SAN, NAS and Object Service Configuration
Domain 2 - SAN, NAS and Object Service Configuration
This domain tests day-2 provisioning across block, file and object on OceanStor: mapping views, initiator and host-group hygiene, multipathing and ALUA, NFS and SMB export design, and S3 bucket policy.
Block services
- A LUN becomes visible only through the full chain: LUN into LUN group, mapping view, port group. The host must exist in a host group whose initiators (WWN or iSCSI IQN) are registered.
- ALUA: array ports report preferred and non-preferred paths. UltraPath or native MPIO must be installed so the host uses preferred paths and fails over without unnecessary ownership trespass.
- iSCSI: keep service and management networks separate, enable CHAP where the baseline requires it, and make MTU consistent end to end - jumbo frames everywhere or 1500 everywhere.
File and object
- NFS exports need host or subnet access control and an explicit root-squash decision; SMB shares need correct AD integration plus the signing and encryption policy your security baseline mandates.
- Quotas, HyperSnap schedules and NDMP backups are configured per file system, and snapshot space is reserved from the pool.
- Object access is controlled by bucket policy and access-key pairs; WORM retention is set at bucket or object level and cannot be shortened once locked.
Worked example
A Linux host with two HBAs sees no LUN. Check in order: initiator WWPNs registered and online; host object created with both WWPNs; host belongs to the host group; the LUN is in the LUN group referenced by the mapping view; the port group contains the target ports zoned to the host; the mapping view is mapped and the LUN is online. If the LUN appears twice, the host is seeing both ALUA paths without multipath - install UltraPath rather than masking one path.
Traps
Mapping one LUN to two independent hosts without a cluster file system; forgetting to re-zone when an array port is added; carrying service traffic on management ports; letting snapshots exhaust the pool because no reservation was made; mismatched iSCSI MTU causing intermittent login failures.