/usr/adm/unix.log errors: ME TOO!

Kevin O'Gorman kevin at kosman.UUCP
Thu Mar 17 08:16:37 AEST 1988


In article <5030 at ihlpg.ATT.COM> bamford at ihlpg.UUCP (Harold E. Bamford) writes:
>In article <294 at icus.UUCP> lenny at icus.UUCP (Lenny Tropiano) writes:
>>What are these errors, and what should I do?
>>
>>/usr/adm/unix.log:
>	.
>	.
>	.
>>HDERR ST:51 EF:40 CL:FFC0 CH:FF03 SN:FF0F SC:FF01 SDH:FF24 DMACNT:FFFF DCRREG:94 MCRREG:9700 Sat Mar 12 14:39:11 1988
>>
>>drv:0 part:2 blk:56415 rpts:1 Sat Mar 12 14:39:28 1988
>	.
>	.
>>Can I do something about that bad block?  Or isn't it bad?

Yep, you can do something about it if it's bad, although just the one message
by itself doesn't prove it one way or the other.  Here's what the message
means:

	HDERR:  Hard disk error
	ST:	WD1010 Status Register.  I dunno what it means.
	EF:	Error register; sum of
			80: bad block
			40: CRC error
			10: ID error (sector header?)
			04: Aborted
			02: Track 0 error
			01: Data Address
	CL:	Lower byte of the cylinder address (ignore the FF)
	CH:	Upper byte of the cylinder address (ignore the FF)
	SN:	Sector number (ignore the FF)
	SC:	Sector count (ignore the FF)
	SDH:	Size, drive and head: for us, always 20 + head number.
	and I don't think the others are useful.

You can use this information to discover the exact sector that gave you
the problem, just by applying a little math to the above.  I used to know
how to remap the sector, but I would have to figure it out all over again
now.  Lately, my disk troubles have been things like seized bearings, and
I haven't had to excercise my skills in this area for some time.

It gets a little tricky with the sloppy use of terminology for logical and
physical addresses.  It helps to know that a logical block is made up from
2 physical blocks (512+512 = 1024 bytes), and that with 17 sectors (physical)
per track on most drives, there are 8 logical blocks, and a spare sector.
If you specify a logical block to be remapped, the system will remap both
of the physical sectors.

I think you use your diagnostic disk for this (though it may be okay to use
iv also).  You may need to know the magic incantation 's4test' to get into
the better parts of the diagnostic -- try it, it brings up a more complete
set of choices about what to do, though it assumes you know more about what
you're doing.

Good luck!



More information about the Unix-pc.general mailing list