ADSM-L

Odd ADSM/3494/3590 behavior

1997-05-20 14:18:28
Subject: Odd ADSM/3494/3590 behavior
From: "Kent L. Johnson" <johnsk6 AT RPI DOT EDU>
Date: Tue, 20 May 1997 14:18:28 -0400
We are running ADSM 2.1.0.12 with an ethernet-connected 3494 and 3 3590 SCSI
tape drives on AIX v4.1.

This morning, I started a 'move data' command on a 3590 volume.  The move
data started ok, and 3590 volumes were mounted.  Later, I entered a 'query
mount' command, to see which volumes were in use, and the system reported
that there were no tapes mounted.  I also executed a 'show mp', and this also
reported no tapes mounted.

The 'move data' command was still active, and the 'query proc' data showed
that data was still being moved.  So, I went to see if the volumes were in
the drives.  They were, and the 'move data' command completed.  The
"dismounting" messages were displayed for the volumes used in the 'move data'
command.

I tried it again with another volume, and the same symptom occurred.

The following are my observations from the activity log and the AIX errpt
output.

1) During both 'move data' commands, the error ANR8776W was logged.  The
error occured with the same volume for both move data commands.  The message
for this error is the following:

   Volume 000043 in drive RMT2 (/dev/rmt2) contains lost VCR data;
   performance may be degraded.

I believe that the 'query mount' lost track of the mounted volumes after this
error.

2) Coincident with each of the occurrances of the ANR8776W error, the
following entry occurs in the AIX errpt log.

0F78A011   0520133597 T H rmt2           RECOVERY LOGIC INITIATED BY DEVICE

3) The second 'move data' command dismounted the volume with the "bad VCR
data" and immediately after the volume dismount (ANR8468I) message, the
following error was logged.

ANR9999D pkshmem.c(564): Warning: attempt to free invalid shared
memory block: Address=302242C0, HdrP=302242B8, HdrP->next=30224700.

So, this seems like a bug to me.  It also raises some questions in my mind.

1) What is "bad VCR data"?  Is there a definition and description of VCR data
anywhere in the IBM documentation?
2) If volume 000043 has "bad VCR data" and is returned to scratch, will it
have "bad VCR data" the next time it is used?
3) Is there any way to fix "bad VCR data"?

I think I'll call this in to IBM ADSM support.

Kent

--
Kent Johnson                        Internet: johnsk6 AT rpi DOT edu
Kent Johnson                        Internet: johnsk6 AT rpi DOT edu
Unix Systems Programmer (VCC 323)      Phone: (518) 276-8175
Rensselaer Polytechnic Institute         Fax: (518) 276-2809
<Prev in Thread] Current Thread [Next in Thread>