top - download
⟦86df23fed⟧ Wang Wps File
Length: 10289 (0x2831)
Types: Wang Wps File
Notes: Spelunked
Names: »~ORPHAN64.00«
Derivation
└─⟦3b8d69e90⟧ Bits:30005794 8" Wang WCS Floppy ( A0306A )
└─ ⟦this⟧ »~ORPHAN64.00«
WangText
3854/rt…02…FIX/0000/USM/0006
…02…APE/830630…02……02…#
FIKS DISK
OPERATION
…02……02…FK 7809
APPENDIX A
FIKS DISK OPERATION…86…1 …02… …02… …02… …02…
T̲A̲B̲L̲E̲ ̲O̲F̲ ̲C̲O̲N̲T̲E̲N̲T̲S̲
1. INTRODUCTION ..............................
003
2. S̲U̲M̲M̲A̲R̲Y̲ ..................................
004
2.1 HARDWARE ..............................
004
2.2 SOFTWARE ..............................
006
3. FIKS DUAL DISK SYSTEM .....................
008
4. INITIALIZING OF FIKS DISK SYSTEM ..........
009
5. USE OF TOS IN FIKS DUAL DISK SYSTEM .......
010
6. S̲Y̲S̲T̲E̲M̲ ̲O̲P̲E̲R̲A̲T̲I̲O̲N̲A̲L̲ ̲P̲R̲O̲C̲E̲D̲U̲R̲E̲S̲ .............
010
6.1 BOOT LOAD OF FIKS .....................
011
6.2 DUALIZING OF DISKS ....................
012
6.3 ISCARDING OF DISKS ...................
012
6.4 DISK STATUS PRINTOUT ..................
013
6.5 USE OF TOS: ...........................
014…86…1 …02… …02… …02… …02…
1. I̲N̲T̲R̲O̲D̲U̲C̲T̲I̲O̲N̲
This document gives an overview of how the FIKS Disk
System functions and how it shall be operated. It is
meant to be read of technicians, system operator and
othr persons, who interface to the FIKS Disk System.
For more detailed description refer to the following
documents.
ESP SYSTEM PSP
Product Specification, FIX/1105/PSP/0046
CR80 AMOS I/O System,
Product Specification, CSS/006/PSP/0006
CR80 DMALink Driver,
Product Specification, CSS/006/PSP/0002
CR80 AMOS File Management System,
System Product Specification, CSS/920/SPS/0001
CR80 Disk Driver,
Product Specification, CSS/006/PSP/0005…86…1 …02…
…02… …02… …02…
2. S̲U̲M̲M̲A̲R̲Y̲
2.1 H̲a̲r̲d̲w̲a̲r̲e̲ (ref. fig. 1)
The disks (CDC and FLOPPY) are controlled of the interface
boards placed in the file processor. These diskcontrollers
are accessed from te file processor via the SCM-mainbus
using the IO-addresses 200,204 (=4(#32, #33)) for CDC
and 8(=2*4) for floppy. In this way they are requested
to perform data transfers to/from the disks from/to
the disk-CACHE-CR80-memory RAM placed page 1, locaions
0 - #4000). These transfers are executed using a SUBBUS.
Transferrence of data between the file- and user processor
is done via the DMA-link.
The CDC-diske (kind = MMD82, unit no = 0) consist of
two submits:
- the section accessed with movble heads (submit no
= 0), where the volume MOVHEAD is placed.
- the section accessed with fixed heads (submit no
= 1), where the volume FIXHEAD is placed.
The Floppy disk (kind = FLOPPY, unit no = 0) is in
the Node/MEDE configuration only placd in BRANCH ONE
(i.e. it can not be used of BRANCH TWO).
There is one submit (submit no = 0), where a volume
with name FLOPPY (normally) is placed.…86…1 …02…
…02… …02… …02…
FIG. 1…01…FRONT VIEW
DISK OPERATION…86…1 …02… …02… …02… …02…
2.2 S̲o̲f̲t̲w̲a̲r̲e̲ (ref. fig. 2)
When an application in the CR80-computer requests a
disk read/write, this command is sent via the IO-system
in the user processor and DMA-driver (DMA00) for the
File Management System (FMS) placed in the file processor.
The FMS translates the commands to disk-sector read/write
commands. These are handed over to the disks drivers
(CDC000, CDC001 for the CDC-disks and FDD000 for the
Floppy), which arry out the interface to the disk-controllers.
Needed data transfers between file- and userprocessor
is controlled by the FMS and executed by the DMA-driver.
When the disk-operation is finished the disk-drivers
are interrupted. They return completon to the FMS,
which forwards this information via DMA-drivers and
IO-system to the application. Certain completians (disk
hardware errors) are handled of the system software
(ESP).…86…1 …02… …02… …02… …02…
FIG. 2
D̲I̲S̲K̲ ̲O̲P̲E̲R̲A̲T̲I̲O̲N̲…86…1 …02… …02… …02… …02…
3. F̲I̲K̲S̲ ̲D̲U̲A̲L̲ ̲D̲I̲S̲K̲ ̲S̲Y̲S̲T̲E̲M̲
Seen from the system operational software point of
view, the FIKS dual disksystem (ref. fig. 1, Node/MEDE-config.)
consist of two logical DEVICES:
- MMO, with volume MOVHED
- FIXO, with volume FIXHEAD
The devices may consist of submits from one or both
of the disk units. If both submits are used then DISK
̲STATUS is said to be DUAL. If only submits from BRANCH
ONE/TWO are used then DISK ̲STATUS s said to be ONE/TWO.
If DISK ̲STATUS is DUAL then the disk read operations
will be performed from one of the disk units and disk
write operations on both units. In case of a hardware
failure in one of the disk units, this one is DISCARDED
and theone unit left is used as single (ONE/TWO). Later
on when the erroneuous disk is repaired, it can be
included in the disk system again - a DUALIZING of
the disk is started. All write operations are now performed
on both units while a copying of disksectors from the
disk used all the time to the disk to be included.
When this copying is finished the units are identical
and DISK ̲STATUS is DUAL.
Both BRANCHE's can use the disks at the same time.
If both branches were allowed to do writing on te disk
units simultaneous, then the disks would soon be corrupted.
Therefore a feature for controlling the ability of
being able to write on the disks is present. When the
FIKS-system is booted, no disk writing can be done
- every disk write operaton is converted to a dummy
'CHECK'-operation. All diskread-operations are executed
in the usual manner. This state of the disk system
is meant to be used when a branch is STANDBY. When
the branch is going to be ACTIVE DISK-writing is enabled.
(WRTDSK).
To keep track of the different states of the disk system
a DISK ̲STATUS ̲WORD is used.…86…1 …02… …02… …02… …02…
15 14 3 2 1 0 bit
no
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
̲ ̲ ̲ ̲
̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲ ̲
̲ ̲ ̲
B̲i̲t̲ ̲o̲n̲ ̲m̲e̲a̲n̲s̲
0: DISK ̲ONE ̲INCLUDED
1: DISK ̲TWO ̲INCLUDED
2: DISK ̲WRITE ̲ENABLED
3: DISKS ̲DUALIZING
This word is checkpointed (i.e. written to disk) to
be retrieved at next initializing of the disk system
for determinaton of DISK ̲STATUS.
It is noticed that in BRANCH ̲ONE correspond DISK ONE/TWO
to the IO-addresses 200/204, and in BRANCH TWO correspond
DISK ONE/TWO to the IO-addresses 204/200.
4. I̲N̲I̲T̲I̲A̲L̲I̲Z̲I̲N̲G̲ ̲O̲F̲ ̲F̲I̲K̲S̲ ̲D̲I̲S̲K̲ ̲S̲Y̲S̲T̲E̲M̲
When the FISK-system is bot loaded BRANCH ̲ID, DISK
̲STATUS is defaulted to BRANCH ̲ONE, DISKS ̲DUAL. The
state of the disk-controllers is examined.
If the cables to the controller or the interface board
itself is missing, then no attempt to use the corresponding
disk unit i performed - the DISK ̲STATUS is set in accordance
to this. When BRANCH ̲ID id keyed in by the operator,
DISK ̲STATUS is updated again (if BRANCH TWO is stated).
The next step in the initializing procedure is to examine
the earlier checkpointed DISK ̲SATUS ̲WORD's (one on
each disk unit): By combining the obtainable DISK ̲STATUS's-
(those from the disk controllers and the checkpointed),
the largest common DISK ̲STATUS is determined. In case
DISK ̲STATUS comes out with "no disk available" the
system perator is asked to key in a DISK ̲STATUS. If
the state of the branch is going to be ACTIVE the disk-write
is enabled.…86…1 …02… …02… …02… …02…
5. U̲S̲E̲ ̲O̲F̲ ̲T̲O̲S̲ ̲I̲N̲ ̲T̲H̲E̲ ̲D̲U̲A̲L̲ ̲D̲I̲S̲K̲ ̲S̲Y̲S̲T̲E̲M̲
Normal operations with the TOS-system will take place
in BRANCH ONE and on DISK ONE-floppy drive is only
connected to BRANCH ONE. It is presmed that some kind
of updating is performed on the disk and last time
FIKS was active the DISK ̲STATUS was DUAL. Therefore
the checkpointed DISK ̲STATUS is automatically set to
ONE when the TOS-system is used. Next time FIKS becomes
active, DISK ̲STATS will be one and a 'DUALIZING of
Disks' - procedure can be performed. If the system
in some manner is reduced (one BRANCH or DISK is missing)
and TOS is used, then the checkpointed DISK ̲STATUS
shall be set in accordance to the case. To do this
thee exist three utility programs DISK ̲ONE.c, DISK
TWO.c and DISKS ̲DUAL.c which overwrites the checkpointed
DISK ̲STATUS. These are manual activated by the operator
(Use S-RUN 'Program')
The FIKS-fileboot may be used together with TOS-boot-no
disk wrting will be performed. The operations that
may be executed in this way can not be predicted, but
at least there is no risk of damaging something.
6. S̲Y̲S̲T̲E̲M̲ ̲O̲P̲E̲R̲A̲T̲I̲O̲N̲A̲L̲ ̲P̲R̲O̲C̲E̲D̲U̲R̲E̲S̲
This section deals with what a system operator has
to do n certain selected cases. All keying in is performed
on the console, - transparent Watchdog mode to the
affected branch is assumed. Keying in is underlined.
6.1 B̲o̲o̲t̲ ̲l̲o̲a̲d̲ ̲o̲f̲ ̲F̲I̲K̲S̲
B1 A̲ ," normal
B2 B̲
ERROR SWITCHOVER ...... etc.
if no response or the system oes respond as stated
then key in
I 3̲3̲ "boot from opposite
B1 A̲ "site branch
B2 B̲
the following may appear
DISK STATUS ?
O̲E̲/T̲O̲/D̲L̲
The system has found that no disk can be used as a
result of the disk investigation performed at te disk
initializing. The operator may then close the proper
DISK ̲STATUS by keying in OE, TO or DL(ONE, TWO or DUAL).
The DISK ̲STATUS used last time the system was FIKS-ACTIVE
can be found in earlier log printing - DISK ̲STATUS
CHANGES IS ALWAYS LOGGD:
Note: The operator has to make a decision that may
have catastrophic effect on the system in case
it is wrong. Is the situation caused by hardware
failure? Wrong BRANCH ̲ID keyed in? Uncontrolled
use of TOS-system.
If an errorreport with errrcode # 4FF (disk system
̲stopped) appears, the just after bootload, then this
may be caused by an unlucky hardware error in one of
the disk unit. It is then better to exclude this unit
from the system. This is done by switching off the
power to th
