Some notes on Dale's image-CD's-to-disk-to-ADSM process.

Written in perl, Dale called it save_cds.pl, which I put in
/dfs/.rw/admin/bin, giving all root_principals R/W access.

To recreate the list of US CD's,
    /jukeman/cdadm survey -s d=s2jb1.dev +p s:p | grep ^[0-9] > /s2jb1.us


To (re)start the program, on each image server,
    rm /var/spool/logs/*logs (or rename them to something else)

Then on 141, I
    cd /images1
    rm nohup.out  (or rename it to something else)
    nohup /dfs/.rw/admin/bin/save_cds.pl /s1jb0.us 1 &
so that the nohup.out file is in /images1/nohup.out, and
    cd /images2
    rm nohup.out  (or rename it to something else)
    nohup /dfs/.rw/admin/bin/save_cds.pl /s1jb1.us 2 &
so that the nohup.out file is in /images2/nohup.out,

On 142, it was
    cd /images1
    rm nohup.out  (or rename it to something else)
    nohup /dfs/.rw/admin/bin/save_cds.pl /s2jb0.us 1 &
and
    cd /images2
    rm nohup.out  (or rename it to something else)
    nohup /dfs/.rw/admin/bin/save_cds.pl /s2jb1.us 2 &
etc.


==============================================================

 6/14 14:30 - Started 2 processes on all servers, with the
              exception of only 1 process on 145 & 146 due to
              lack of space for /images2 file system.  This
              meant we had 11 processes total, running.
              Also, the 1, supposedly US CD on 147's s7jb1,
              was not US data.

 6/15 19:00 - There were various problems with the script,
              such as the replaying of the log file, and the
              creation of the patinfo file, so I stopped
              everything and restarted all 11 runs.
              After running for about 24 hours, we put 500 GB
              into ADSM (about 17% of capacity) and
              processed 235/3389 US CDs (about 7%).
              This extrapolates into 5 days to fill ADSM
              tapes to 85%, and 3 batches of tapes to finish
              the US CDs.

 6/19 13:45 -