These are the instructions to move the rtserver to a different pc.  Based on August 2003 
move of rtserver.



Jane Nachtman
update: 5/23/05
(updated the locations of various files: Bill Badgett, Mousumi Datta, 3/29/06)

1. people to page
   a. silicon -- they will have to restart their monitoring (both imon and svxinter)
   b. COT (Aseet)-- they will have to restart monitoring on cot2
      -- before doing anything with COT monitoring, need COT on standby or OFF
   c. Level 3 -- there may be config files which need to be modified.               


2. files to change

   a.  /cdf/onln/code/cdfprod/smartsockets/config/unix.cm, java.cm, vxworks.cm 
	-- change server_names
   b. same files on solenoid pc, on the MCS-S disk, under smartsockets-config
        -- call Steve Hahn
   c. /cdf/onln/code/cdfprod/smartsockets/v6_0-r2/standard/rtserver.cm
        -- change server_names

3. startup script on old machine and new machine must be modified so that rtserver is 
   started in the correct place

4.   the crate cpu's may need a hostAdd line in the ~vxworks/boot/front-end/startup.*
   files.  There are several of these being used by the various crates, so you have 
   to figure out which to change.  There are also svx monitoring processes and L3 
   files which may need a hostAdd line in a config file. Experts should do this, but may
   need a reminder.

5. procmon needs to be modified to watch the new machine

6. rtserver web pages which need to be modified to point people to the new machine

Recommended Procedure:

A. page people in 1, tell them what you are doing
B. start rtserver on new machine, change files in 2a, 2c
C. get ace and whoever else you can to start restarting everything in
   a coordinated fashion...
   -- every crate must be rebooted
       -- even teststand crates
   -- all the procmon-monitored processes 
   -- Level 3
   -- consumers?  xmon? svxmon
   -- dqmon, tevmon, daqmon, run control
   -- restart modified procmon watching new rtserver.  ss_conntest.c needs to be modified 
      for procmon check.
   -- kill sscheck on old machine, start on new.
D. monitor connections on the old and new rtserver, this will help you to see what you 
   still need to do
E. in parallel with C, change pc files in 2b. Then get silicon and COT people to restart 
   their monitoring, and restart icicle.  See details below.
F. Notes about what needs to be done for silicon -- don't do these unless silicon 
   person is present :
   -- reboot silicon daq crates
   -- reboot Svxinter caen controller (this gave some problems last time)
   -- restart PSGUI and IMON
F. once all connections are switched over, turn off old rtserver, do test runs, calibrations,
   etc.
G. send mail to cdfonline list letting them know what you did.



Details for special processes on windows pcs, from 2003.  Get the experts
to do COT and SVX.

   a. icicle (on solenoid, kill icicle by X, click on icicle icon to start)

   b. PSMon and TempMon on the svxicon(?) pc (second to last row in 1st floor
counting room), type q in window and wait a few
minutes for them to quit, then click the icons to restart.
Andy Hocker is the expert on this.

   c. COT2 (right away as you go into the 1st floor counting room) - click on
"Smartsockets Running" to stop it, then click on same button (says "Smartsockets S
topped" now) to restart.  If you get a memory error, you have to stop and
restart the MCS program.  There is a chance that the pc will crash, too, in
which case you would have to call Aseet or page on-call person.  Aseet is the
expert on this program, probably best to call him before you do anything.