hutching@utah-cs.UUCP (Brad Hutchings) (04/07/88)
I have a puzzle for all you apollo netlanders:
I am working with the following configuration:
DSP-80 1MB (Paged off of a DN3000 running SR9.7)
Danford 300MB disk (MSS-300, SN106, SN176)
There are two disks, the problems listed here are for SN106 although
they function similarly.
I am having the following problem:
When I attempt to invol the disk, I get the following message:
Writing logical volume 1.
Formatting... % complete:
20
40
60
80
block=4490B, bat_daddr=20E6E, index=72(d), bit=0(d)
blk @ 8C8000
bat consistency error
Run aborted.
When I originally received this disk, it had been invol'd and was mountable.
Since I got this disk used, I decided to check it out by running fbs.
fbs ran fine (about 20 passes) and found 9 or 10 badspots. However,
I cannot invol the disk. I have tried invol revisions 9.4
and 9.7 with about the same results. 9.7 results below, 9.4 above.
I tried these different versions of invol in both a crp'd process as well
as at the md prompt.
Writing logical volume 1.
Formatting... % complete:
20
40
60
80
Writing all blocks... % complete:
20
40
60
80
Reading all blocks... % complete:
20
40
60
80
block=448DB, bat_daddr=20E6E, index=70(d), bit=16(d)
blk @ 8A8000
bat consistency error
Run aborted.
The dsp-80 is paged off of a DN3000 running SR9.7 software.
Also, the bad blocks list is intact and matches the flaw list
that is attached to the side of the disk.
I have tried invoking invol with different logical volume sizes, etc.,
mainly to speed up the formatting process (since it doesn't work anyway) and
get virtually the same results each time. The numbers in the error message
vary somewhat, for instance, for small logical volumes (1000-10000 blocks),
the block=<number> and bat_daddr=<number> are smaller (probably since the
logical volumes are located in the lower block numbers of the disk) and I
haven't yet detected any pattern in the index and bit numbers. I have
tried adding the block=<number> and bat_daddr=<number> to the bad block
list and re-running invol. Invol still fails but the block numbers
that were added to the bad blocks list no longer appear in the error
message; they are replaced by new block numbers. I tried this 5 or 10
times with the same results each time (hey, even I can see trends :-) ).
One other thing that I have noticed is no matter what the block numbers
are in the error message (large or small logical volumes) the
number in blk @ <number> is always a 6 digit hex number that starts with 8
usually followed by an 8 or a A or a C (generally) and
then followed by 4 zeros.
I have two identical disks that give about the same results. Since the
disk came in a mountable state, I am more or less assuming that the disk
is probably OK (boy, am I waffling or what :-) ). The disk controller card
that is resident in the DSP-80 appears to have been manufactured by intel.
I have run the SAX utility on the DSP-80 and it checks OK.
I think that's about it.
I would appreciate any suggestions or pointers.
Brad L. Hutchings