> A backup tool that knowingly produces corrupt backups without exiting
> with a non-zero error code is buggy -- pure and simple.

Pure, yes.  Simple, no.  Who gets to decide what counts as corrupt?

Answer: the admin choosing the backup tool.  Not (except for a very
small set of cases, irrelevant to the general discussion) you or me.

> One of the obvious things to do is to ignore the files which change,
> and another is to stop whatever changes them.  If the latter is
> impossible yet those changing file must still be backed up with a
> best-effort level then create a static copy before starting the
> backup (and still ignore the live, changing, copy).

Or just do what you easily can and let it go at that, which is probably
what the tar in question does (it's certainly what my tar does).  And
if that's good enough to satisfy the relevant admin, who is anyone else
to gainsay that decision?  (If the admin chose that program without
investigating what it did, it was probably a foolish choice, but I
don't see it as a tool author's place to refuse to make an otherwise
useful tool available just because it can be ignorantly misused for
applications it's not suited for.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B