JC Simonetti wrote:
I have different values for the filemarks, measured with another
program and not amtapetype. Do you know if the filemarks are
I don't have a DDS3 drive, so I just made one up, that was a good
approximation. Amanda uses the filemarks to get a better calculation
of the tapecapacity. If the filemark is small compared to the
tapelength, and you don't have many DLE's to write to tape, they
are not that important. The variations in tapelength between tapes
or the number of retries when writing with a dirty head could be more
important.
application-dependant, tape-dependant, tape-and-taper-dependant ???
The IBM fms software tells that filemarks are
tape-and-taper-dependant. Do you know more? Do you have any opinions
concerning that?
Amtapetype tries to measure the size of a filemark by measuring
the difference in capacity when writing 100 files vs. 200 files to tape.
The fact that it sometimes even calculates a negative size (reported
as 0 too), means that the filemarks are not important in this case.
Some tapedrives do have a significant amount, however.
Just to be sure we are talking about the same thing. When you write
data on a tape, you write sequences of "data / mark / data / mark"
wehre the marks are here to make a separation between the different
flows of data, the "fsf 1" parameter of mt, the "EOD" (End Of Data)
mark. A tape contains a sequence like that: BOT.(data.EOD)*.EOT | BOT
= Begin Of Tape | data = well... | EOD = End Of Data | EOT = End Of
Tape
I'm not a tape hardware specialist, so don't believe me blindly.
How a tape separates the different data blocks on a tape could be
implemented in different ways. Some tapedrives use a gap, resulting
in lost capacity equal to the length of the gap. The gap distance is
used to stop the tape, and is dependent of the speed of the tape
(could be a few inches, when I used to work with 9inch reel tapes on
an IBM mainframe). Others use a distinct pattern. It could even be
implemented out-of-band on e.g. a seperate track on the tape, resulting
in no capacity loss at all.
EOT is often implemented as twice EOD.
With access to the really low level commands of a tapedrive, you could
read the data that was once written behind the EOT mark. Data rescue
centers use these techniques, if you pay them enough.
Some tapedrives (QIC tapes) write in "serpentine" tracks to the tape:
When hitting the physical end of tape, the heads are shifted a little,
and the tape is reversed, writing the data between the previous tracks.
I believe some can do 4 passes over the tapelength, each time shifting
the heads a little. While writing there is a seperate head that erases
the tape. The "erasing" head erases the tape over the complete width
of the tape, so writing a quarter of such a tape erases it completely.
--
Paul Bijnens, Xplanation Tel +32 16 397.511
Technologielaan 21 bus 2, B-3001 Leuven, BELGIUM Fax +32 16 397.512
http://www.xplanation.com/ email: Paul.Bijnens AT xplanation DOT com
***********************************************************************
* I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, F6, *
* quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, *
* stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, *
* PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, *
* kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... *
* ... "Are you sure?" ... YES ... Phew ... I'm out *
***********************************************************************
|
