Note
When the system first initializes using the new disk packs, PLATO will start but has some non-fatal issues. You may safely ignore any of those issues for now.
Overview
In this article, we begin the process to upgrade all other parts of the system:
The PLATO build process relies on several procedures. The ones shipped with Release 1 are insufficient and will be replaced during this phase.
plmods 1-b bu 1-c mod files 1-d retro1.org tpmods gogomods maintp s0maint1-b ReadMe1-c maintp maintsub o are the “Original” blocks which should be the same as the ones running at the start of this procedure if all prior steps were done correctly. Blocks without the o prefix are the ones which will replace their respective contents in the deadstart tape.getall, getsingle and tgetselect update the contents of the o code blocks in this file with different conditions. getall updates the o blocks with the contents of the currently running system.getsingle selectively updates individual blocks from the currently running system contents.tgetselect to retrieves selected blocks from a deadstart tape which is not currently running.1-b ReadMes0maint maintdf dayfile contains this output."Just because you can, doesn't mean you should"
This step should be very carefully considered. cdc.io is well-tested but should be used sparingly until this first stage is complete.
Using ''CDC.IO'', all other files that need to transfer from any other installation may be moved dynamically to the new installation.
Danger
Do NOT transfer files to a live system unless you really know what you are doing.
Before PLATO can be built, the core maintenance procedures need to be updated.
maintp.c (code block maintp)jobstat to monitor its progress.Warning
The dayfile from this procedure is written to listing block dayfile in file maintp so don't return to this code file from jobstat until the job is completed.
Once the job has successfully completed, update the backup (as in Plan B) procedures which may be used through the console to rebuild PLATO.
maintsubc (code block fix-builds)jobstat to monitor its progress.Warning
The dayfile from this procedure is written to listing block dayfile in file maintsub so don't return to this code file from jobstat until the job is completed.
EXECUTE lesson s0maint (enter the name then press DATA)
Choose Option:
a. Full Assemblies <Press NEXT> r. Full System Assembly <Press NEXT>
Shift-LAB to invoke personal defaults.
Shift-NEXT to build the batch job.
Shift-LAB to submit the job.
NEXT to view the job in lesson jobstat.
Take a BIO1) Break
Review the dayfile output in code file maintdf block dayfile.
At This Point
PLATO has been built, but does not exist on the deadstart tape. If there is a need to restart the system, the old deadstart tape will still work properly.
Transferring new binaries AND the new configuration elements defined in maintsub to a new deadstart tape is the final stage in performing the upgrade.
Use the show_tape dtCyber command to see the status of the loaded tapes:
At the dtCyber console, attach a new blank tape to NOS Tape unit 51 (C31 E00 U01) using the emulator Operator> prompt. We assume that you have created a dedicated deadstart tape directory called deadstart relative to the location of the cyber.ini file. It is also recommended that you use the naming convention shown below to maintain naming of the deadstart tapes in time sequence:
load_tape 13,0,1,w,YYYYMMDDX.ds.tap
Where YYYYMMDD is the current date and X is the sequence letter A through Z.
maintsub.1-f mkdstart.jobstat but your attention should be on the NOS Virtual console.assign,<jsn>,050. (Where <jsn> is the Job Sequence Number of the requesting job.)
assign.<jsn>,051. (Where <jsn> is the Job Sequence Number of the requesting job.)
This job does very detailed reporting of what it is doing and its output will be generated to the system printer when complete.
You may now detach your new deadstart tape at the dtCyber operator prompt:
unload_tape 13,0,1
Modify ''cyber.ini''
Once the new deadstart tape has been successfully created you should edit cyber.ini (The Deadstart Tape Location) to identify the new deadstart tape file. The other option is to copy/rename the original deadstart someplace and then copy the new tape over the old one.
Preserve Deadstart Images
The backup of a deadstart tape is essential since you will always want a known good deadstart to fall back on if things go wrong.
At this point, the new deadstart tape should have been properly built and the currently running system may now be shut down and restarted.
Once the deadstart has completed, on the NOS console issue the E,C. command to check the status of the new disks:
Confirming that there are no errors, proceed to start CYBIS by issuing the NOS console command CYBIS..
Note that there will still be the ERROR - SYSTEM ID .NE. ROUTING ID. Ignore that for the moment and connect a PTerm session.
You will see a number of “success indicators”:
You may proceed to log in as admin/s.
You will see an error message from s0init: system *ncc* not in network table. This is a good thing!
Now proceed to:
Let CYBIS reinitialize and the following should appear:
You may now re-connect the PTerm session and login as admin/s:
maintsub.1-g productupd.jobstat.
When this job completes, the PLATO binaries will have been incorporated into the NOS file PRODUCT/UN=INSTALL on the chance that it may be needed for future enhancements. This file is used by the NCC distribution to maintain an image of the deadstart tape during product installation.
Deadstart again… Just to convince yourself that everything is working properly.
Next Step: Move on to version management and Create a Backup Repository
Next Step: NDL Modification
Next Step: NOS Configuration Summary