Checklist for installing Digital Unix 4.0A.
  SJ 96/12/04

 Most actions in this checklist should be done on the console through
 Jedgar.  In the steps below, the commands will have several prompts:

      SU>        Run as super-user from any terminal.
      su>        Run as super-user from console.
      #          Run in single-user mode on console.
      >>>        Run in monitor mode on console.

 
0] Save files in OS area to be restored on new vanilla 4.0 system.
    su>  SDIR=/tulsa/transfer/`hostname`
    su>  mkdir $SDIR
    su>  cd /var/spool/cron
    su>  tar cf $SDIR/cron.tar .
    su>  cp /etc/sia/shadow_matrix.conf $SDIR
    su>  cd /usr/spool/mqueue
    su>  tar cf $SDIR/mqueue.tar .
    su>  cd /etc
    su>  tar cf $SDIR/etc.tar fstab fdmns routes hosts rc.config lipw fstore.cf

    For Alcott, which does not install enough products via Tulsa,
    we need to save many files in /usr/local which were installed
    by hand.
    su>  cd /usr/local
    su>  tar cf $SDIR/local.tar netscape bin lib etc home

  Check for any special products or symbolic links in the /usr/local
  filesystem which are not installed via tulsa.  

  Dismount any local filesystems which are exported to other systems.
    su>  cat /etc/exports
  Dismount filesystem on remote systems first.
    su-other> /sbin/umount ... 
  Then dismount on local system.
    su> /sbin/umount ...

1] Run "disklabel" to verify that the sizes of the partitions on the
   system disk are correct for "/", "/usr" and swap.
      We want 

      su> /usr/sbin/disklabel rz8

  a:   423000        0    unused        0     0         # (Cyl.    0 - 29)
  b:  1535800   423000    unused        0     0         # (Cyl.  300 - 1389*)
  c:  3933040        0    unused        0     0         # (Cyl.    0 - 2789*)
  d:        0        0    unused        0     0         # (Cyl.    0 - -1)
  e:        0        0    unused        0     0         # (Cyl.    0 - -1)
  f:        0        0    unused        0     0         # (Cyl.    0 - -1)
  g:  1974240  1958800    unused        0     0         # (Cyl. 1389*- 2789*)
  h:        0        0    unused        0     0         # (Cyl.    0 - -1)


1]  Shutdown to monitor.

    melville> setargus down ${HOST}

    SU>  shutdown -h +15 "Shutdown to upgrade OS to DEC OSF1 V4.A"

    Attach cdrom drive. (check for unit number conficts!)
    Set boot flags:

    >>>   set boot_osflags ""
    >>>   set server on

    Verify and note presence of CDROM.  You'll need to convert the
    device name into a special file name for the mount in [6].

    >>>  init
    >>>  show device
    >>>  show config
 
2]  Put Alpha Systems Firmware Update 3.8 CDROM in drive.
    >>>  boot dkb600
    Bootfile: 
    ? show
    ? update

    Power off computer for 15 seconds.
    Power on computer.

This next section may be skipped if you are doing a clone 
installation.  Ifso, go to step [17].

3]  Put Digital Unix 4.0A OS CDROM in drive.
    Boot.  
    >>> boot dkb600
      Select Custom install.
      Select "US" and then "Pacific" for timezone.
      Select "advfs" filesystem types for "/" and "/usr".
      Put "/usr" on partition "g".

      Install these subsets:
         21 24 26 27 32-34 39-47 51-54 63-64 84 86

      Follow instructions on rebooting the system for kernel build.
        >>>  set boot_osflags "A"
        >>>  set bootdef_dev DKB0
        >>>  boot

      Select these kernel build options:

        NTP V3 Kernel Phase Lock Loop (NTP_TIME)
        STREAMS pckt module (PCKT)
        X/Open Transport Interface (XTISO, TIMOD, TIRDWR)
        ISO 9660 Compact Disc File System (CDFS)

      The kernel will be built, installed and the system will reboot.
      If you are familiar with "ed", you can do the below changes
      when prompted to edit the configuration file.  Otherwise, you'll
      need to login, reconfigure the kernel and again reboot.

       Set maxusers to 256
       semmsl          40
       kentry_zone_size  26214400  
       zone_size       104857600   

4]  Remove the "/proc" filesystem
      # umount /proc
      vi /etc/fstab

5]  Install products from the two optional products disks.
        
      Disk 1 of 2.
        #  mount -dr /dev/rz14c /mnt
        #  cd /mnt/ALPHA
        #  ls
        #  setld -l DEC_Ada_RTL
        #  setld -l DEC_Fortran_RTL
        #  setld -l DEC_Pascal_RTL

      Add PAK for "MONITOR-UNIX", Cut from the 
      /tulsa/proc/cslg_paks_1997/reg_paks.csh file.

        # lmf register ...
        # lmf load ...

        #  setld -l  Performance_Manager
        #  setld -l Porting_Assistant
        #  cd /
        #  umount /mnt
      Disk 2 of 2.
        #  mount -dr /dev/rz14c /mnt
        #  cd /mnt
        #  ls
        #  cd AFA401
        #  setld -l kit
        #  cd /
        #  umount /mnt

6]  Shutdown system and remove cdrom drive.
        # halt
        >>> show device
        (remove cdrom drive) 
        >>> sho device

    Boot single user.
        >>> set boot_osflags ""
        >>> boot

7]   Add the "/usr/var" and "/tmp" filesets to the "usr_domain"
     domain.
      #  mkfset usr_domain var
      #  chfsets -b 250000 usr_domain var
      #  chfsets -B 250000 usr_domain var
      #  mkfset usr_domain tmp
      #  chfsets -B 100000 usr_domain tmp
      #  chfsets -b 100000 usr_domain tmp

     Mount the /tmp filesystem.
      # vi /etc/fstab
      usr_domain#tmp   /tmp   advfs rw 0 0
      usr_domain#var   /usr/var advfs rw 0 0

      # mount /tmp

     Save the /usr/var files temporarily.
      # cd /usr/var
      # tar cf /tmp/var.tar .
     Remove the /usr/var/files.
      # cd /usr
      # rm -r var
      # mkdir var
     Mount the /usr/var filesystem.
      # mount /usr/var
     Restore the files.
      # cd /usr/var
      # tar xpf /tmp/var.tar 
      # rm /tmp/var.tar
     Link /usr/tmp to /tmp
      # rm -r /usr/var/tmp
      # ln -s /tmp /usr/var/tmp

8]  Load software licenses
      Go to multiuser
      # ^D
      (Login to root)
      su>  /usr/sbin/setup
      (select "License Manager")
      Register packs for OSF-BASE and NET-APP-SUP-250

9] Configure the network
      su> /usr/sbin/setup
      (select "Network Configuration Application")
        Configure the front door and back door interfaces.
        Enable the "routed" daemon.
        Exit and start the network.
      su> /usr/sbin/setup 
      (select BIND Configuration Application")
        Configure as a client.
        Configure bind ahead of local (/etc/hosts).
        exit
      su> vi /etc/resolv.conf
        (Add search domains and nameservers as is done on 
         our other systems.  Where is the tulsa file at? )
      su> /usr/sbin/setup
      (select NFS Configuration Application")
        Configure the nfs daemons and add the jrr1 filesystems.
        exit

      Set search path in resolv.conf instead of domain.
      su> vi /etc/resolv.conf

      Set hosts entry in svc.conf to be "bind,local" to fix
      gethostbyaddr().

      su> vi /etc/svc.conf


10] Mount the tulsa filesystems
      Add symbol link for /tulsa filesystems and modify 
      /etc/fstab to use links instead of mount points.
        su> mkdir /nfs /nfs/jrr1 /nfs/jrr1/tulsa
        su> ln -s /nfs/jrr1/tulsa /tulsa
        su> vi /etc/fstab
        su> mount -a

      If the mount hangs, you may need to stop and restart the 
      routed daemon to get it to add routes obtained from both
      interfaces.

11]  Add the CSLG paks.
        su> /tulsa/proc/cslg_paks_1997/reg_paks.csh
        su> /usr/sbin/lmf reset
 

12]  Make directories for local software installation.
       su>  mkdir /usr/local
       su>  cd /usr/local
       su>  mkdir bin etc lib home include man share  man/man8
 
13]  Install 'srcmaint'
       su>  cd /tulsa/obj/du40/srcmaint*
       su>  make install
 
14]  Install 'tulsa'
       su>  cd /tulsa/obj/du40/tulsa*
       su>  make install
 

15] Install layered products.

    Mount filesystem from swdist.cac
    or use layered product cdroms from CWSOFT.
    The commands below have '/software/' in the path
    assuming that the install is from swdist.cac.

       su> mount swdist.cac:/sw/j/cslg /mnt

    (From disk #1)
    C++
      su> /usr/sbin/setld -l /mnt/software/cxx*/kit
          (all)

    Pascal
      su> setld -l /mnt/software/dpo*/kit
          (all)

    Fortran
      su> setld -l /mnt/software/dfa*/kit
          (all)

    DCE
      su> setld -l /mnt/software/dce2*/kit
        DCE Application Developers Kit 
        DCE Application Developers Manual Pages
        DCE Command Reference Manual Pages
        DCE Runtime Services

    (From disk #2)

    ADVFS 
      su> setld -l /mnt/software/afa4*/kit
          (Already done by OS install!)

    Capacity Planner
      su> setld -l /mnt/software/ecp*/kit
          (all)

    Performance advisor and collector
      su> setld -l /mnt/software/psdc30b/kit   
          (all)
      su> setld -l /mnt/software/pspa30b/kit    
          (all)

    DEC migrate
      su> setld -l /mnt/software/dmg120/kit
         

16]  Install the  du40a_patches product from /tulsa/bin to
     allow for High Uids.
       su> /tulsa/tools/tulsa_prodinst du40a_patches      
       su> cd /sys/conf
       su> ./config UWUCS
       su> cd ../UWUCS
       su> make depend
       su> make
       su> mv vmunix /vmunix.new
       su> init s
       # rm vmunix.old
       # ln vmunix vmunix.old
       # mv vmunix.new vmunix
       # sync; sync; sync
       # halt

16.b] Remove network configuration and hostname for Clone disk.
      If you are not preparing a clone disk, go to step [18].

      Unconfigure network interfaces.
        # /usr/sbin/netsetup

      Remove hostname, IP address configuration and change swap
      location.
        # vi /etc/rc.config
        # rm /sbin/swapdefault
        
This section is for cloning the system disk which has been
installed through step 14.  If you are not cloning a system 
disk, then proceed to step 18.

17] Clone the system disk.
    Detach scsi devices with the exception of the target
    system disk.
    Attach the clone disk to system as unit 0.
    Attach the target system disk as unit 1.
    If these numbers conflict with other SCSI units which
    you've detatched, you may need to do some extra work
    later when those disks are to be re-attached.
    >>> show device
    Note scsi disk model name for "lun 0" and "lun 1".
    >>> show config
    >>> set boot_osflags "I"
    Boot the system
    >>> boot dkb0
    If booting a 3000/400 or 3000/700, you may use the 
    "vmunix" kernel, otherwise use the "genvmunix" kernel.

    If the system is a 3000/300, the "rz" unit numbers below will"rz
    be "0" and "1" instead of "8" and "9".

      # mount -u /
      # mount /usr
      # mount /tmp
      # mount /usr/var
      # disklabel rz8 > /tmp/rz8.label

    If the clone and target disks are the same disk model as shown by
    the "show device" you did, then do step [17.1]
    Otherwise do step [17.2]


17.1] Clone the entire system disk.
      # disklabel -R -r -t advfs rz9 /tmp/rz8.label rz
      # dd if=/dev/rz8c of=/dev/rz9c bs=65536  &
      Ignore error on block 0.
      (Wait 2 hours)

17.2] Make custom label.
    Examine the label on the target disk and
    determine whether is is acceptable to hold the root,
    swap and usr filesystems.
      # disklabel rz9 > /tmp/rz9.label
    If no label on disk, put on default label first.
      #  disklabel -wr  rz9 rz
    Then, again:
      # disklabel rz9 > /tmp/rz9.label
      # cat /tmp/rz9.label
    Edit the label as necessary to set the A and B partitions to
    be the same and the "G" partition to be the remainder of the 
    disk.  The "C" partition should be the entire disk.
    The "A" partition should be type "advfs" and the others should be
    marked as unused.  
      # vi  /tmp/rz9.label
      # disklabel -R -r -t advfs rz9 /tmp/rz9.label rz
    Clone the "a" partition.
      # dd if=/dev/rz8c of=/dev/rz9c bs=65536 count=3305
      (Wait 15-30 minutes for copy to complete.)
    Setup the "b" partition.
      # disklabel -s rz9b swap
    Copy the "g" partition.
    Setup "g" as new usr_domain.
      # /sbin/mkfdmn -o /dev/rz9g usr_domain1
    Make filesets on "g" partition.
      # /sbin/mkfset usr_domain1 usr
      # /sbin/mkfset usr_domain1 var
      # /sbin/mkfset usr_domain1 tmp
    Mount usr_domain1#tmp on /mnt
      #  /sbin/mount usr_domain1#tmp /mnt
      #  chfsets -b 250000 usr_domain1 var
      #  chfsets -B 250000 usr_domain1 var
      #  chfsets -B 100000 usr_domain1 tmp
      #  chfsets -b 100000 usr_domain1 tmp
    Copy the files
      #  /sbin/umount /mnt
      #  /sbin/mount usr_domain1#var /mnt
      #  /sbin/vdump -0f - /usr/var | (cd /mnt; /sbin/vrestore -xf -)
      #  /sbin/umount /mnt
      #  /sbin/mount usr_domain1#usr /mnt
      #  /sbin/vdump -0f - /usr | (cd /mnt; /sbin/vrestore -xf -)

18] # shutdown -h +1
    (Power off disks and renumber to switch new clone with
     previous system disk.)
    >>> set boot_osflags ""
    >>> boot

    When boot succeeds, shutdown again and remove the unit 1 
    for use in cloning another system.  (Reset unit number to zero.)
    # halt
    >>> show device
    (remove disk unit 1)
    >>> show device
    >>> boot

19] Custom configuration done in single-user
    [ You may use script /DU40.step19 ]

    Reconfigure hostname and network for this node.
      # TERM=vt100
      # export TERM
      # mount -u /
      # mount /usr
      # mount /tmp
      # mount /usr/var

    Disable sendmail.
      # rm /usr/lib/sendmail /usr/sbin/sendmail
      # rm /usr/sbin/mailq /usr/sbin/newaliases
      # touch /usr/lib/sendmail
      # ln -s /usr/lib/sendmail /usr/sbin/sendmail
      # chmod 000 /usr/lib/sendmail

    Add whatever routes are required for this host 
    and configure network devices.  You'll ordinarily need to add
    a "default" route.
      # /usr/sbin/netsetup
    When you exit from netsetup, restart the network with this command:
      # /usr/sbin/rcinet restart

    Edit the /etc/hosts file to add the fully qualified domain
    name ahead of the unqualified name for the front and backdoor
    addresses.  

      # vi /etc/hosts

    Edit the /etc/svc.conf file to ensure the hosts lookup is set
    to check "bind,local".

      # vi /etc/svc.conf

    Add nfs fstab entries for tulsa.
      # mv /etc/fstab.tulsa /etc/fstab

    Set "xdm" for login, not "cde".

      # /usr/sbin/xsetup

    Setup nfs as client, and also configure at least one 
    filesystem for export if will also be nfs server. 

      # /usr/sbin/nfssetup

    For Alcott, also setup a caching named.

      # /usr/sbin/bindsetup

    Goto multi-user and login on console as root.
      # ^D
      (login)
      su>

Step [20] is also for clone systems only.

20] More custom configuration done in Multi-user.

    Add device special files.
      su> cd /dev
    Add pseudo ttys for an interactive server (Many logins)
      su> ./MAKEDEV pty0 pty1 pty2 pty3 pty4 pty5 pty6 pty7 pty8 pty9 pty10 \
          pty11 pty12 pty13 pty14 pty15
    Add SCSI disks devices.
    If a 3000/300
      su> ./MAKEDEV rz0 rz1 rz2 rz3 rz4 rz5 rz6 rz7  
    otherwise
      su> ./MAKEDEV rz8 rz9 rz10 rz11 rz12 rz13 rz14 rz15

    Restore local disk information from /tulsa/transfer/`hostname`/etc.tar
    files.
      su> cd /tmp
      su> tar xf /tulsa/transfer/`hostname`/etc.tar
      su> cat fstab
      su> vi /etc/fstab  # add additional local mounts
      su> mv fdmns/{????} /etc/fdmns  # Restore additional local domains.

     The lipw and fstore.cf files should have been saved in the 
     tar file restored to /tmp. 
      Ifso, use those instead of copying one from another system.

      su> mv /tmp/fstore.cf /etc/fstore.cf
      su> mv /tmp/lipw /etc/lipw

      Otherwise, get a copy from another sibling of this host.
      su> rcp 'sisterhost':/etc/lipw /etc/lipw
      su> rcp 'sisterhost':/etc/fstore.cf  /etc/fstore.cf

21]  Tulsa product configuration.

     Verify host name configured for tulsa files:
     .  /tulsa/tulsa.cfg
     .  /tulsa/fstab.cfg
     .  /tulsa/hosts/crontab/crontab/
     .  /tulsa/hosts/symbols
     .  /tulsa/tfx/std/DEFAULT/.rhosts

     In step [24] you will add the  "du40" tree to the configuration
     for this host in "tulsa.cfg".



    Build crontab
     su>  /tulsa/tools/mkcrontab
 
    Make filesystems and exports
     . Ensure host is configured in /tulsa/fstab.cfg
       su>  /tulsa/tools/decosf/bldfs

     . Verify that mount points are empty directories, dismount
       filesystems as needed.
 
     If this is a new host, rebuild exports on other systems which
     export filesystems to this host.
       su> /tulsa/tools/decosf/bldfs
 
22]  Mount local and remote filesystems.
     The nfssetup was done in step 19, so we are ready to mount
     the NFS filesystems.

       su> /sbin/mount -a     # May have all been done by bldfs, but anyhew.

23]  Add symbolic links to locate cluster accounting files
     . Link /cluster
       su> rm /cluster

       su> ln -s /sy27 /cluster         # UA Galaxy
      or
       su> ln -s cl00  /cluster         # Becker

     . Link /sy99
       su> rm /sy99 /usr/local/bin/sy99.bin
       su> ln -fs /cluster/sy99 /sy99
       su> ln -fs /sy99/bin.du40 /usr/local/sy99.bin
 
     Add symbolic links to NFS mounted directories.
     . Link /usr/spool/mail
       su> mv /usr/spool/mail /usr/spool/mail.dist
       su> ln -s /sy29/mail /usr/spool/mail    # Saul
     or
       su> ln -s /cl01/mail /usr/spool/mail    # becker

     . Link /public
       su> ln -s /sy39/public/osfalpha /public
     . Link /usr/ftp (saul only)
       su> ln -s /sy31/ftp /usr/ftp

    Add symbolic links to uwmsgs directory for this cluster.
     .  Link /usr/local/uwmsgs to cluster common directory.
        (This should be in tulsa!)
        su>  ln -s /sy30/uwmsgs /usr/local/uwmsgs   # Saul
       or 
        su>  ln -s /cl00/uwmsgs /usr/local/uwmsgs   # Becker

24] Run tulsa_daily to see what wants to get installed.

    Run the tfx_symbols script and verify the flags are set 
    correctly for this host.  Check the output against what
    you get on a sister host if you are unsure.  Check to make 
    sure that the "idf.dat" file is available or the script will
    fail to set the "group" definition right.
     
     su> rcp -p ${sisterhost}:/usr/local/bin/idfflags /usr/local/bin/idfflags
     su> /tulsa/tools/tfx_symbols
     su> cat /sy99/dat/idf.dat

     su> PATH=/usr/local/bin:$PATH
     su> export PATH

    Make a temporary copy of the "tulsa.cfg" file for this install.  
    Add the "du40" tree for the host, and disable the "sendmail" 
    install.  Pount the environment variable "TULSA_CONFIG" to it.  
       su> cd /tulsa
       su> cp tulsa.cfg tulsa.cfg.`hostname`.tmp
       su> vi tulsa.cfg.`hostname`.tmp
       su> TULSA_CONFIG=/tulsa/tulsa.cfg.`hostname`.tmp
       su> export TULSA_CONFIG

       su> /tulsa/tools/tulsa_daily
       su> grep -v /usr/local  /tulsa/logs/*`hostname`* | more
     Note which files are 'newer' and verify that the tulsa files
     are correct.

     These files are known differences and the changes have been
     installed in tulsa for the new DU40 systems.  To get the 
     newer files, the upgraded systems must get the "du40" tree
     configured for the host in "tulsa.cfg".

       .   /tulsa/bin/std/DEFAULT/etc/services
       .   /tulsa/bin/sy99/du40/sbin/init.d/security
       .   /tulsa/tfx/maintenance/DEFAULT/etc/shells
       .   /tulsa/tfx/init/du40/sbin/init.d/nfs
       .   /tulsa/tfx/std/du40/usr/share/lib/termcap
       .   /tulsa/tfx/std/DEFAULT/etc/resolv.conf
       .   /tulsa/tfx/std/du40/usr/share/lib/termcap

     These Digital Unix 3.2C files are also expected to be 
     installed in place of DU40 files:

       .   /tulsa/bin/more/decosf/usr/bin/more
       .   /tulsa/bin/more/decosf/usr/bin/page
       .   /tulsa/bin/pwd/decosf/usr/bin/pwd

     For these, save the vender programs as ".dist".
       su> mv /usr/bin/more /usr/bin/more.dist
       su> mv /usr/bin/page /usr/bin/page.dist
       su> mv /usr/bin/pwd  /usr/bin/pwd.dist

     Remove ftpd, as it has a gaping security hole.

       su> rm /usr/sbin/ftpd
  
     Change permissions on ping as it can be used to attack
     other sites computers.

       su> chmod 700 /usr/sbin/ping
 
 
25] Install ALL files from tulsa.

    Run 'tulsa_weekly -x' in the background, and monitor its progress.
    It has a lot of work to do, so you can expect it will take 
    some time before producing the cmp.log listing.
    (You likely need /usr/local/bin in your path, see [24])

     su>   /tulsa/tools/tulsa_weekly -x  &

     Save log files and repeat, then verify that log files 
     for both runs are identical! 

     su>   /tulsa/tools/tulsa_weekly

     Look for directories installed with the wrong permissions.
     su>   ls -lLd /usr/*
     su>   ls -lLd /usr/local/*

     su>  find /usr/local -type d -perm 000 -exec /usr/sbin/chmod 755 {} \;
     su>  chmod 700 /usr/local/cops

     
26] If an existing cluster do 26.1      else 26.2       

26.1] Adding a sibling to the cluster.

      Create the "/etc/nologin" file to discourage logins.

       su> touch /etc/nologin

      Update 'cluster-sites' file for this cluster member.
       su> vi /cluster/etc/cluster-sites

      Update '/etc/passwd'.

      . Copy password files another cluster member.
       su> rcp sister-host:/etc/passwd /etc/passwd
       su> rcp sister-host:/etc/shadow /etc/shadow

      . Build dbm files with /usr/local/etc/mkpasswd.
       (These could take a few minutes if you've got a lot of accounts)
       su> /usr/local/etc/mkpasswd /etc/passwd
       su> /usr/local/etc/mkshadow /etc/shadow

      . Test password entry retrieval.
       su> cpw -f noyd

      The second field should contain the encrypted password.
      If it does not, then the "/etc/sia/matrix.conf" symbolic
      link is pointing to the wrong file, or the shared library
      didn't get installed via "/tulsa/proc/sy99/usr/local/lib/libshd.a".

26.2] Generate password files.
     . Create the "/etc/nologin" file to discourage logins.

       su> touch /etc/nologin

     .  Sort /etc/passwd by uid and correct uid & group information
        as needed to conform with our standard. 
        su> vi /etc/passwd

     .  Create /etc/shadow password file.
        su> vi /etc/shadow

     .  Build passwd dbm datafiles.
        su>  /usr/local/etc/mkpasswd /etc/passwd
        su>  /usr/local/etc/mkshadow /etc/shadow

     .  Verify password retrieval.
        su>  ln -fs shadow_matrix.conf /etc/sia/matrix.conf
        su>  /usr/local/bin/cpw -f noyd

     .  Run cpwsync.
        su>  /usr/local/etc/cpwsync 

     .  Verify home directories and shells are correct.
        su>  tail /etc/passwd

      Build /sy99/dat files.
     . Obtain AUF and CDF information for each account
       to build /sy99/dat/auf.dat and /sy99/dat/cdf.dat.
       Sort auf.dat by uid and cdf.dat by account.
     . Touch rmf.dat.

 
27] Verify that the system startup procedures will invoke the
     startup of local products.
     . Check to be sure that the "syslog" startup is after 
       the network startup in /sbin/rc3.d, so the name
       service is available.
       su> ls /sbin/rc3.d
       su> rm /sbin/rc3.d/S10syslog

     . Install and enable sendmail.

       Enable the installation of sendmail in your temporary fstab.cfg
       and rerun for the install of sendmail.
      
       su>  touch 0101010170 /usr/lib/sendmail
       su>  vi /tulsa/tulsa.cfg.`hostname`.tmp
       su>  /tulsa/tools/tulsa_weekly


28] Verify some additional configuration files.

     .  Verify these files are correct:
          /cluster/etc/UWauthorize
          /etc/motd
          /etc/services               (product std)
          /etc/inetd.conf             (product inetd-config)
          /etc/syslog.conf            (product syslog)
          /etc/fstore.cf              (product fstore)
 
29] If 'archive' product is installed, verify that host is configured
     properly in nameservice (forward and reverse lookup) and listed
     in /usr/Unitree/etc/arcftp.hosts.

31] Run catman to build man page index.  This should have been
     done by tulsa, but may not have been?

     su> /usr/sbin/catman -w

     . Test 'man' command on local man pages.
     su> man perl

32] Add additional symbolic links that are preserved from previous
    filesystem migrations.  (UA galaxy only)

     su> /tulsa/transfer/saul_links.ksh

    That script can also be found as "/ux01/noyd/saul_links.ksh".


    Copy additional uwin files.  (SAUL, BECKER)
      saul8> cd /usr/uwin
      saul8> tar cf /tmp/uwin.tar config scripts

      su> rcp saul8:/tmp/uwin.tar /tmp/uwin.tar
      su> cd /usr/uwin
      su> tar xpf /tmp/uwin.tar

    Or, install directly from uwindev.u:

      % su2 uwin-adm
      % cd /usr/uwin-adm/install
      % rdist  -m uwinadm@${HOST}

    
    Copy all of /usr/local/bin to get non-tulsafied binaries. (ALCOTT)
         
      su> rsh ${sisterhost} tar cf /tmp/ulb.tar /usr/local/bin
      su> rcp ${sisterhost}:/tmp/ulb.tar /tmp/ulb.tar
      su> rsh ${sisterhost} rm /tmp/ulb.tar
      su> cd /
      su> tar xpf /tmp/ulb.tar

    Copy the /usr/lib/X11/rgb.* files from another system, or 
    install the files via the appropriate DEC product.


33] Shutdown and reboot

    Goto single user mode
      su> init s

    Remove the "/etc/nologin" file to enable logins.
      # rm /etc/nologin

    Halt
      # sync;sync;sync;halt

    Set BOOT_OSFLAGS to boot multi-user

      >>> setenv BOOT_OSFLAGS "A"
      >>> boot

34]  Add "uwmsgs" message to announce changes.


35]  Test "passwd" program to verify that it works.  If you get
     complaints about DCE, then disable the DCE password flag
     in the idf.dat file.

36]  Set up for Argus.

     melville> setargus up ${HOST}