[Prev][Next][Index][Thread]

SUMMARY: System crashed and my disk went with it

I previously wrote:

> Sometime during the night one of our DEC 3000/400 workstations
> went south. I strongly suspect a faulty tape drive causing the
> problem as we have seen this before. Anyway, the boot disk is
> hosed to the point where I get the following error:
> AUDIT_BOOT_STARTS ...
> AUDIT_CHECKSUM_GOOD
> AUDIT_LOAD_BEGINS
> AUDIT_LOAD_DONE
> can't open osf_boot
> 
> I am not exactly sure what this means but here is what I tried to 
> do to fix the problem. First, I booted from the CD-ROM and
> went into the system management shell. I then ran fsck on the
> disk and it fixed A LOT of errors and all of the partitions on
> the disk EXCEPT the a partition cleaned up successfully. When 
> fsck'ing /dev/rrz14a I get the following:
> Phase 1 - Check Blocks and Sizes
> Phase 2 - Check Pathnames
> ROOT INODE UNALLOCATED
> ALLOCATE?  [yn] y
> cannot find inode 2
> 
> So, given these two errors I thought that maybe the bootblocks
> were gone/hosed. So, I tried writing a new disklabel:
> disklabel -w -r /dev/rrz14a rzxx. (I inserted rzxx into the
> disktab file.) Then, I tried rebooting. But, I still get:
> can't open osf_boot
> 
> Does anyone have some ideas that will get this system disk back
> up short of me having to reinstall the whole operating system?
> Thanks in advance for any help you man have to offer.
> -- 

It turns out that osf_boot is a file in /  (duh!, forgot that)
that was missing. I managed to save all of the partions on the
disk except the root partition. fsck gave me a missing inode
error when trying to repair it and puked. In fact after mounting
/dev/rz14a (root file system) and trying to cd to it, I would get:
/mnt: not a directory (or something like that) So, I couldn't seem
to get osf_boot to copy to the old root partition to save my...well...
you know. 
So, it seems that 2.0 has many more system managemet utils than 
did 1.2 or 1.3 included with it. One of which was tar. Yipee! So, 
I tarred the rest of the partions and re-installed from scratch. :( 

What did I learn?:
1. Even tho I use tar as my means of backup, I will now regularly (weekly?)
   make dumps of the file systems too. Especially root!
2. Don't keep putting off buying those extra backup tapes and head cleaners.
3. If you suspect a flaky tape drive take it OFFLINE before it hoses
   a filesystem.
4. Keep hard-drives on one SCSI bus and cd-roms and tape drives on the
   other. Any thoughts on this? This isnt the first TLZ06 drive to
   flake out on me and hose a disk.
5. That at least three helpful DECies work weekends:
   sarasin@xxxxxxxxxxxxxxxxxxxx
   urban@xxxxxxxxxxx
   murphy@xxxxxxxxxxxxxxxxxx
6. One can drink his/her weight in Mountain Dew in 24 hours.
7. That I learn more about unix from disasters than anything else.

Thanks for the help.
-- 
****************************************************************************
* Ed Jones                            |     ejones@xxxxxxxxxxxxxxxxxxxx    *
* Department of Psychology            |     ejones@xxxxxxxxxxxxxxxxx       *
* Signal Detection Lab                |     ejones@xxx  (Bitnet)           *
* Wright State University             |                                    *
****************************************************************************