autofs, fstab and an unavailable NAS

NFS or SMB determines how files are transferred. autofs or fstab determines when the OpenATV mount dialog arranges the connection. Either protocol can use either mode.

Choose a mode

	autofs – first access	fstab – startup
Mount trigger	Access to the share path	Startup scripts attempt the configured mount
NAS off during boot	No remote export needs mounting until accessed	Connection attempts can delay startup
NAS becomes available later	A later access can attempt mounting again	May require mounting again
No longer used	Can expire after inactivity if nothing keeps it busy	Normally remains mounted until unmounted
NAS fails during access	The active NFS/SMB connection can wait or return errors	Also depends on protocol failure/retry handling

autofs is a useful starting point for a NAS that is sometimes off. It does not guarantee fast boot: a plugin, movie list or storage check accessing that path during startup still triggers the connection then.

Local paths and HDD replacement

Checked OpenATV selection	Path
autofs	/media/autofs/NAME
fstab, without HDD replacement	/media/net/NAME
fstab, with HDD replacement	/media/hdd

NAME is Local share. With autofs the path remains under /media/autofs even if HDD replacement was selected. Prefer the explicit recording directory for NAS recordings. /media/hdd is not available for another mount when a local HDD already uses it.

Relevant configuration files

The checked manager reads and writes network mounts in /etc/auto.network and /etc/fstab. Older guides often mention /etc/enigma2/automounts.xml; its import is disabled by default in this source version. Do not assume old configuration recipes apply to the new manager.

/etc/auto.master connects the map to its base path. The test receiver uses:

/media/autofs /etc/auto.network --ghost

--ghost can make directory names visible before the remote server is mounted. A visible folder does not establish connectivity.

These are alternative NFS definitions for one share. Replace address and export path; do not create both:

/etc/auto.network

nas-recordings -fstype=nfs,rw,proto=tcp 192.0.2.10:/srv/recordings

/etc/fstab

192.0.2.10:/srv/recordings /media/net/nas-recordings nfs _netdev,rw,proto=tcp 0 0

_netdev identifies a network filesystem; it does not cap waiting time. The final two 0 fields belong to fstab, not an autofs map. Use the menu editor for ordinary changes. Saving can regenerate recognised manual network entries, so avoid conflicting definitions maintained through multiple tools. SMB mount files may also contain passwords.

The checked image uses classic init scripts. x-systemd.automount from systemd instructions is not a replacement for this autofs arrangement. nofail is not a general instruction to never wait for the NAS either.

Why does a spinner appear?

Even checking a directory can trigger network access. Opening a movie list, thumbnails, free-space checks and stat() or path-existence queries can all do so. The checked mount manager uses such an existence check for autofs targets.

If an operation waits in the interface’s execution path, Enigma2 can display its spinner. This initially means busy; it does not establish a crash or faulty skin. Even retrieving a file’s size can be delayed by an unavailable NAS.

An autofs idle timeout controls expiry of an unused mount. It is not a deadline for a blocked file operation. NFS timeo, retries and soft/hard mounts operate at other levels; see NFS options. An active recording keeps storage busy and normally prevents idle expiry.

During an outage

1. Note whether recording, playback or timeshift was active and when the fault began. Avoid further access to the same unavailable path.
2. Check the NAS, switch/router, cable and address. Restore the same server and export where possible; this can release waiting requests.
3. Once the interface responds, stop affected applications normally. Do not start new recordings on the failed destination.
4. If the NAS will remain off, move recording, timeshift, EPG and plugin paths to reachable storage and disable the unused mount definition.
5. Collect debug and kernel evidence for recurring problems and report the fault. A GUI restart does not reliably remove a network request still waiting in the kernel.

Read registered network mounts through SSH:

grep -E ' (nfs|nfs4|cifs|autofs) ' /proc/mounts

This list does not prove that the server currently responds. Unrestricted df -h, ls or stat may access the failed share and wait themselves. Do not force unmounting during writes. An outage may already have damaged a recording even if the interface later resumes.

Verification: File mappings and autofs configuration checked. The existing NAS mount was not deliberately interrupted; no fixed delay is presented as a measured result.

Sources: OpenATV mount management and status checks, autofs master map.