SUMMARY -- Cleaning Exabytes (repost)

From: Carlo Tiana (carlo@white.stanford.edu)
Date: Fri Jul 19 1991 - 14:35:01 CDT


Apologies for previously posting this without a subject line, as someone
rather rudely pointed out. Here it is again, in case you search for it by
subject line in the future.
As an addendum, the only physical difference I notice between my cleaning
tape (not Exabyte brand) and normal magnetic cassettes is in the large
hole in the center of the box - on the "bottom" side if you will - that
on the cleaning tape is coated with a shiny substance (looks like foil or
something) whereas in the normal tapes it is not. It also looks as if part
of the tape drive head mechanism contains an LED/tiny light bulb and sensor
that rise into this hole when the tape gets loaded. If this is indeed the
mechanism to detect a cleaning tape (the "doohickey" many of you
mentioned), I am impressed at how complicated a system this is (rather than
a simple hole somewhere, like for write protection). This is all
speculation, so I may be far off base here - I just thought I would include
it as a curiosity.

Earlier today I wrote:
>
> We have a couple of Exabyte drives that have undergone a lot of
> use by now, and the time to clean them has come. So we have equipped
> ourselves with the cleaning kits and are now asking ourselves the
> question - how? Something like "mt -f /dev/rst0 clean" ? :-) I mean,
> how do I get them to run through the tape a little bit when there
> is no "Play" button on the front? I have not tried just dumping
> to a cleaning tape rather than a magnetic tape, assuming it won't
> work, but maybe that's what I need to do? Any suggestions/tips
> would be appreciated. As usual, I'll summarize.
> Carlo.
>
Well, thank you for the numerous replies, many of which are still coming
in, and will continue to do so I am sure. All credits at the end of this
message. Here's the gist of the replies with excerpts of the more original
ones.

Practically all of you said basically: "pop in the cleaning cartridge, the
drive does the rest". People were split about evenly on the subject of
whether the drive "recognizes" a cleaning tape as such and uses it, or
whether the drive simply winds back and forth a bit in search of a BOT
mark, doesn't find one and spits the tape out - achieving much the same
effect. Either way, the heads get clean.

Some people suggested that indeed one should do one of the following
"software-based" things, in order for the cleaning to occur:
retension the tape (with the mt retension switch)
write/read data to/from the cleaning tape

Some of you said "absolutely only ever use Exabyte tapes", other said any
cleaning tape would be ok. The Exabyte tape comes in 2 flavors, one good
for about 3 cleanings, one good for about 12, the latter reportedly only
works on "later model" or "< 2yr old" units.

Then some of you offered various other related advice, as follows:

From: Mike.McCann@eng.clemson.edu (Mike McCann)
Rule of thumb for cleaning your tape heads: 30Gb or 30days, whichever comes first.

From: rick%pgt1@Princeton.EDU (Rick Mott)
We have actually had a lot of experience with these things, and found out
the hard way that they are VERY sensitive to dirt/dust in the environment,
especially if the enclosure's cooling fan sucks instead of blows. Ours does.
Symptom: write errors start creeping in after a few months, then get a lot
worse rather rapidly. Make sure your enclosure has an adequate filter on
the fan, and clean/replace it regularly. If it ain't there, buy a piece
of open-cell foam air conditioner filter (any hardware store, 79 cents)
and tape a double layer across the fan. That's what we did, and haven't
had a problem since...

From: tyen@mundo.eco.utexas.edu (Anthony Yen)
.... who uses checks the output of dump and checks/keeps track of:
1) the number of blocks written, and 2) the number of retries
the drive sensed. The latter is read with a
'mt -f /dev/nrst0 status'. If the retries/total_blocks ratio
goes over 2-5%, or the total blocks written since the last
cleaning cycle reaches 30 Gb (whichever comes first), it's
time to clean.

From: deltam!tigger!jt@uunet.UU.NET (jim wills)
Carlo,
The firmware , and the sensors of the Exabyte know when a cleaning cartridge is inserted . The tape drive will go through a 20 second cleaning routine and will then spit the tape out.

This will only work with the Mx Card Firmware of 4$22 and up. (Drive Two years old or newer.

------------------------------------------------------------------------
------------------------------------------------------------------------
Thanks to:
From: daniel@canr.hydro.qc.ca (Daniel Hurtubise)
From: button@alc.com
From: Melissa Metz <melissa@broccoli.cc.columbia.edu>
From: pbh@CFSMO.Honeywell.COM (Paul Henninger)
From: Chip Christian <chip@allegra.att.com>
From: Joe Angelo <angelo@enterprise.arc.nasa.gov>
From: Mike.McCann@eng.clemson.edu (Mike McCann)
From: Randy Born <randy@ai.iit.nrc.ca>
From: holle@ASC.SLB.COM
From: scott@dsg.tandem.com (Scott Hazen Mueller)
From: ept@eptsun1.ctd.ornl.gov (E P Tinnel)
From: tom@yac.llnl.gov
From: bks@okeeffe.Berkeley.EDU (Brian K. Shiratsuki)
From: doug@perry.berkeley.edu (Doug Neuhauser)
From: phillips@athena.Qualcomm.COM (Marc Phillips)
From: Peter Smith <peter@sucia.stanford.edu>
From: feldt@phyast.nhn.uoknor.edu (Andy Feldt)
From: canuck@rice.edu (Mike Pearlman)
From: David Fetrow <fetrow@biostat.washington.edu>
From: rick%pgt1@Princeton.EDU (Rick Mott)
From: trinkle@cs.purdue.edu
From: Angela Thomas Hoynowski <angela@everest.tandem.com>
From: tyen@mundo.eco.utexas.edu (Anthony Yen)
From: Kenton C. Phillips <kenton@space.mit.edu>
From: hal@mpl.UCSD.EDU (Hal Skelly)
From: deltam!flyer!mark@uunet.UU.NET (mark galbraith)
From: deltam!tigger!jt@uunet.UU.NET (jim wills)
From: eddieb@EE.MsState.Edu
From: herman@galileo.csc.ti.com (Herman Schuurman)
From: wallen@cogsci.UCSD.EDU (Mark R. Wallen)
From: Gary Barber <gary@seismo.gps.caltech.edu>
From: librainc!ho@uunet.UU.NET (Alan K. Ho)
From: gerry@jtsv16.jts.com (G. Roderick Singleton)
From: jmcrowell@ucdavis.edu (John M. Crowell)
From: markg@shasta.nextwave.com (Mark Glasser)



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:20 CDT