition fails pre-boot validation. During an upgrade, the new image is written to the alternate bank; if the UEFI/Secure Boot environment fails to validate the bootloader signatures (DB/DBx lists) or encounters firmware-level delays during the initial handoff, the system flags the partition as BANK_STATE_INVALID.
In environments utilizing HPE hardware, this is frequently tied to outdated UEFI firmware or Service Pack for ProLiant (SPP) versions that are not fully compatible with the specific signatures introduced in ESXi 8.0 U3k.
If the host is currently stuck, performing multiple cold boots may eventually allow the validation to pass. To resolve the underlying cause and prevent recurrence across the cluster, perform the following steps:
If an immediate firmware update is not possible, a manual cold boot (powering the server off and then on via iLO or physical button) may be required up to three times to force the UEFI environment to successfully initialize the bootbank.