Article 174270 of comp.os.vms:

The scsi error bits are in the read/write err recovery page.
The page code is 01 (tho' reports may show 81 hex also;
bit 7 is another field.

Layout of that page is:

byte	bit7	6	5	4	3	2	1	0
0	PS	----------PAGE CODE (01)-------------------------
1	-----------------=PAGE length (often 0A)-----------------
2	AWRE	ARRE	TB	RC	EER	PER	DTE	DCR
3	-----------------READ RETRY COUNT------------------------
4	---------------CORRECTION SPAN---------------------------
5	-----------------HEAD OFFSET CNT-------------------------
6	-------------------DATA STROBE OFFSET CNT----------------
7	-----------------RESERVED--------------------------------
8	--------------WRITE RETRY CNT----------------------------
9	-------------------RESERVED------------------------------
10	(MSB)-------------RECOVERY TIME..........................
11	..............LIMIT--------------------------------------

AWRE: 0/1 - DON'T/DO RELOCATE badblks on write
ARRE: 0/1 - don't/do reloc bad blks on read
TB: 0/1	-   don't/do transfer unrecoverable blk
RC: 0/1 -   don't/do invent data to avoid delay
EER:0/1-    min err miscorr prob/use most expedient err recov
PER:0/1 -   don't/do report recovered errors
DTE:0/1 -   don't/do stop transfer on a recovered err
DCR:0/1 -   allow/disallow use of ECC error correction 


Hopefully this will help folks reset their drives. I've hacked around with
this in isolated cases but recommend people just set their drives (use
rzdisk if you must). Also note that VMS 7.1 tolerates a great deal
more variability in these bits than earlier drivers do...sticking with
Alpha VMS 1.5 is something I recommend against.
Glenn Everhart

Foregoing info is in the SCSI-2 standard, by the way. This just describes
the mode page.