Re: vfs_mountroot panic on tru64 v5.1A

From: Tony (tony.bolton_at_geest.co.uk)
Date: 05/08/03

  • Next message: Bob: "Windows Single Sign On (SSO) insecure?"
    Date: 8 May 2003 13:57:34 -0700
    
    

    Hi again,

    All sorted now. Basically our root partition is mirrored, and there
    was a problem with the hardware database between the two disks. We
    had to rebuild the root domain on one of the disks, then trick LSM
    into re-syncing the opposite way ( as the ok device was the 2nd drive
    ). This worked, and thankfully it's all up again now.

    However, even though I did read the patch notes, I missed a section on
    - wait for it - SDLT 160/320 drives - whereby scsi termination on
    these drives is not handled correctly. This can be sorted by adding
    lines to the ddr database apparently. I'll post details on this when
    I go back to work tomorrow.

    Hope this all helps someone else,

    Cheers,

    Tony

    tony.bolton@geest.co.uk (Tony) wrote in message news:<fd3d6c2e.0305070746.31d2d826@posting.google.com>...
    > Hi,
    >
    > First of all, my apologies if this sounds a little vague.
    > Unfortunately the error I am getting is on the actual server console,
    > which is in a secure location (ie. no internet etc.)
    >
    > We're running tru64 v5.1A, and I have just completed installation of
    > patch kit 4. On Wednesday last week, we had a new SDLT drive added to
    > the system, along with a new SCSI card to run the drive off. The
    > following 2 days, I experienced intermittent problems using the drive
    > - particularly with the device name 'disappearing' from within the
    > /dev/tape or /dev/ntape directory. I put this down to a minor issue,
    > and on rebooting the box it came back and worked fine.
    >
    > However, I installed the patch kit on Tuesday - something I've done
    > many times in the past. I have never had the following happen though!
    > When I tried to reboot, bcheckrc hung, stating that there was a
    > problem with the database on line 294 (which had the string /dev/tape
    > within the error message - but it shot past so quickly I didn't have
    > time to make a note). It would then stop in single user mode, with
    > root mounted as ready only.
    >
    > On contacting Compaq, they decided that the hardware database was the
    > main cause, so I was asked to delete the relevant files and reboot in
    > order to rebuild the database. Great, I thought.
    >
    > It booted back fine - no problems, and I could see everything again
    > (ie. domains etc).
    >
    > I thought at this stage 'time to make a backup', so decided to use the
    > new SDLT. This time, the SDLT drive 'remapped' itself to tape1 (it
    > was tape 2 before - when it worked that is). So, I started to backup
    > / using vdump.
    >
    > This is when the major problem kicked in - the machine hung, blue
    > screened, then said there was a panic on vfs_mountroot. It didn't
    > give any specifics, and just went back to the machine prompt (>>>).
    > No matter what I do - single user, different vmunix etc. - it still
    > panics.
    >
    > My initial thoughts are that it is almost definately the new SCSI
    > device. The fact that I patched the system is probably just
    > coincidence, and so far Compaq seem to agree that it certainly looks
    > hardware orientated.
    >
    > My question is : has anyone else had the same problem in the past -
    > and if so what was done to resolve the issue? We running on a DEC
    > ALPHA 4100, with 2 HSZ50 controllers in a storageworks cabinet. It
    > might also be worth mentioning it's a dual processor machine.
    >
    > Another interesting point is that I can boot from a v5.1 OS Cd ok, and
    > get to the shell window. However, I tried to scan all devices using
    > 'scu scan edt', but that hung and threw me out to a blue screen all
    > over again.
    >
    > Compaq are looking into the problem as I write this, but I want to
    > cover all avenues and see if anyone else has had similar problems,
    > particularly with SDLT units.
    >
    > Cheers,
    >
    > Tony
    > Ps. We also have an additional tape drive on the system, but on
    > totally seperate card - it's a TZ88.....


  • Next message: Bob: "Windows Single Sign On (SSO) insecure?"

    Relevant Pages

    • Re: HEADSUP: geom_vinum committed
      ... it may have been in your case that the root volume failed to mount ... two Vinum test drives. ... plex name usr.p1 org concat vol usr ...
      (freebsd-current)
    • Re: Hard drive weirdness.
      ... then I got a grub Error 17: ... I changed root from to ... Now I tried removing both sata drives. ... My problem is that raid ...
      (Ubuntu)
    • Re: user failures
      ... I had my boot drives partition table zeroed out last night by something ... to the filter screen, but when I click the apply button, anything I've ... I've nuked /root/kmailrc, and then restarted kmail as root, with no ...
      (Fedora)
    • Re: Peer to Peer sharing
      ... Windows XP by default created administrative shares for the root folders of the fixed drives of the computer. ... You can create your own shares to share your drives from the root of the drive, but Microsoft highly recommends that you share only the folders that you need to, rather than the entire drive. ... Windows Vista displays the properties for the drive with the Sharing tab selected. ...
      (microsoft.public.windows.vista.networking_sharing)
    • Re: Grub Manual ... Solved
      ... definition of a root directory has been tested and it is right. ... It is doable but the stuff in info grub is not ... defines the drive and partition number on that drive where grub can find its ... that ones broken too basis, currently running 2.6.23 (no sata drives here, so ...
      (Fedora)