Re: [rdiff-backup-users] Version 0.11.1 released

[dean gaudet]

>        Directory statistics file:  Although the session statistics file is
>        still generated, the directory statistics file no longer is,
>        because the new code structure makes it less inconvenient.

this is kind of unfortunate... i use the directory_statistics file to see
what portions of my filesystem are changing the most.

since i have to do post-processing on the dir stats file anyhow (i.e.
sorting) it's no problem for me to collect the data from an even larger
raw file than i process currently.

prior to the built-in dir stats i was using a find, but the find takes a
long time :)  and find methods also make it more of a pain to discover the
NewFileSize statistics.

would you consider adding a --file-stats option maybe, which logs the
stats on a per-filename basis?  something like:

# RelativeFileName NewFileSize DeletedFileSize IncrementFileSize

where only one of those three numbers are non-zero... that's certainly
enough to recreate the data which i currently view daily.

-dean

p.s. this is all i currently do at the end of my backups:

dirstat="`ls -1tr root/rdiff-backup-data/directory_statistics.* | tail -1`"
echo ""
echo "NewFileSize over 5MB:"
zcat $dirstat | awk '!/^#/ && $7 >= 5000000 { printf "%15u %s\n", $7, $1 }' \
        | sort -nr
echo ""
echo "IncrementFileSize over 5MB:"
zcat $dirstat | awk '!/^#/ && $14 >= 5000000 { printf "%15u %s\n", $14, $1 }' \
        | sort -nr