[H-GEN] Any recommendations for Tape Backup software ?

[Greg Black]

[ Humbug *General* list - semi-serious discussions about Humbug and     ]
[ Unix-related topics. Posts from non-subscribed addresses will vanish. ]

On 2003-11-10, David Jericho wrote:
> Greg Black wrote:
> 
> >Well, maybe.  The ability to do that depends on the hardware and
> >firmware on the tape drives; the capabilities of the operating
> >system's drivers (often far from state of the art); and other
> >arcana such as the phase of the moon.
> 
> Sorry? I don't understand what you're getting at. I was under the 
> impression that the software at hand simply was smart enough to 
> understand an end of file, or end of tape, and ask for a new one.

As I said, it depends on the hardware/firmware (and, to a lesser
extent, on the OS).  There are many tapes that will eventually
report an end of medium status -- but it's difficult to know,
from the application end, exactly how much data was successfully
written from the last chunk and it's not guaranteed that a
subsequent attempt to read the data back will result in the same
data that you thought you wrote.  Merging the end of such a read
with wherever you started the next tape is not always easy and
it's sometimes impossible if you just started writing the next
tape from the apparent point of failure.  Not all tape drives
have this issue, but lots of units in current service do.

> I'm not using any tape drive smarts to do the spanning, it's all in the 
> backup software. Dump can do it, tar can do it, Bacula can do it. Amanda 
> can't do it.

With the tape drives and OS variants at my disposal, I can
demonstrate many cases where dump and tar can definitely not do
it; I have no experience with Bacula, but it would impress me if
it could manage this correctly.

One of the difficulties with testing these things is that you
really have to write full tapes and, for most systems, that
takes time measured in hours rather than minutes.  And you have
to do repeated tests, because sometimes luck is against you and
the thing seems to work because of flukes in data block sizes,
the length of the particular tape, the number of errors and
re-writes on the tape, etc.

> >I'm not fond of amanda, either, but for a far more complex set
> >of reasons than this particular issue which is beyond the reach
> >of software fixes unless you have the right combination of stuff
> >for it to play with.
> 
> Of course, my reasons are far more in depth that I have interest in 
> posting to this list. Far too time consuming for one. And I've already 
> got a reputation as a ranter, I don't want to enhance it yet further 
> again :)

I don't understand the point here.  Anyway, in case you were
asking for more information, the reason I gave up on amanda was
the difficulty of finding the right tape to use to restore a
specific bit of data -- especially when you are relying on the
untrained office staff in a customer's office on the other side
of the world.  Actually, I'm pretty sure there was an additional
reason, but it escapes me now.

Cheers, Greg

--
* This is list (humbug) general handled by majordomo at lists.humbug.org.au .
* Postings to this list are only accepted from subscribed addresses of
* lists 'general' or 'general-post'.  See http://www.humbug.org.au/