Tape Backups are NEVER Reliable - EVER

On Tue, 22 Jun 2004 20:33:16 GMT, "Ron Reaugh"
wrote:


"mschlack" wrote in message
. com...

-snip

One thing to clarify: do you have both restore and archive needs?
Restore would be to rebuild after a disaster or after losing or
corrupting specific files. Archive would be for
long term retention -- rarely used data that's taking up space
otherwise but would need to be occasionally mounted at some future
point (like parts drawings for obsolete products). If you're truly
archiving, then tape probably is a must (or optical), since the
reliability of data that's never read on disk drives can't be assumed
for many years (those little old bits can flip on you). SDLT is a
solid choice, in any event.

Tapes are NOT for long term storage. Only some opticals have been legally
approved.

Errr... false.

Ron is mangling the fact that for archive purposes, the physical
characteristics of WORM opticals have resulted in that media being
approved as a tamper-proof archive.

But tape has been approved by people like the National Media Lab for
long-term storage.

Malc.

"Malcolm Weir" wrote in message
...
On Sat, 19 Jun 2004 21:14:47 GMT, "Ron Reaugh"
wrote:

[ Snip ]

In short, I need some type of recommendation, in writing, in some type
of
white paper, from some type of credible sources, that SDLT tape backup
drives, at least for the purpose of long term archiving

Tapes have NEVER been considered a viable "long term archiving" medium.

Neither, though, has anything else, except for paper.

No, some opticals.

Tape is, and has been proven to be, better than most things, and
particularly better than disk whose failure modes tend to result in
massive data inaccessibility more often than those of tape.

However, this issue is not tape-vs-anything else, it's on-line and
on-site vs. off-line and off-site.

Not these days as on-line and off-site is becoming viable.

Hey what about off-line and on-site...there must be a market for that or is
that called DAT.

[ Snip ]

Reliability is always the sum of all such factors and any backup strategy
should look more towards the least common denominator...Murphy....an
automatic corrollary to Murphiy's law is that tapes are unrelaible. The
proof of that is the incredible cycle strategies that have developed over
the years for tape backups. That comes from the fact that too frequently
the tape isn't usable for any one of a number of reasons.

Substitute "backup medium" for tape and there would be something
useful in the above...

Nope, I got it right with "tape".

On Fri, 02 Jul 2004 02:19:19 GMT, "Ron Reaugh"
wrote:


"Malcolm Weir" wrote in message
.. .
On Sat, 19 Jun 2004 21:14:47 GMT, "Ron Reaugh"
wrote:

[ Snip ]

In short, I need some type of recommendation, in writing, in some type
of
white paper, from some type of credible sources, that SDLT tape backup
drives, at least for the purpose of long term archiving

Tapes have NEVER been considered a viable "long term archiving" medium.

Neither, though, has anything else, except for paper.

No, some opticals.

Nope. While the media may be (relatively) stable, the mechanisms to
process that media is the problem! How many 12-inch WORM drives do
you have?

Tape is, and has been proven to be, better than most things, and
particularly better than disk whose failure modes tend to result in
massive data inaccessibility more often than those of tape.

However, this issue is not tape-vs-anything else, it's on-line and
on-site vs. off-line and off-site.

Not these days as on-line and off-site is becoming viable.

It's been "viable" for decades, for some value of "viable".

Hey what about off-line and on-site...there must be a market for that or is
that called DAT.

Off-line, on-site is a typical tape archive.

Reliability is always the sum of all such factors and any backup strategy
should look more towards the least common denominator...Murphy....an
automatic corrollary to Murphiy's law is that tapes are unrelaible. The
proof of that is the incredible cycle strategies that have developed over
the years for tape backups. That comes from the fact that too frequently
the tape isn't usable for any one of a number of reasons.

Substitute "backup medium" for tape and there would be something
useful in the above...

Nope, I got it right with "tape".

If you had much of a clue (which is debatable), you'd understand that
*any* backup medium is inherently "unreliable", in the sense that it
can (and does) fail. All you are trying to argue is that one subtype
of medium is intrinsically worse than others, which is obviously
nonsense given the different possible implementations of that subtype.

For example, no-one with any integrity would claim that, say, Zip
disks were a "reliable" media, but it takes a certain amount of
ignorance to extrapolate from that data point to the conclusion that
removable disks are unreliable (which is what you've done with tape).

Add to the mix the issue of undiscovered problems, and you are
*beginning* to touch on why we have complex backup medium strategies.

Malc.

"Malcolm Weir" wrote in message
news
On Tue, 22 Jun 2004 20:33:16 GMT, "Ron Reaugh"
wrote:


"mschlack" wrote in message
. com...

-snip

One thing to clarify: do you have both restore and archive needs?
Restore would be to rebuild after a disaster or after losing or
corrupting specific files. Archive would be for
long term retention -- rarely used data that's taking up space
otherwise but would need to be occasionally mounted at some future
point (like parts drawings for obsolete products). If you're truly
archiving, then tape probably is a must (or optical), since the
reliability of data that's never read on disk drives can't be assumed
for many years (those little old bits can flip on you). SDLT is a
solid choice, in any event.

Tapes are NOT for long term storage. Only some opticals have been
legally
approved.

Errr... false.

Ron is mangling the fact that for archive purposes, the physical
characteristics of WORM opticals have resulted in that media being
approved as a tamper-proof archive.

Nonsense as anyone who reads this whole thread can see for themselves. TDK
says that their CD-R are good for over 50 years.

But tape has been approved by people like the National Media Lab for
long-term storage.

Wacko, they have nothing to do with the government but are a tape
manufacturer. Also I notice that their controversial pages on media lifetime
aren't on their website any longer.

"Malcolm Weir" wrote in message

Off-line, on-site is a typical tape archive.

On that we agree with the thread title.

On Fri, 18 Jun 2004 20:22:31 GMT, "Jolly Student"
wrote:

Okay Folks:

........................
I work for a mid-sized company (600 employees) whose "Technology Director"
has openly said that "Tape backups are not reliable". This director had a
"consultant" come in to back up his assertion, a consultant who asked to
check his email via his "AOL" account (indeed, his email address is
something like ).

......................

Roger.


Roger,

Most reasonable storage architects realize that tape is a part of a
whole storage solution. Just because your compaly bought a swiss-army
knife doesn't mean that the knife is gonna meet all your company's
needs. You seem to sense it but there will be expectations of that
RAID box which it will not be able to meet and that time will come
sooner and not later. That RAID5 box will get used up quicker than
you'll think. Think of a carpenter, he has all the power tools but he
still has his old handy dandy hammer.

If your storage capacity is even enough for a few years, it won't meet
the storage needs of what you think now, later. Your manager will
eventually have to step up to this blunder when a huge litigation or
the IRS pops up and demands data from six years back. Can you imagine
the management head ache of six years of data on one RAID box even if
it could hold that much? I'd say you may be livin easy now but you
and your manager are lookin at a major ass kicking later on. Your
manager will get fired and you will have to fix that problem later if
you don't get fired too.

I'd try to nip that in the bud, now! Go talk to a reputable storage
architect and get the real skinny on backup and disaster recovery.
Also, research your state's corporate data retention policy and the
IRS. By law, there a lot of data companies are expected keep and
retain for a certain period of time such as email, HR, accounting and
more. Use this as your justification against a pure disk backup.

Boll Weevil

On 18 Jun 2004 16:33:22 -0400, (Al Dykes) wrote:



My take on this is _nothing_ is reliable (as in never screws up).
(cite: Rules, by Murphy)



Yea.. so what? We'll all die too.

On Sat, 19 Jun 2004 02:19:38 GMT, Faeandar
wrote:
Start looking for another job and soak up the sun while you can.
These guys are going to eat it sooner or later and if you're still
around you will take the blame, somehow someway.
Even if you have a dated e-mail showing your recommendations and how
this course of action will lead to the company going under.

Sell any stock you have and get the hell out.

But to address your question:
I don't know of any white papers that are written for someone so
idiotic. All I've seen assume you actually want to recover from a
disaster. And most of those assume you some semblance of
understanding when it comes to reliability and recoverability.

If "the director" looks at you like a bug when you try to explain what
will happen in the event of a plane/earthquake/virus then you've
already lost. Unless you can get the ear of someone further up you're
screwed.

Good luck on that, I'd be interested to hear how it pans out.

~F


I totally agree with Faeandar. If you haven't already done so,
document all your recommendations now especially if you plan to keep
your job. That director of IT will be comin to kick some ass and at
least you can defend yourself.

Boll Weevil

Malcolm Weir wrote:

On Sat, 19 Jun 2004 21:14:47 GMT, "Ron Reaugh"
wrote:

[ Snip ]

In short, I need some type of recommendation, in writing, in some type
of white paper, from some type of credible sources, that SDLT tape
backup drives, at least for the purpose of long term archiving

Tapes have NEVER been considered a viable "long term archiving" medium.

Neither, though, has anything else, except for paper.

Papyrus, clay tablets, engraved gold, and chiseled stone also have track
records.

Tape is, and has been proven to be, better than most things, and
particularly better than disk whose failure modes tend to result in
massive data inaccessibility more often than those of tape.

However, this issue is not tape-vs-anything else, it's on-line and
on-site vs. off-line and off-site.

[ Snip ]

Reliability is always the sum of all such factors and any backup strategy
should look more towards the least common denominator...Murphy....an
automatic corrollary to Murphiy's law is that tapes are unrelaible. The
proof of that is the incredible cycle strategies that have developed over
the years for tape backups. That comes from the fact that too frequently
the tape isn't usable for any one of a number of reasons.

Substitute "backup medium" for tape and there would be something
useful in the above...

Malc.

--
--John
Reply to jclarke at ae tee tee global dot net
(was jclarke at eye bee em dot net)

On Fri, 02 Jul 2004 03:44:10 GMT, "Ron Reaugh"
wrote:

"Malcolm Weir" wrote in message
news
On Tue, 22 Jun 2004 20:33:16 GMT, "Ron Reaugh"
wrote:

Tapes are NOT for long term storage. Only some opticals have been
legally approved.

Errr... false.

Ron is mangling the fact that for archive purposes, the physical
characteristics of WORM opticals have resulted in that media being
approved as a tamper-proof archive.

Nonsense as anyone who reads this whole thread can see for themselves. TDK
says that their CD-R are good for over 50 years.

Poor Ron, he gets so confused! One might assume, from reading the
above, that TDK's CD-Rs had been "legally approved"!

Which is, of course, false.

But tape has been approved by people like the National Media Lab for
long-term storage.

Wacko, they have nothing to do with the government

Errr... the US Government provided the funding for them, under the
National Technology Alliance program, which was a DoD/CIA think (which
became a NIMA thing, and is now a NGA thing, but all the players
remained the same, just the acronyms changed).

but are a tape manufacturer.

And TDK *isn't* a CD-ROM manufacturer????

So Ron would have us believe that TDK's claim that their CD-Rs are
"good for over 50 years" is somehow valid, and the claims of the
government-funded NML research managed by 3M/Imation are not.

Hmmm...

Still, anyone who believes that CD-R is magically stable (despite
being dye-based) while magnetic tape isn't stable has fundamentally
failed to understand the issues involved. Neither are stable... but
both can be made pretty darn good *if* you treat them right!

Also I notice that their controversial pages on media lifetime
aren't on their website any longer.

The NTA restructured about 2 or 3 years ago, with the NML, the NCAT,
and whatever-it-was-called-at-Sarnoff (NIDL, I think) being rolled up
into a new entity managed by a single corporation (as opposed to three
separate management contracts).

So the government-funded NML ceased to exist at that time. But their
research hasn't gone away, even though their funding source has.

Sorry, Ron, you'll have to get a clue!

Malc.