I often discover empty recordsdata as an alternative of roughly essential paperwork. Thankfully no actual injury to date as a result of I’ve many backup methods (together with good previous paper).
The file exists with its identify intact however the content material is empty. The file is displayed by the finder with “Zero bytes measurement” (to be appropriate: it needs to be “Zero byte measurement”).
The file kind is often jpg or pdf. These are the categories I embody in LaTeX paperwork and seek advice from with Unix onerous hyperlinks (ln).
I’ve been monitoring this downside for a very long time however discovered no rationalization and no approach to reproduce systematically.
I even wrote a script to be able to detect the alternative of a full file by an empty file with the identical identify at backup however but couldn’t pin it down.
Has anybody observed the identical phenomenon?