Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!rice!sun-spots-request
From: henry@zoo.toronto.edu
Newsgroups: comp.sys.sun
Subject: Re: Defect Lists
Keywords: Miscellaneous
Message-ID: <5747@brazos.Rice.edu>
Date: 13 Mar 90 19:54:46 GMT
Sender: root@rice.edu
Organization: Sun-Spots
Lines: 20
Approved: Sun-Spots@rice.edu
X-Refs:  Original: v9n79, Replies: v9n81
X-Sun-Spots-Digest: Volume 9, Issue 81, message 4

>Some defects don't show up at format time. They wait until the most
>inopportune moment...  Disk defects don't usually go away. The set of
>actual defects on an old disk is a superset of the original set of
>defects, recorded in the manufacturer's defect list. 

Our experience, with Fujitsu Eagles that have been in production for most
of a decade, is quite the opposite:  the two lists seem to be almost
entirely unrelated.  The story we heard, way back when, was that the
manufacturer's defect lists on such drives are generated by analog means
-- he has no idea how you're going to format the drive -- and say very
little about digital results.  This was from George Goble, who I am
inclined to consider a reliable source.  The fix is to run your
formatter's diagnostics for *several days* per drive, and map out anything
that it hiccups on.  That was G.G.'s recommendation, and it worked just
fine for us too.

I think it's important to distinguish what kind of drive we're talking
about.  The situation for SCSI drives is probably very different, since
there much more of the formatting is the manufacturer's job.  I've never
worked with those and can't comment on how that affects things.