ADSM-L

Backup failures due to tape mount wait

1999-02-19 15:31:34
Subject: Backup failures due to tape mount wait
From: Thomas Denier <Thomas.Denier AT MAIL.TJU DOT EDU>
Date: Fri, 19 Feb 1999 15:31:34 -0500
My site is running an ADSM 3.1.2.1 server under MVS. All client backup files
are sent directly to a 3494 tape library with a Virtual Tape Server. Client
backup sessions consistently start with messages like the following:

Incremental backup of volume '/u06'
Normal File-->     109,568 /u06/oradata/DBA1/control1.ctl [Sent]
Normal File-->     109,568 /u06/oradata/DBA1/control2.ctl [Sent]
Normal File-->  20,987,904 /u06/oradata/DBA1/rbs01.dbf  ** Unsuccessful **
ANS1114I Waiting for mount of offline media.
Retry # 1  Normal File-->     109,568 /u06/oradata/DBA1/control1.ctl [Sent]
Retry # 1  Normal File-->     109,568 /u06/oradata/DBA1/control2.ctl [Sent]
Retry # 1  Normal File-->  20,987,904 /u06/oradata/DBA1/rbs01.dbf [Sent]

Having this happen at the start of a backup is relatively benign. I have also
seen a much nastier variant of the problem. ADSM started putting a 300
megabyte file on a tape that already had some data on it. The tape filled up
partway through the file. At that point ADSM reported that the backup was
unsuccessful, waited for an empty tape to be mounted, and retried the backup,
started at the begining of the file. All of the time and machine resources
used in transfering the part of the file that fit on the first tape were
simply thrown away. I suspect that a client file too big to fit on a single
file would create a still nastier problem: the backup would fail once when it
hit the end of a partially full tape, and then fail repeatedly as it hit the
end of each of a series of initially empty tapes. Since the Virtual Tape
Server simulates 3490 media with a capacity of 800 megabytes per volume, this
scenario is a real possibility at my site.

A search of the ADSM archive turned up some messages from people who appeared
to have a related but more serious problem, in which messages like the ones
quoted above were accompanied by some sort of internal server error. I did not
find any messages matching my symptoms exactly. Has anyone seen the variant of
the problem described above?
<Prev in Thread] Current Thread [Next in Thread>
  • Backup failures due to tape mount wait, Thomas Denier <=