Veritas-bu

[Veritas-bu] DLT Tapes

2001-02-15 09:53:37
Subject: [Veritas-bu] DLT Tapes
From: Andrew Shinkarev shinkara AT pprd.abbott DOT com
Date: Thu, 15 Feb 2001 08:53:37 -0600
On Wed, Feb 14, 2001 at 06:46:24PM -0800, Raymond Jackson wrote:


Not quite sure but as far as I know it's Fuji and Maxell 

We are using STK L700 libraries with Quantum DLT 7000 tape drives
inside and Imation DLT IV tapes
Imation brand contains Maxell and Fuji mix, 4 to 1 or so.
Maxell made tapes from this mix gave us 1 % bad tapes, Fuji - 2.5. 
Most tapes became unusable after 3-4 mounts and 20-50% capacity used. 
Based on serial number it's possible to say who made a tape:
SH949M3901 = Maxell
     |
MA003F640 = Fuji
     ^ 
Sometimes suppliers offer free replacement for bad tapes.

Another factor is bad tape drives Then a tape drive goes bad you 
might get a bunch of frozen media without concern who made these tapes  
and this might make all your calculations worthless.
So, I would love to buy some tapes from both Fuji and Maxell
and count bad tapes after 6 months of using them.  
I got these numbers based on using 3300 Imation DLT IV tapes.
If anyone has his own number  please, please post it. 

In January  this issue was disscused already:

[Veritas-bu] FW: media write errors - 84

Thorpe, Alastair at78748 AT GlaxoWellcome.co DOT uk
Thu, 11 Jan 2001 10:59:24 -0000 

      Previous message: [Veritas-bu] Compression Issues on L1000 [with DLT7000] 
      Next message: [Veritas-bu] Cleaning Tapes 
      Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] 



        We have four similar installations (although each is independent - a
unique NB Cluster) using the Sun L11000 (Aka ATL P3000), although we are
still using NetBackup 3.1.1 with Solaris 2.6 on E3500's.  All four
installations are identically configured, even down to the firmware
revisions on the devices.   For only one of these installations (physically
co-located with one of the others), we have been plagued by 'media write
errors (84)' and are still conducting our own root cause analysis. We
followed the same path of swapping out drives that were giving the most of
these errors but the problem was not eradicated. 

        For what it is worth, we involved our media supplier, who brought in
media specialists from Imation (who currently do not sell DLT Media so are
reasonably impartial).  We were told that: "it is possible for a dodgy DLT
drive (perhaps heads are misaligned) to be capable of writing data to a DLT
tape 'successfully', then if this tape is loaded into another perfectly good
DLT drive, then this can have problems reading the tape, and perhaps the
cleaning LED comes on.  This can lead you to see errors in other perfectly
good drives when this 'poorly written' media is loaded and make you believe
there is a problem with them, when in fact it was a different drive that
originally wrote the data".

        Once pragmatic step we have taken on the media suppliers advice,
which are resulted in a significant reduction (6x reduction) was to switch
DLT tape brand from a Fuji to Maxell manufactured media.  Some brands (e.g.
Quantum) co-source so our supplier checks the batch numbers. Apparently,
there are currently only two manufacturers world-wide, licensed by Quantum
to make DLT media (FUJI and Maxell).  Both manufacture to strict standard
defined by Quantum and employ rigorous Quality Control.  If there had been
problems with a certain batch then this would be known to them.  So there is
no suggestion that the original FUJI media we were using in this one
installation is anything but perfect.  In fact, we have re-deployed
elsewhere and it works fine. I do not know how to explain the improvement,
but just know for this particular installation (a complex interaction of
hardware and software components, plus environmental conditions) it has made
a positive improvement.  However, we still see 1-2 Media write errors/week,
on a hard-worked L11000, and would ideally like to eradicate them altogether
if this was possible.

        So this discussion thread is very interesting to us also.  Further
actions we are planning are to upgrade NB to 3.2 or 3.4 and are also looking
at upgrading the firmware in the L11000.   

        Regards
>       Alastair 
>       GlaxoSmithKline plc
> 



> 
>   I'm told by the folks who support our libraries that there
> are only two factories that make DLT tapes.  Any vendor you
> buy them from has bought them from one of those factories.
> I remember that one of them is Matsushita's (sp?) but I can't
> remember the other.  One of the codes on the back of the
> cartridge indicates which factory it came from.  When I asked
> which was better I was told that each has had bad runs, but
> in general they're about the same.
> 
> Ray
> 
> Dennis Dwyer wrote:
> > 
> > I've had more than 20 tapes fail within the last three months and all of 
> > them were Fuji. For my money I'm going to stick with Quantum or Compaq 
> > tapes going forward.
> > 
> > Dennis Dwyer
> > Tampa Electric Company
> > 
> > >>> "McCormick, Diane S" <Diane_McCormick AT AIMFUNDS DOT COM> 02/14/01 
> > >>> 16:22 PM >>>
> > Would like to know if there are any preferences, comments, data, etc. on
> > best brand of DLT tapes.....we currently use Quantum but are looking at
> > Fuji.....thanks
> > 
> > Sam McCormick
> > Production Control
> > Data Center Operations Support
> > AIM Management Group, Inc.
> > www.aimfunds.com
> > ph:  713.986.9057
> > email:  Diane_McCormick AT aimfunds DOT com
> _______________________________________________
> Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
> http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

-- 
----
Andrew Shinkarev
Network Systems Specialist
Abbott Laboratories
(847) 9387559
shinkara AT pprd.abbott DOT com



<Prev in Thread] Current Thread [Next in Thread>