Checklist

[Dan "The Man" Colpi <colpi_(_at_)_sneezy_(_dot_)_bio_(_dot_)_purdue_(_dot_)_edu>]

Greetings and Salutations,

A week or two ago I submitted a list of all the things you must to
to convert a DEC 3100 to OpenBSD given one that already has OpenBSD
on it.  I have since imporved this list as I have done a few more now.
So this is it, a mindless checklist of everything I did to "convert 
a DEC 3100 to the Dark Side"

If you can think of something I have missed, please throw me an e-mail.
Also, I am going to be converting a DEC 5000 soon as would like input
asto what I might have to do different.

Thanks

Daniel Colpi
colpi_(_at_)_purdue_(_dot_)_edu


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

So, you want another OpenBSD computer hu?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
(assuming same disk setup, target is disk #0, and source is disk #1)
(also assuming rrz0a -> /    ; rrz0b -> swap ; rrz0c -> whole drive
               rrz0d -> /usr ; rrz0e ->/usr/target  )
(Note: I am on a /bio backbone, this isn't going to be the same for you)
(Note: /bio/scratch3 is a temp space that I can write to)

___Save Files
   ___dump root to scratch (just in case)
                             -->% dump -0f /bio/scratch3/dump /dev/rrz0a
   ___/etc/passwd --> % mv /etc/passwd /bio/scratch3/passwd.target
   ___/etc/crontab
      ___any self created scripts you made to run with crontab
   ___/etc/fstab
   ___/etc/export.fs.[n]
      (if you have a drive that is set up the way you want for target
      machine, you might want to save it to a file to save you some
      time)
      (Note: you might want to do this on a machine that is already
             converted)
   ___save your partition list  -->% chpt -q /dev/rrz0a >> temp_part
                                OR % disklabel -r /dev/rrz1a >> temp_part
                                   % mv temp_part /bio/scratch3
___Shut down target
___Connect source drive
___Make sure drives are jumpered not to conflict (target=>0 ; source=>1)
___Plug in and power up
   ___Make sure clock is right time (some old computers lose some time)
   ___Make sure you see both drives (ie drive 1 & drive 0)
___Run disklabel  (I had drouble labeling the disk, but this worked)
        -->% disklabel -r /dev/rrz1a >> temp
           % disklabel -B -R /dev/rrz0a temp
           % disklabel -e /dev/rrz0a
       secs/track_______ * tracks/cil______ = secs/cil______ (the int
       # of cils________                                      multiple)

   ___ / (rz0a)           = _______________
   ___ swap (rz0b)        = _______________
   ___ drive (rz0c)       = _______________
   ___ /usr (rz0d)        = _______________ / ______________ (target / source)
                                            (remove src if desired)
   ___ /usr/target (rz0e) = _______________
        (if you have a "cookie-cut" version you can just:)
        -->% disklabel -B -R /dev/rrz0a /bio/scratch3/temp_part
___Run newfs  -->% newfs /dev/rrz0[ade]
___Transfer files
          --->% mount /dev/rz0a  /mnt
              % cd /mnt
              % df -k .     (make sure enough room)
              % df -k /
              % dump 0f - /dev/rrz1a | restore -rf -
              % rm restoresystable   (if you don't do this you'll be sorry)
              % cd /
              % umount /mnt
              % mount /dev/rz0d  /mnt                ---+
              % cd /mnt                                 |
              % df -k .                                 |  Repeate this 
              % df -k /usr/                             |  Repeate this 
              % dump 0f - /dev/rrz1d | restore -if -    |   section if
      restore > add [all dir's you want to keep]        |     needed
      (probally all but: CVS, src & source_machine      |       on
      restore > extract                                 |     other
      restore > set owner/mode for /./? [yn] n          |   partitions
      restore > quit                                    |
      (Note: it will say "DUMP not completed" if you    |
       did not restore everything.  This is not a bad   |
       thing, ignore it)                                |
              % rm restoresystable (might not be there) |
              % cd /                                    |
              % umount /mnt                          ---+
___Configure  (BE CAREFUL to configure /mnt/etc  NOT /etc !!!!)
   ___mount /usr --->% mount /dev/rz0d  /mnt
   ___make dir /mnt/target
   ___cd /mnt
   ___replace all scripts made to be run by crontab
   ___cd /
   ___umount /mnt
   ___mount /   --->% mount /dev/rz0a /mnt
   ___cd /mnt
   ___remove /mnt/etc/ntp.drift
   ___replace rest of scripts to be used by crontab
   ___move /bio/scratch3/export.fs.[n] /mnt/etc/export.fs.[n]
   ___hostname
      ___/mnt/etc/hostname.le1
      ___/mnt/etc/exports        (or run /usr/local/etc/mk-exports)
      ___/mnt/etc/hosts  (change IP number too) (may not have to)
      ___/mnt/etc/myname
   ___/mnt/etc/rc   (maybe not if (flavor == vanilla))
   ___/mnt/etc/mygate
   ___/mnt/etc/fstab    1 -> 0  & add any needed lines
   ___/mnt/etc/passwd
   ___rm -r /mnt/etc/rdist
      ___change root's home directory to /root from /
      ___reconfigure passwd file (for correct format)
          __ % mv /bio/scratch3/passwd.target /mnt/etc/passwd.orig
    same /   % nawk -F: ' { print $1 FS $2 FS $3 FS $4 "::0:0:" $5 FS \
    line \__ $6 FS $7 } ' /mnt/etc/passwd.orig >> /mnt/etc/passwd
             % cp /mnt/etc/passwd /mnt/etc/passwd.old
             % pwd_mkdb -c /mnt/etc/passwd
             % pwd_mkdb -d /mnt/etc/ /mnt/etc/passwd
             % pwd_mkdb -d /mnt/etc/ -p /mnt/etc/passwd.old
             % chmod 600 /mnt/passwd.orig
   ___do you want permissive /mnt/root/.rhosts (bilbo, biotest & gimli)
   ___how about cooperative /mnt/etc/ttys   (%s/network/network secure)
   ___if your source machine boots off disk 1 and target boots off
      disk 0, you need to make sure the correct kernal is at /mnt/bsd
   ___cd /
   ___umount /mnt
___Power down
___Remove source drive
___Correct jumpers if nessary
___Power up
___Before starting OpenBSD at Boot Ram type:
        >> printenv
        >> setenv bootpath rz(0,0,0)/bsd
___Boot off of target disk
   ___restore user file system
      --->% cd /usr/target
          % rsh bilbo ls /usr/amanda/\*
       __    (now in order)
 same /   % rsh bilbo cat /usr/amanda/target.date.[0->(n-1)] \
 line \__   < /dev/null | resore -rf -
             (now when done with above loop (IN ORDER!!!!!))
          % rm restoresystable
   ___recreate crontab
      --->% crontab -e
___Add target to any rdist entries nessary (probally source machine)
___Checks
   ___ping source machine (to make sure not same IP addresses)
   ___run ntpq -p 
   ___finger @localhost
   ___finger @target
   ___finger @target_IP_address
   ___print something
   ___cd into /bio/scratch3/
   ___run date
   ___telnet to target from another machine
   ___finger user_with_home_dir_on_target_(_at_)_target from another machine
   ___ftp one file from another machine
   ___cd to /bio/target from another machine
___Go home early

Nessary Tools:
===============
   Screwdrivers (if internal drives)
   Postit-notes (for keeping drives straight)
   One hell of a lot of coffee