Use zridebldr.com to build this.

This system lets you control almost any arbitrary SCSI disk
as a normal VMS disk. It uses io$_diagnose to read disk size
and read and write blocks. If it gets errors long enough it
will try to reset the SCSI bus it is on (using the device
name to infer that DKB4: for example would be on bus PKB).
If this fails it may act weird but generally will fail to yank
reset. Then I/O may eventually just fail. It mimics mount
verify this way...many retries with resets if they seem
to be needed.

Use:
Build, copy zrdriver.exe and zridehost.exe into somewhere
like a part of sys$system. Copy zridehost.cld there too.

$ sysman io conn zra0:/noada/driver=sys$system:zrdriver
$ sysman io conn zra1:/noada/driver=sys$system:zrdriver
 ... and so on for all ZR units you want. These are what 
	you mount.

Run a batch job or whatnot so that it includes the elements

$ set command sys$system:zridehost
$ zridehost zra0: dkb5:
	for setup of zra0: on physical device dkb5:. Note that
	it is perfectly ok to use gkb5: or whatnot instead; both
	dkdriver and gkdriver accept io$_diagnose.

	Note too you may need
	$ sysman io set exclude=(dkb0,dkb1,dkb2,dkb3,dkb4,dkb5)
	or something like it to avoid having those devices 
	autoconfigured if you don't want dkdriver configured
	for them. If you do that THEN you need e.g.
$ sysman io conn gkb0:/noada/driver=sys$gkdriver
	to define gkb0:
if you want that.

The dtdoany.com file is designed to run batch files as needed,
use:
$ submit/param=(zra0:,dkb5:) dtdoany

to get the effect of the above.
(I use submit/noprint)
Glenn Everhart


ps -
Be sure and put 
$dism zra0:

and so on
commands in syshutdwn.com in sys$manager. This will make the zr
disks dismount and will cause the zridehost image to exit, thus
freeing the "host" disks (e.g. dkb5: in the example).
