cephfs-data-scan: scan_links is not verbose enough
cephfs-data-scan is used to recover cluster. Depending on the fs size to recover it may run for hours or even days, and a possibility to estimate how long it would run is very useful for users waiting for recover. For this we run cephfs-data-scan with --debug-mds=10 setting, and from logs may calculate how many objects (inodes) processed so far and estimate the progress. It works well for "scan_extents" and "scan_inodes" steps, but not well for "scan_links", because it does not log enough data about its activity.
This can be easily fixed by adding some debug messages to the tool.
I am creating a ticket for this simple task just to be able to backport the change.