Amanda-Users

Re: Tape DDS-3 values

2002-09-12 11:34:07
Subject: Re: Tape DDS-3 values
From: Gene Heskett <gene_heskett AT iolinc DOT net>
To: Galen Johnson <gjohnson AT trantor DOT org>, Kablan BOGNINI <kbognini AT yahoo DOT fr>
Date: Thu, 12 Sep 2002 11:12:48 -0400
On Thursday 12 September 2002 10:43, Galen Johnson wrote:
>Kablan BOGNINI wrote:
>>Hello,
>>
>>I am using HP DDS-3 tapes for my backup. I've tried to
>>get the correct values for my tape. But tapetype gives
>>this result:
>>define tapetype HP-DDS3-DAT {
>>    comment "just produced by tapetype program"
>>    length 9860 mbytes
>>    filemark 0 kbytes
>>    speed 840 kps
>>}
>>
>>I think this is not correct because I've a DDS-3 125M
>>tape with 12GB.
>>
>>Could someone give me more accurate values for this
>>tape or point me to doc ?
>>
>>Thanks in advance.
>>
>>___________________________________________________________
>>Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français !
>>Yahoo! Mail : http://fr.mail.yahoo.com
>
>I'm gonna go out on a limb and say you have hardware compression
> turned on...as any search through the archives will state, "turn
> if off and leave it off".  Software compression will most likely
> give you more than the tape advertises (just ask Gene).  If you
> are running linux, run "mt -f /dev/tapedevice compression 0" . 
> If you still get the same result you may have to play with data
> density settings for the drive which will require perusing the
> manuals.
>
>=G=

While you can use mt to turn it off, on most drives is a switch that 
will turn it off, a bit more dependable than remembering to do it 
with mt, if your platforms mt even has that ability.

Which brings me to the next subject regarding compression.

Most modern drives can autoswitch, and will switch it back on if it 
finds, in the tapes own, you can't read it, header, that the 
compression is on for that individual tape.  So basicly it doesn't 
do you any good to turn it off if the *(&% tape itself turns it 
back on during the tape recognition as each tape is inserted into 
the drive.

What I've found to be a way to fix that is to rewind the tape, then 
dd the amanda label block out to a scratch file, then rewind the 
tape again, use mt to shut the compression off and write 20 megs or 
so of data from /dev/zero to it using dd.  Rewind it and rewrite 
the scratch file back to restore the label.

By doing this long garbage data write, you are going to force the 
drive to flush its buffers.  When this is done, the drive will 
re-write those hidden headers, doing it with the compression flag 
now set to off.  If such a flush operation is not forced, the drive 
doesn't update those headers and you're stuck with xx gig tapes 
that only hold less because the compression is on regardless of 
your wishes.  Sometimes much less when you feed the drives rll 
compressor with a few gigs worth of bz2 files.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
99.15% setiathome rank, not too shabby for a WV hillbilly