disk transfer error

[Alan's Home for Wayward Notes File.]

In article <781 at larry.sal.wisc.edu>, cathy at larry.sal.wisc.edu (Cathy Accettura) writes:
> Last week I was doing a dump and I got this error
> 
> DUMP: (this should not happen) bread from /dev/rra1c [block 151984] :
>       count = 6144, got -1
> 
> I am running Ultrix 3.0 and have a MicroVaxII.  The disk that is 
> causing me problems is a Fujitsui Eagle.  
> 
> Is there anyway I can use badsect or bad144 to get around this problem
> without format the disk.  I thought that badsect was my best bet but
> I could not figure out how block and sector related to each other.

	That fact the disk is called "ra" suggests you have it
	on a DSA controller or a controller that makes ULTRIX
	think it is a DSA controller.  In this case badsect and
	bad144 aren't applicatable, ULTRIX should be doing BBR
	to replace the block.  Read the documentation on radisk(8).
	If it was a bad block then probably got replaced and has
	been marked with "Forced Error", which basically means
	ULTRIX couldn't get a good copy of the data to write on
	the new block.
> 
> The uerf program does show something but no sector number
> ---- UNIT INFORMATION ----
> UNIT CLASS			DSA DISK
> UNIT TYPE			RA81
> CONTROLLER NO.		      17.
> UNIT NO.		      0.
> ERROR SYNDROME			DISK TRANSFER ERROR
> 
	You may be able to get more information if you use the
	"-o full" option on uerf(8).  It may depend on how much
	the Eagle really looks like an RA81.

> Should I give up and reformat or is there a way to write the bad sector
> into a file.

	Doing a scan with radisk(8) will reveal where the forced
	error is if there is one.  You can then use radisk(8) to
	clear the error and start tracking down the file the block
	belongs to.

	When running radisk(8) the system has be running single
	user and all the file systems possible should be unmounted.
> 
> cat
> cathy at larry.sal.wisc.edu
> Space Astronomy Lab-University of Wisconsin.


-- 
Alan Rollow				alan at nabeth.enet.dec.com