How to Identify and Manage Archives with All-Zero File Chunks?

[jk-1]

When using Borg Backup, what is the best practice for:

1. Listing all archives that contain one or more all-zero file chunks?

2. Identifying/listing which source files have an invalid backup in a specific archive that contains one or more all-zero file chunks?
   How can we determine the original source files from which these all-zero file chunks were compressed and deduplicated?

[jk-1]

It seems that

borg [common options] list [options] [REPOSITORY_OR_ARCHIVE] [PATH...]

has a

key health: either “healthy” (file ok) or “broken” (if file has all-zero replacement chunks)

what is exactly what I need.

One question still: when is this health set / updated?

My guess is that during backup it is set "healthy" and updated ONLY during

borg check --verify-data --repair [REPOSITORY_OR_ARCHIVE]

Is this correct or are there other borg check option combinations / borg commands that can update "health" bit?

[ThomasWaldmann]

It's updated when borg check --repair detects new file chunks missing or missing file chunks reappearing.

--verify-data is not needed for that, because only presence of the chunks is determined.

BTW, that way of dealing with missing chunks might change in borg2 (see #8568).