Revectoring bad blocks on RA81 disks

[Chris Torek]

Important announcement:  I have set up a new mailing list,
info-uda50 at maryland, a.k.a umcp-cs!info-uda50.  Anyone who wants
to subscribe, please send mail to info-uda50-request at maryland
(or umcp-cs!info-uda50-request).
------
(enough of that)

Speaking of forced error data errors, I found that doing a replace
operation with a forced error modifier made our UDA50 controller
say "yes, I did the replacement", but the bad block descriptors
were unchanged.  Removing the forced error modifier made it work.
Since then the controller microcode has been "upgraded", so I
suppose it could have been just another UDA50 bug....

The cylinder group summary information must (obviously) be rewritten
every time the cylinder groups change, which is certain to be very
often, but it can be regenerated, so I wouldn't worry about losing
it.  Interesting that the sectors containing that info are the ones
to first start showing errors.  Perhaps the oxide is getting
attracted to the disk heads :-).
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at maryland