The <A HREF="calendar.htm">calendar</A> indicates which steps are 
to be done on which days.

<PRE>
Date: Mon, 12 Aug 1996 16:24:15 -0700 (PDT)
From: Steve Jones <noyd@u.washington.edu>
Subject: Digital Unix 4.0 upgrade.

Schedule to upgrade Digital Unix 3.2C hosts to Digital Unix 4.0

0.  Upgrade saul8.                                                     Complete
    . Compose procedures for upgrades.
      . Update installation.   (Takes too long, not smooth.)
      . Full installation.     (Quicker and smooth.)
      . Clone installation.    (Fastest)

1.  Upgrade saulfs0? 3000/700 servers in saul cluster.
    Prepare
    .   Install ftp software on saulfs01.u.
    .   Ask netops to switch CNAME for ftp.u to saulfs01.u.
    .   Wait until ftp access to saulfs02 has subsided and
        name service has fully switched over.
    .   Take down saulfs02.
    .   Upgrade saulfs02.
    .   Announce down time for saulfs01.
        (No email delivery or access.)

    Downtime
    .   Disable pine, elm and mail on saul.
    .   Take down saulfs01.
    .   Replace saulfs01 with saulfs02 and reconfigure saulfs02
        to be saulfs01.
    .   Re-enable pine, elm and mail on saul.

    Afterwards
    .   Upgrade saulfs02 again.
    .   Have netops switch ftp.u back to saulfs02.

2.  Upgrade the Becker Cluster.  (Long downtime, upgraded as a unit.)
    Prepare
    .  Announce Becker to be down all day.
       (Expected downtime is 12 hours.)
    .  Add 3pack of 2GB disks, and move /bk01 and /chkout to new drives.
    .  Remove 3 2GB drives replaced by 3pack.

    Downtime
    .  Care must be taken to avoid running cron, sendmail, inetd
       during upgrade.
    .  Upgrade becker1.
    .  Upgrade becker2.
    .  Restart cron, sendmail, inetd.
    .  Make new OS announcement.


3. Upgrade the Saul cluster.  (After saulfs upgrade)
    Prepare
    .  Add chkout disk to saulfs01,
       and begin migration of chkout reservations to new disk.
    .  Announce downtime for saul.
    .  Notify "cron & at" users about extended downtime for
       cron and at.
    .  Upgrade first half of saul. (1,2,4)
      .  Save cron files for reinstallation.
      .  Upgrade saul1,2,4, disabling loadr, inetd.
    Downtime
    .  Shutdown cluster and announce OS upgrade
      .  Shutdown saul3,5,6            (Saul is down)
      .  Enable loadr on saul1,2,4.    (Saul is up)
      .  Readd cron & at files.
      .  Add announcement of upgrade.
    .  Upgrade second half.
       .  Upgrade saul3,5,6, and make available as upgrades are 
          completed.

4.  Upgrade the inpho system.
    . Announce downtime.
    . Upgrade system, avoiding running inetd,sendmail until
      upgrade is complete.

5.  Upgrade walker.
    . Announce downtime.
    . Upgrade system.

6.  Upgrade jrr1.
    . Announce upgrade and temporary move of filesets.
    . Add disks to saulfs02.
    . Move /sx00,/sx01,/ux01 filesystems to 3GB drive on saulfs02.
    . Modify /tulsa/fstab.cfg for locations of copied filesets.
      Run bldfs to export filesets from saulfs02.
      Run bldfs to remount filesets from saulfs02 instead of jrr1.
    . Upgrade jrr1.
    . Copy filesets back to available jrr1 disks.
    . Restore /tulsa/fstab.cfg.
      Run bldfs to export filesets from jrr1.
      Run bldfs to remount filesets from jrr1 instead of saulfs02.
    . Remove disk from saulfs02.

7.  Upgrade weber.  (No downtime)
    Prepare
    . Take saul8, preserve system disk and add upgrade as weber3.
    Upgrade
    . Upgrade weber2.
    . Re-enable loadr
    Afterwards
    . Remove weber1 and upgrade to be saul8 again.

8.  Upgrade alcott.
    Prepare
    . Upgrade alcott.
    . Copy missing /usr/local files from another alcott.
    . Re-enable loadr.

9.  Upgrade neruda, last DU32C system.
    Prepare
    . Backup httpd files to spare disk somewheres.
    Downtime
    . Upgrade neruda.
    Afterwards
    . Restore httpd files.
</PRE>