Linux Storage

Linux presents persistent storage through layers. A physical drive, virtual disk, or remote logical unit appears as a block device. Optional partitioning, software RAID, encryption, and LVM create new block devices above it. A file system organizes one of those devices, and a mount attaches that file system to the single Linux directory tree.

A path such as slash var slash lib identifies files in a mounted file system. The mount table connects that path to a file-system device. That device may be an LVM logical volume, which maps extents from a volume group built from physical volumes. Those physical volumes may be partitions or whole devices reached through SATA, NVMe, iSCSI, Fibre Channel, or another transport. Troubleshooting works from the path downward and from the device upward until both views meet.

Read the stack before changing it

The same storage can have several names at different layers: a kernel device name, a partition, a device-mapper name, an LVM logical volume, a file-system UUID, and a mount point. These names are not interchangeable. Short kernel names can change when discovery order changes, while UUIDs and LVM names are intended for durable configuration.