> > Personally, I'd say don't fu*k around with something you don't get
> > source code to.
>
> Well, you certainly have a good point. Possibly it is worth it for me
> to
> amend my statement...
>
> > I sincerely wish we'd gone with GNU tar or afio or amanda. I think
> > management decided they needed something to point a finger at; they
> > didn't give free programs much of a chance. Now we're MAJORLY
> invested
> > in networker.
>
> The point that I was trying to make was that I understand tar and cpio
> are
> very unreliable, in terms of trying to rescue a botched tape, or
> screwed
> up archive. I'm not really familiar with the algorithm that amanda
> uses.
This is not true with cpio. I believe cpio to be one of the most
reliable. It can and will recover data from a tape with media errors.
Naturely the data where the media error occur is unrecoverable, but cpio
will skip to the next uncorrupted file.
>
>
> Also, its great that you have programmers at your disposal, and too
> bad
> that management doesn't allow you to make use of them. I know that
> the
> bru developers are probably on this list, and certainly on the
> linux-tape
> list, and actively pursue its development.
>
> Also, there is a demo copy available, and tech support, at least for
> installation, available for bru. If it doesn't work on your platform,
> at
> least you tried, then use amanda otherwise..
>
> > Interpreted languages are not fundamentally less reliable than
> compiled
> > languages. Both can be used well, or used poorly.
>
> True, but that really isn't an issue. I just meant don't use
> someone's
> shell scripts that are a wrapper for tar/cpio/etc, because the
> infrastructure is not sound. I think GNU tar now handles device
> files,
> like those in /dev ok, but normal vendor's tar chokes here. (Possibly
> I'm
> wrong -- I do know that there are pros and cons to both tar and cpio)
>
> Dave