UPGRADING MAIN SSP TO 3.1
------------------------- 

1. Verify spare SSP is configured as spare: 
  
  - There are two things to check in verifying that an ssp is spare 
  
    a. tftp service is commented out in /etc/inetd.conf

    b. there is no ssp_startup.sh entry /etc/inittab

  - If the spare is not configured correctly, shutdown the spare 
    before doing anything else.  It'll be reconfigured after it is
    upgraded.

2. main-ssp# init 1 

  - This is essential to prevent the daemons from attempting to restart
    during the upgrade.
    
3. main-ssp# /cdrom/ssp_3_1/ssp_upgrade /cdrom/ssp_3_1

  - The ssp_upgrade script was modified to include the ssp_terminate
    command as an internal function.  The command referenced in the
    install doc, /opt/SUNWssp/bin/ssp_terminate, has a bug which will 
    not terminate the daemons properly so there is no need to run it
    prior to ssp_upgrade.  There is no harm in running the command.

  - The ssp_upgrade script will then pkgrm and pkgadd all the ssp packages
    in the correct order.

  * NOTE: pkgadd assumes ssp is main and will update inittab to contain
	  the ssp_startup.sh entry.

  Verify the following:

  - /etc/inittab and /etc/inittab.main should be identical and both 
    should contain the ssp_startup.sh entry.
      
  - /etc/inittab.spare should not contain the ssp_startup.sh entry.
    If necessary, manually edit this file and remove the entry.

4. main-ssp# lockfs -fa

  - If the kernel patch is NOT 103640-12 or better, do not reboot 
    without issuing this command.  Use uname -a to get kernel rev.
  
  - Consider installing the latest Kernel Patch (currently 103640-17).

5. main-ssp# reboot

  - Allow the ssp to go into multi-user and give it time start all the 
    daemons before doing anything further.  Feel free to login and 
    monitor progress by watching the platform messages file.

	tail -f /var/opt/SUNWssp/adm/messages

  - Look for cbs/cbe connection established message (port 1025) and then
    proceed.

6. Install patch 105873-01. Adding this patch removes the requirement 
   of rebooting the domains prior to a DR operation.

  - If you have problems installing patch, reread step #5.


UPGRADING THE SPARE SSP TO 3.1
------------------------------

A. spare-ssp# init 1

B. spare-ssp# /cdrom/ssp_3_1/ssp_upgrade /cdrom/ssp_3_1

  - Same as above.

  Verify /etc/inittab.main and /etc/inittab.spare as in 3 above.

  - Do not modify /etc/inittab.  Only modify inittab.main or inittab.spare 
    if necessary.  Running ssp_config (next step) will take care of 
    /etc/inittab.

C. spare-ssp# /opt/SUNWssp/bin/ssp_config

  - This is required since pkgadd assumes the ssp will become main.  See
    bugid #4112713.

D. Verify spare is configured correctly.

  - Same as #1 above

E. spare-ssp# lockfs -fa

  - Same as #4 above.

F. spare-ssp# reboot

  - Allow ssp to come up into multi-user mode and verify that no daemons 
    are started up.  If they do, spare was not configured properly.

G. Install patch 105873-01 - Same as #6 above.


PROBLEM/SOLUTION
----------------

P. Customer wants to use spare ssp as an install server and requires
   tftp service enabled on the spare.

S. Remove hard links to cbe.ima and cb_port in /tftpboot.  Uncomment 
   tftp service in /etc/inetd.conf, kill -1 inetd.

   - The links in tftpboot will be XXXXXXXX and XXXXXXXX.cb_port, where
     XXXXXXXX is the hex representation of the control boards IP 
     address. Do NOT remove cbe.ima and cb_port.

   - To revert back, you will need to rerun ssp_config with the cb
     option to supply the control board IP addresses again.  This will
     rebuild the links to cbe.ima and cb_port.  See bugid #4112732.


P. cbe hang, runtsk error (-11)

S. run cb_reset and see bugid #4043945

   - Probable cause is that the ssp and control boards are not in a 
     private network.  This is not a subjective solution. Nothing else
     should be on the hub except the ssp and control board period.


P. daemons are running on spare ie cbs, snmpd, straps, edd, etc...

S. Check /etc/inittab and rerun ssp_config as root if necessary.

   - Recheck /etc/inittab, lockfs -fa if necessary and reboot.
   

P. Can't run bringup

S. Ensure that cbe has established connection with cbs (port 1025)

   - If connection does not get established, try running cb_reset and
     monitor the platform messages file for connection.


P. Bringup/hpost fails in lbexit phase.

S. Ensure that edd is not interfering with the hpost run and try again.  
   If problem persists, escalate problem.

   - Again monitor the platform messages file.  If edd is checking for
     thermal calibration, wait until it is finished or stop edd with
     edd_cmd -x stop.

   - Install patch 105873-01.


P. Cannot install patch 105873-01

S. Ensure that you uncompressed and untar'd as ROOT and then ran
   installpatch as root and in multi-user mode.

   - Permission problems will cause installpatch to fail with the
     following if not untar'd as root:

     pkgadd: ERROR: checkinstall script did not complete successfully

   - Installpatch will fail with "process interrupted" if not in
     multi-user mode.


P. Still getting errors.

S. Log a call with SunService or escalate the problem.