[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Help! Big time disk problems
Help! Does anyone out there have experience with the LMFS
'remove partition' operation? I'd like to hear from you.
Here's my situation ....
I am getting disk errors ("%DISK-ERROR-SEARCH") on one of my machines.
The machine is a Symbolics 3670 running Genera 7.2. The errors only happen
on a single surface on one of the three disks on the machine. Symbolics
support tells me that there probably is a problem with the controller or
the head for that surface. Now, before I can stop using the disk and get
the thing serviced, I have to remove two LMFS partitions from that disk.
I have been clearing the decks for action and I now have enough space in
other LMFS partitions on other disks to hold the contents of the partitions
on the bad disk. I know that there are still damaged files on the bad disk
and, worse, some of the files have damaged headers so they cannot be deleted
(at least by conventional means).
Now, the questions ...
1) What will the 'remove partition' operation do in the face of disk errors?
2) Is there a way to delete the files with damaged headers?
3) I have already tried to run 'remove partition' in a mode where it susposed
to clear the partitions but not remove them. Curiously, it did nothing.
I have found the code responsible for this (lmfs:fsmaint-remove-partitions),
and it has a clause that specifically aborts when no partitions are picked
to be removed. I could easily hack around this clause, but perhaps its
there for some important reason I don't understand.
4) Pending a final solution to my problem, I would like to rig the LMFS so
that it only wrote files to the good partitions. Has anyone done this
and lived to tell about it? I think I have figured out a way to do this
-- it looks like there is a flag I can set in the struct's that represent
partitions -- but I am wary of unforseen consequences.
5) Would I be better off backing up everything twice over, nuking my filesystem
and starting from scratch? Say it ain't so!
6) Is there any way I could lock out the bad parts of the disk and continue
to use the rest of it?
Thanks in advance
- Steve Olson
MIT Lincoln Laboratory
olson@juliet.ll.mit.edu