Re: [rdiff-backup-users] atomic increment files?

[Matthew Flaschen]

Marcel (Felix) Giannelia wrote:
> OK, after writing a few little utilities, I was able to investigate this
> further last night and found that the inability to run du is not
> entirely rdiff-backup's fault -- one of the machines that's backing up
> to this server has a misconfigured sendmail that's (long story short)
> causing about 15000 modified files every day in /var/spool/clientmqueue.
> That translates into one million files in the same directory in
> rdiff-backup's increments (accumulated over 154 runs), so it's no wonder
> everything that does directory listings is choking there.
> 
> However, I stand by what I said earlier -- making thousands of little
> files on purpose is a bad design.

It seems to me you're very wrong.  A typical restore is only going to
touch a few files.  Now you tried to address this with, "accessing an
archive file's directory structure is likely faster than doing the same
in a part of the filesystem containing many thousands of files per
directory."   But you provided no evidence whatsoever for this very
non-obvious statement.  If that were true, why don't people use AVFS as
their primary filesystem?

Matt Flaschen