UDA50 Problems on VAX 11/780 under 4.1 BSD

[uno at RAND-UNIX.ARPA]

We have been experiencing problems with an RA81 (and now an RA60) dropping
offline on VAX 11/780s under 4.1 BSD.  We get series of messages like:

	uda0: hard error, disk transfer error, unit 0, grp 0x0, hdr 0x24
	uda0: hard error, SDI error, unit 0, event 0353, hdr 0x0
		. . .
	uda: unit 0 online, size=400176

for each disk that drops offline.  The "SDI error" message is repeated six
times.  The disks drop ready (evinced by drive interface lights going out)
and eventually come back to life.

Our UDA50s contain boards M7485 and M7486 (which supposedly fix some
problems).

We are running a version of uda.c at revision 4.3 dated 82/03/14.

We have several systems configured as follows:

	VAX 11/780 with 1 RA81 and an RM05
        VAX 11/750 with 1 RA81, 2 RA60s (newly added on same UDA50) and an
		RM05
	VAX 11/780 with 1 RA60 and 1 RA81 on same UDA50

The first system has experience this error about once every 2.6 days over
the last year.  We have tolerated this.

The second system (VAX 11/750) doesn't seem to have had any (noticeable)
problems with this error.

The third system, which we put into production tonight experience the
problem on both the RA60 and RA81 simultaneously seven times over a 2.5 hour
period this evening!!!

Has anyone had any success with fixing the problem under 4.1?  Is there some
hardy soul out there running the same set of UDA50 boards who has overcome
this problem via driver mods or ???

Previous messages mentioning mysterious hardware munging that DEC might
perform or unnumbered ECs have not given us anything solid to pursue.

				Thanks,
				Stan Uno
				The Rand Corporation
				uno at rand-unix  (ARPA)
				decvax!randvax!uno  (UUCP)