When using VIO normally the hdisks are servered by a VSCSI interfaces to the LPAR.
These vscsi interfaces change their state, if one VIO server within a dual vio environment gets bootet.
On normal behaviour the hdisks should be set in the state “failed” and will be rechecked by the interval set in the option hcheck_interval on the hdisk.
It might be possible that a vscsi adapter gets into a defined state (i.e. a VIO is longer not reachable). If this happens, the hdisks get into a “missing” state. This state is not being updated by the AIX system. It needs to be checked and changed after the VIO server serving the path to the hdisk is reachable again.
Be aware of this when updating VIO servers, as it might cause dataloss on the running LPARs.