zfs-import-{cache,scan}: change condition to FileNotEmpty

[nabijaczleweli]

Motivation and Context

Sometimes(?) zfs generates an empty cache file for me, and the import fails, which is suboptimal, since this means that dracut fails, and I have to type in zpool import -a to boot, delete the file, and regenerate+reinstall the initrd

Description

This works around this by ignoring a clearly bogus cache, while still supporting systems that do utilise one

How Has This Been Tested?

Applied equivalent change to the services on my laptop, regenerated and rebooted for each of the following: no cachefile => -scan, empty cachefile => -scan, non-empty cachefile => -cache. This is superiour to the previous behaviour of empty cachefile => -cache (failed).

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly. – I didn't, but I don't think there is any, so
• I have read the contributing document.
• I have added tests to cover my changes. – N/A
• I have run the ZFS Test Suite with this change applied. – dunno how to, but I don't consider it relevant here
• All commit messages are properly formatted and contain Signed-off-by.

[behlendorf]

The reason for this is that while the kernel provides interfaces to read/write existing files (from within the kernel), it does not provide one to remove a file. So in the case where we'd want to remove the cache file because all pools have been exported it's instead truncated to zero length. The zpool utility already treats a zero-length cache file and a missing cache file the same, so it makes good sense to apply this same reasoning to these service files.

[rlaager]

This looks good to me. My only concern was whether ConditionFileNotEmpty was new (in systemd). I checked that and it has been available for a long time.