HD Problem...Is it possible to fix HD to be able to mount it?

HD Problem...Is it possible to fix HD to be able to mount it?

Post by Boyd Lynn Gerb » Sat, 16 Sep 1995 04:00:00



        Sorry to the trouble the net, but I know it has to be possible to
solve my problem.  SCO support tells me it is not possible, but I think
they mean it is going to take a lot of work.  AN there is not way to know
how much I may be able to recover.  I know there has to be a way.  I lost
my primary root hard disk because of a power related failure that got
through the UPS.  The drive was toast and the power failure also destroyed
my tape backup unit.  I know I should use the backups, but the two units
I have borrowed from others are not able to read my tapes.  I would like
to be able to fixed the disks and mount them.  I am using ODT 3.0 with a
EAFS file system.

        I have placed a new drive in the system and it is now working.  In
the past I could just add the various devices by hand with mknod and mount
them.  I am able to accesss the entire hard disks with dd and the nodes I
created with mknod.   When I try to mount them I get unknown file system
type and then

mount: cannot stat '/dev/disk1' where disk1 is the node I made.  

        I am able to dd the contents of the entire disk.  There has to
be a way to fix the information!!  I need to mount them.   They are 1.8GB
and 400MB.  This is and ODT 3.0 system.  What can I do to be able to mount
these drives.  I know this may take a lot of work.  Could someone point me
to where to look to get the information to fix and mount these drives.  
        My only other back up was lost in a Release 5 upgrade that went
bad.  I am willing to spend what ever time I need to to get this working  
any ideas.

Thanks,

Boyd


 
 
 

HD Problem...Is it possible to fix HD to be able to mount it?

Post by Scott A. Marti » Sun, 17 Sep 1995 04:00:00



(deletions)
>...  I am able to accesss the entire hard disks with dd and the nodes I
>created with mknod.   When I try to mount them I get unknown file system
>type and then

>mount: cannot stat '/dev/disk1' where disk1 is the node I made.  

>    I am able to dd the contents of the entire disk.  There has to
>be a way to fix the information!!  I need to mount them.  
(deletions)
>Thanks,

>Boyd



Boyd,

Sounds like you know your way around pretty well so this
note may in vain -- when you are attempting to mount these
disks, are you using an inode pointing to the entire drive,
the active partition on that drive, or the filesystem(s) on
the drive?  I've sometimes forgotten to distinguish between
these when "pulling in" additional drives.

If these drives have valid structures intact, you could run
"mkdev hd" to bring them on-line.  Just "quit" your way
through the various steps (e.g., fdisk, divvy) if they see
valid info for the drive.  If you can dd your way through,
maybe the partition and divvy info are intact.

If you don't gain access via this method, call me. I know
you can run fdisk again on the drive without data loss, but
I would want to try the divvy re-creation in my lab before
you attempt it (i.e., run divvy, set "prevent" on all
divisions and save).

Let me know if I can help.

Scott

Scott A. Martin
NOESIS Open Systems
Seattle, WA


 
 
 

HD Problem...Is it possible to fix HD to be able to mount it?

Post by Timothy D. Kue » Mon, 18 Sep 1995 04:00:00




Quote:>Sounds like you know your way around pretty well so this
>note may in vain -- when you are attempting to mount these
>disks, are you using an inode pointing to the entire drive,
>the active partition on that drive, or the filesystem(s) on
>the drive?  I've sometimes forgotten to distinguish between
>these when "pulling in" additional drives.

The procedure I used last time I did this with (working) hds was:

mkdev hd
reboot
mkdev hd
quit fdisk and anything else which wants to
    init the drive, but name the fs-s in divvy
mkdev fs for the various fs on the new drive
mount the fs's to bring stuff in

Tim Kuehn

 
 
 

HD Problem...Is it possible to fix HD to be able to mount it?

Post by 80 » Mon, 18 Sep 1995 04:00:00


I got my hard disk back up and working.  I used the IT article on dparam
and a duplicate identicle HD for the SCSI after configuring it.  I was
able to brut force find the bytes that I had to fix.  I did the same with
the other except I dd the entire disk to the new scsi created a new file
system on it and found the bytes that I need to fix.  The power failure
over wrote the dparam values on the disk and part of the bad block area.

Thanks for all the suggestions.

Boyd


 
 
 

1. Why am I plagued with these HD problems?

  I had installed the SLS1.0 release soon after it came out, but was never
able to keep it up for long. I installed series a, b, c, and x on an
ext2fs. After some time (from a few minutes to several days) I would get
either an HD timeout and controller reset message or something of the
following form:

  general protection: 0000
  EIP: 0008: 00118700
  EFLAGS: 00010282
  fs: 0017
  base: 00000000, limit: 0000
  Pid: 1148, process nr: 12
  66 39 48 0c 75 2e 39 58 08 75
  xargs: /bin/rm: terminates by signal 11

These problems then quickly gain momentum until I am kicked out and unable
to log on. My first thought was that this was because the e2fs disagreed
with my EISA system. After trying several more times in vain to install the
system, I remembered that I had moved the HD controller card to a different
slot, and never reconfigured the EISA part of the CMOS (I don't really
understand how this works, any insight appreciated).

  I then reconfigured the card, installed the system again, and lo and
behold, it worked! After happily linuxing away for several weeks, I recently
copied the latest SLS release from a friend, figuring that with all those
great fixes and new libraries it could only get better. However, soon after
installing it I encountered the same degradation. So I tried to go
back to the old system. Still no good. What gives? What did I do to get
lucky the time that it did work? And what exactly is the problem?
Is my system just not compatible? Is this an EISA specific problem?

  Somebody please save me before I turn back to *GASP* !CHOKE! MESS-DOS.

My controller is a DTC2290 connected to a QuantumLPS240. The system is a
486 DX50 with 8MB.

If Someone even has a clue or might be able to help if I give more detail
please let me know.

Jason LaPierre
Department of Cognitive Science
University of Rochester

P.S. The SLS release was fantastic after some tweaking, and when I wasn't
having the aforementioned problems. And the new installation script is
great though I didn't get much chance to use the system 8)

2. : Mounting AppleTalk server on Unix machine

3. It`s IMPORTANT!!!

4. redirect DISPLAY from SUN-box to Linux-box

5. Please Help....it′s a very small thing!!!!.PLEASE

6. remote X (newbie)

7. Help me !!!!!! I have no Idea it′s crucial

8. tcpd problem?

9. Am I the only person around with a 40 gig HD?

10. It`s IMPORTANT!!!

11. Hint for WD HD users: Possible HD failures

12. Green HD (HD spindown simulator)....Is it possible?

13. Cannot mount HD with AHA29 ad Segate HD