[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Lisp stopped itself (still)
Date: Tue, 17 Sep 1991 11:15 EDT
From: Scott_Busse%mindlink.bc.ca@Warbucks.AI.SRI.COM (Scott Busse)
(* Thanks to everyone who has helped me with diagnosing this problem.
(* One thing to note is that there's *no* way for me to get into Lisp
(* at all.
(* So far:
(* Scanning and using the tests.flod revealed some sticky bits in
(* one of the 2 MW boards. It went like so:
Test Simple main-memory
Testing memory board in slot 7. = #o7
Testing memory board in slot 8. = #o10
Testing memory board in slot 25 = #o31
At least 1 location has bit 22. stuck on
At least 1 location has bit 32. stuck on
(* and this last statement was repeated for bits 36,37,38,38,40,
2. locations have bit 22 stuck on; addr-ior 63124426, addr-and
6312046, xor(addr-adn, addr-ior) 4000
233. locations have bit 32 stuck on; addr-ior 63120777, addr-and
63120400, xor(addr-and, addr-ior) 377
(* this last statement was repeated for bits 36,37,38,39,40,41,42,
(* all with the same numbers ie. 233. locations...addr-ior #,
(* addr-and #, xor #, including the 377 at the end
Testing memory board in slot 26 = #o32
(* this last statement was repeated for slots 27-30, showed no
(* Then I performed a Test Main-memory
Testing memory board in slot 7.
Testing one location per page.
(* many numbers here, nothing conspicous *)
Testing powers of 2.
(* and this was reported for slot 10,31,32,33,34,35,36 with no
(* errors apparent. Is this testing some board other than the
(* 2 MW boards?
Sounds like one of your memory boards is bad.
Label for unit 0: 1635 cylinders, 15 heads, 24 pages/track,
Id = 0, Fast mode = 1, Root at 30(8)
Pack name: DE-00355
Testing disk unit 0
Error reading cylinder 1635 head 0 sector 0 on unit 0
Too many retries for unit 0. Last error was: Timed out waiting for
seek to cylinder 1635
(there appears to be 1635 cylinders)
Error while mounting unit 1: Select error on unit 1 during unit
(* This last line was repeated for units 2,3,4,5,6,7 since there's
(* only 1 drive I suppose.
(* It seems that that last cylinder should have been omitted at
(* format time.?
You can ignore that error, it always happens. For some reason, Test
Disks always tries to access one extra cylinder (maybe it's just
verifying the information in the label).
Testing every 177(8) locations: 0 1770 3760 5750 7740
Testing power of 2 locations
(* seems to be no problem here even with the bad mem board in.
(* So from the load microcode report...
Setting up 2MW board in slot 7
Card 1 being assigned to slot 30.
Card 2 being assigned to slot 29.
Card 3 being assigned to slot 28.
Setting up 2MW board in slot 8
Card 1 being assigned to slot 27.
Card 2 being assigned to slot 26.
Card 3 being assigned to slot 25. (* the bad guy!? *)
2 2MW memory boards found
(* I removed the slot 8 2MW board and redid the Simple
(* test main-memory, with no errors then reported.
(* Attempting the boot sequence still comes up with
(* the same error at the "Start" command, so I assume
(* the memory has nothing to do with the disk error.
Test Disks doesn't do an exhaustive test (it would take hours per disk),
so it won't detect individual bad blocks. It's mostly useful for
detecting head crashes or other problems that affect large portions of
the disk. It still sounds like you have a bad spot on your disk.
(* Since I have an IFS tape from the other 3650, I tried to Load
(* Microcode from CART:>3650-fpa-mic.mic but got an error
If you're talking about the IFS tape that is labeled with the serial
number of the disk, you can't load files from this tape. Most of the
files on this tape (except for the .FEP files and the root directory)
are empty placeholders.
Load Microcode (default is FEP0:>3650-fpa-mic.mic)cart:>3650-fpa-
(* I knew the file would be on the tape as I was able to read the
(* FEP root on the other 3650 and document its contents...
(* Anyway the error was...
Error while opening cart:>3650-fpa-mic.mic: Data error or bad
block not located
(* Should I assume this means the tape is bad?
(* So I loaded the microcode from the disk, as before, and then
(* tried to load world from this IFS tape.
Load world (default is FEP0:>genera-7-1.load)cart:>genera-7-1.load
Error: Unbelievable (129480788) number of sparse entries
These errors suggest to me that I was right above about the tape you're
trying to read. You should be using the "Genera 7.1 Distribution
Worlds" tape, not the IFS tape for a disk drive.
(* I realize this was a shot in the dark...I would probably
(* have to have the tape in perfect position for this...
(* So I went ahead and load worlded genera-7-1-5-combo.load,
(* Enabled IDS, and Started, and I noticed that the first time
(* the error message came up it was DISK-ERROR-ECC instead of
(* DISK-ERROR-SEARCH. The numbers were the same as when I did
(* continue, though it switched back to DISK-ERROR-SEARCH.
(* BTW, what do the <211><211><211> mean?
Those are TAB characters. The FEP is has *very* limited knowledge of
I assume that the only way out of this disk error is to get an
IFS tape from Symbolics for this machine. Is this tape able to
deal with the bad block (does it create a new base load file)?
Can I get a tape that is compatible with genera 7.1 still?
You should have received the IFS tape with the machine. However, if the
disk has developed a new bad block, it won't be listed on the IFS tape.
Finally, if you're going to try to reformat the only disk in the
machine, you'll also need a Breath-of-Life tape, which is used to
bootstrap a machine without a disk.