I have a direct repository with the assistant running (v4.20131002). The repo is in the backup group and should be grabbing every known annex file. Yet git annex status says:
local annex keys: 1386
local annex size: 94.53 gigabytes
known annex keys: 1387
known annex size: 94.53 gigabytes
As you can seem there is one more known annex key than there are local ones. Neither git annex get nor git annex sync changes the numbers. According to tree the repo dir contains exactly 1387 files, which means that the missing file must exist on disk.
1) How do I find out which is the file in question? How do I get it synchronized?
Yesterday there were 1376 known annex keys. Today there are 1387. For some reason the number of keys went up by 11, and I would like to know how and why. The log suggests that a couple of files have been added from inside the repo directory, but I am positive that these files already existed in the annex directory yesterday, and that I ran git annex sync manually without the number of local and/or known keys going up.
2) Why weren't the files synced before? Is there a way to find out whether they were or were not part of the annex yesterday - maybe list all files sorted by the date they were added?
I think you're stressing way too much about a number.
It's quite possible for the known and local key counts to diverge while no data is missing. For example:
I don't think I'm stressing too much. Rather, I'm inquiring about why git-annex is giving me wrong (or at least unclear) output that makes me worry about losing data. I plan to trust git-annex with a lot of important files, and in order to do that I want to make sure it's trustworthy. And that's where the inaccurate key numbers get in the way.
So why do the local and known numbers diverge anyway? Would you consider adjusting the status output to be more accurate, or at least less confusing? I just want to be able to check that every single file in the repo is being backed up.