[comp.os.vms] Volume set strangeness: Can anyone explain this?

hamm@BIOVAX.RUTGERS.EDU (01/18/88)

Here's one of those UFO's again.  I apologize for the length.

I'm setting up a volume set for the first time on my VMS system, and
was doing some experiments last night to see how it worked before I
"go public" with it.  Having initialized the disks (two RA-81s) and
done the initial MOUNT/BIND, I then was trying to make sure I understood
how to mount and dismount the packs properly, i.e., whether it made a
difference what order I mounted them, whether or not to use a single
MOUNT command, etc.

In doing this, I got into a state I've never seen before or since, and
can't reproduce.  I'm hoping someone else can explain what silly error I
made, so I can be sure it won't happen again.  The state was this:  I 
mounted the volumes, one after the other, like this, as nearly as I 
can remember:

        $ MOUNT DUB1: "Users.1" $users
        $ MOUNT DUB2: "Users.2"

I did some tests to see what happened if I created files on $users, and
was dismayed to see them all appear on DUB1:.  I then did SHOW DEVICE/FULL
on the packs, and now comes the strange part.  Normally, I'd see this
at the bottom of the SHOW output for each pack:

  Volume status:  subject to mount verification, write-through caching enabled.
  Members of this volume set are DUB1: (rvn 1), DUB2: (rvn 2).

But *this* time, I got the following for DUB1:

  Volume status:  subject to mount verification, write-through caching enabled.
  Members of this volume set are DUB1: (rvn 1).

and this for DUB2:

  Volume status:  subject to mount verification, write-through caching enabled.
  Members of this volume set are DUB2: (rvn 1).

I'm not *absolutely* certain that the rvn was 1 for DUB2:, but I think it was,
because I remember being surprised that the "relative volume number" displayed
above this in the SHOW output was correct for each pack, i.e., 1 for DUB1: and
2 for DUB2:.  F$GETDVI also showed the correct rvn for each pack.

At this point, it was late, and I thought I'd best get some sleep and have
another look this morning (and probably re-init the packs, etc.)  This
morning I came in, did

        $ mount DUB1:,DUB2: "Users.1","Users.2" $users

and everything worked perfectly.  I thought perhaps I mounted them last night
in the wrong order, so I dismounted them again, and tried:

        $ mount DUB2: "Users.2"
        $ mount DUB1: "Users.1" $users

Everything fine again.

The only other possibly relevant detail I can remember is that the
first time I tried mounting them as a volume set, I screwed up one of the
volume id's, so that an operator request was generated to mount the correct
volume.  But then I canceled that, the operator request was cancelled, and
I went on from there.

At this point, everything works, but I have an insecure feeling.  Can anyone
explain this behavior?  

Thanks,

Greg
------------------------------------------------------------------------------
Gregory H. Hamm                           || Phone:  (201)932-4864
Director, Molecular Biology Computing Lab ||  
Waksman Institute/CABM                    || BITNET: hamm@biovax
P.O. Box 759, Rutgers University          || ARPA:   hamm@biovax.rutgers.edu
Piscataway, NJ 08854 * USA                ||
------------------------------------------------------------------------------
------