Bad Block Forwarding?
Alan's Home for Wayward Notes File.
alan at shodha.dec.com
Thu Aug 24 01:12:39 AEST 1989
In article <15164 at dartvax.Dartmouth.EDU>, steve at avalon.dartmouth.edu (Steve Campbell) writes:
> Does Ultrix 3.1 (RISC) perform automatic bad block forwarding (revectoring)
> on SCI disks such as RA90's? What strategy is used? Does anything prevent
> runaway forwarding in the instance of false bad block errors that are caused
> by other electronic failures?
> Steve Campbell
> Dartmouth College
Well first the RA90 isn't a SCSI disk. In V3.0 and V3.1
the SCSI support doesn't include dyanmic BBR. There is a
program called rzdisk that will let you do this by hand.
I wouldn't be suprised if a future version supports dynamic
BBR in the SCSI drivers.
The DECsystem 5400 and 58xx, which run V3.1, use the KDA50,
KDB50, HSCxx and other DSA controllers do support dynamic
BBR. The exact stratagy varies from disk to disk, but on
most of the disks there is one extra sector per track that
is used for revectoring. Some disks may have more than one.
Except for the time I tried to use an RD52 that was unformatted
I've never seen "runaway BBR". I don't know if there is anything
built into the algorithms to watch for this, but in the worst
case the control tables on the disk will fill up or become
corrupted you'll see messages like:
Unit x write protected. Backup media and reformat.
I don't remember if this was catastrophic electronic failure
or a head crash, when it happened to me. There weren't that
many blocks that had been revectored, though.
--
Alan Rollow alan%nabeth.dec.com at decwrl.dec.com
More information about the Comp.unix.ultrix
mailing list