LEARN. PRACTICE. GROW.Concept and mechanism
Storage has several layers. Growing an LV from 32 to 48 GiB requires another 16 GiB; if the VG has 12 GiB free, four are missing in this simple example. Free space inside the filesystem is not the same as free VG extents. After expanding the LV, confirm whether filesystem growth also occurred. Some options coordinate both; extension without resizing can leave df showing the old size. Identify type, device, and mount before choosing a tool. Mounted XFS can grow using its supported tool; resize2fs belongs to another filesystem family. mkfs creates a filesystem and is not a growth operation preserving existing data.
Guided application
Persistence and availability also need verification. An old fstab UUID may prevent a required mount; nofail allows certain startup failures to be tolerated but does not prove required data are present. Prevent an application from inadvertently writing into the empty directory beneath a mountpoint. Currently active swap needs startup configuration when persistence is required. With autofs, a mount may expire while idle and return when the path is accessed; test access before declaring failure. In NFS, root_squash limits remote root, so intended identity and service permissions still matter. For performance, correlate latency, queues, and workload; iowait alone does not size a disk purchase.
lvs shows 100 GiB and df 80 GiB on XFS whose extension omitted resizing: validate growth at the filesystem layer.
Common pitfalls
Confusing total GiB with increment; filesystem with LV; nofail with available data; remote root with unlimited access.
Related topics: Commands and operational evidence · Identities, ACLs, and resource limits
Confirm identity and state at every layer before and after any storage mutation.
Reference: lvextend(8) · LFCS current five-domain outline; exact edition date unconfirmed