When a single datastore is present on multiple ESXi hosts within a vSphere cluster, those hosts should reflect consistent storage configuration. This article describes some common areas of configuration inconsistency.
Block Storage (VMFS)
- In most cases, all hosts in a cluster should have access to the same LUNs.
- All hosts accessing a LUN should use the same Path Selection Policy (PSP) and Storage Array Type Plugin (SATP) configuration.
- When using iSCSI, all hosts accessing the same targets should use the same advanced configuration options.
- When using iSCSI, all hosts accessing the same targets should use the same iSCSI port binding configuration.
- All hosts accessing the same datastore should use the same VAAI configuration
- Hardware accelerated locking, hardware accelerated init, and hardware accelerated copy configuration should be consistent.
NFS Storage
- All hosts accessing the same NFS export should use the same NFS version.
- All hosts accessing the same NFS export should use a common acceleration configuration.
- The VAAI plug-in should be consistent on all hosts accessing the export.
- If NFS queue depth has been adjusted, it should be adjusted on all hosts accessing the export.
- If any advanced configuration options for NFS have been adjusted, they should be adjusted for all hosts accessing the export.
For assistance and guidance, see the "Additional Information" section for links to other knowledge base articles and product documentation.