Hi Andy,

Wouldn't this whole exercise be easier if a filespace argument of nothing
(like incremental) matched all the local filespaces for the client
node......

Can we raise a design change request for this.......???? And how about
having archive use the include/exclude file as well ...

regards Tonyr


For AIX, I use the following scripts (stupid) script..

#!/bin/ksh
#
# ADSM Archive a node to a monthly milestone queue.
#
# Date:   21/1/99
#
# Desc:   This script is intended to be used from any node that requires a
#      milestone archive. The script will select all local mounted
filespaces and
#      archive the files.
#
# Input:  none
# Output: Success or Failure Code
#
#
# Revisions:
#
#

# Offsite mail userid
USERID="root"


ARCHMC=milestone_month

# Determine the filesystems to be archived
FILESPACES=`df | grep \/dev\/ | awk '{print$7}'`

echo $FILESPACES
echo $ARCHMC

ARGLIST=""

# Build Filespace arg list
for i in $FILESPACES
do
if [ $i = "/" ]; then
     ARGLIST="\"/*\""
else
     ARGLIST="$ARGLIST \"$i/*\""
fi
done

# Execute the dsmc archive command

echo "dsmc archive archmc=$ARCHMC -subdir=yes $ARGLIST"
dsmc archive "-archmc=$ARCHMC" "-subdir=yes" $ARGLIST




Andy Raibeck <storman AT US.IBM DOT COM> on 08/12/99 02:11:56

Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>

To:   ADSM-L AT VM.MARIST DOT EDU
cc:    (bcc: Tony Rynan/Australia/IBM)
Subject:  Re: archive (gotta get it all)




Hello Dave,

Couldn't you run something like this:

   SELECT NODE_NAME,FILESPACE_NAME,FILESPACE_TYPE
          FROM FILESPACES

Then create a script to read the output and create
individual DEFINE SCHEDULE and DEFINE ASSOC commands
for each node and its filespaces? Example schedule
definition:

   DEFINE SCHEDULE domain Y2KARCHIVE_nodename
          DESC="Y2K ARCHIVE"
          ACTION=ARCHIVE OBJECTS="fs1\ fs2\ fs3\"
          OPTIONS="-SUBDIR=YES -ARCHMC=Y2KMC"
          STARTDATE=mm/dd/yyyy
          STARTTIME=hh:mm PERUNITS=ONETIME

   DEFINE ASSOCIATION domain Y2KARCHIVE_nodename
          nodename

The lower-case items would be filled in by the script.
You could use SELECT NODE_NAME,DOMAIN_NAME FROM NODES
to identify which nodes belong to which domains (unless
you have only one domain which makes things easier).

Some complications: If you have clients (especially UNIX)
that have a large number of filespaces, you might want to
create multiple schedules, putting a smaller numer of
filespaces into each schedule.

If you have any Windows clients where the filespace names
are still identified by the drive label (pre-3.1.0.5),
your script can probably determine this by checking the
first two characters of the filespace name for the "\\"
characters. If the first two characters are not
"\\", then you can use the "{LABEL_NAME}\" notation to
specify the filespace names in the OBJECTS parameter of
the schedule; otherwise you can use the "\\machname\c$"
notation to identify the filespaces.

There would almost certainly be little "gotchas" like
this to work out, but it should be do-able.

Regards,

Andy

Andy Raibeck
IBM Storage Systems Division
ADSM Client Development
e-mail: storman AT us.ibm DOT com
"The only dumb question is the one that goes unasked."

Well, I wish that I could set the record straight (whatever it is).  But, I
don't have the knowledge to determine the automated archive process from
the
server schedule!!  Help someone?

Dave Sanders
Sr. Technical Consultant
MassMutual / The Blue Chip Company
1295 State St, E060, Springfield, MA 01111
413-744-5095