An administrator manages a server with several iSCSI LUNs attached. The server reboots after a kernel update, and one of the LUNs, which previously appeared as /dev/sdc, is now enumerated as /dev/sdd, breaking an /etc/fstab entry that references /dev/sdc. The administrator wants to make the mount configuration resilient to device-name changes and to avoid boot delays if the LUN is temporarily missing. Which two actions should the administrator take? (Choose two.)
The nofail option tells systemd and mount to treat a missing device as a non-fatal condition during boot. Without it, a temporarily absent iSCSI LUN can drop the system into emergency mode or cause long timeouts. With nofail, the boot continues and the mount is simply skipped, which is the desired behavior for storage that may not always be present at boot time. It is commonly paired with persistent identifiers.
Why this answer
Persistent identifiers such as UUIDs or by-path symlinks decouple the mount configuration from the kernel's enumeration order, and nofail prevents a temporarily missing iSCSI LUN from delaying or blocking boot. Together they make the fstab entry resilient to device renumbering. Changing to a device-mapper path, adding _netdev, or forcing a udev rename do not address enumeration stability and can introduce new problems.
Exam trap
The trap here is believing that a udev rename or a device-mapper path provides stable naming, when the supported and reliable approach is to use the existing by-uuid or by-path symlinks plus nofail.