Diagnose storage by identifying the failing layer: disk, array, filesystem, NAS service, storage path or application workload.
- Correlate SMART and controller events with symptoms; a single health indicator is not a complete diagnosis.
- For SMB, NFS and iSCSI, compare addressing, service reachability, permissions and server-side capacity.
- For performance or capacity incidents, measure latency, I/O load, growth and the underlying pool or thin-provisioned storage.
Command examples
Adapt placeholders, permissions and target systems before using an example. These examples are documentation; no command runs from this page.
lvs -a -o +data_percent,metadata_percentGet-Counter '\PhysicalDisk(*)\Avg. Disk sec/Read'Check storage connectivity
Use the configured service and target address. Common service ports include SMB TCP 445, iSCSI TCP 3260 and NFS TCP 2049; a successful port check does not validate the complete storage session.
Collect Windows storage evidence
Record volumes, physical disks, relevant storage events, iSCSI sessions and MPIO path state. Windows performance-counter names depend on the operating-system language.
Collect Linux storage evidence
Record filesystems, block devices, mount state and the relevant LVM, ZFS, RAID or multipath status. Interpret the example against the storage stack actually in use.
RAID and backup serve different roles
RAID can maintain availability through some disk failures. It does not provide an independent recovery history for deletion, corruption or ransomware.
Identify a missing path
Establish which host, target, adapter or network path failed before restarting a host or an array.
Keep capacity for operations
Snapshots, merges and rebuilds may need temporary capacity beyond normal data growth.
Correlate SMART with logs
Compare disk indicators with controller, NAS and operating-system events before concluding that a disk is healthy or defective.
Review the storage network
Check the intended storage network, segmentation and contention before treating an iSCSI problem as a disk failure.
Validate recovery
A successful backup job is one signal. A restore test establishes whether the required data and application state can be recovered.