| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 34.1 | Version 9.0 Console | KERNEL::SOWTON | Diagnosis does it down the phone.. | Fri Apr 28 1989 00:24 | 428 |
|
Console version 9.0 is now out...here's an extract from the
release notes.
================================================================================
FEATURES and BUG FIXES for VERSION V9.0 / VAX RELEASE 36
--------------------------------------------------------------------------------
*IMPORTANT NOTE*
The Revision Control commands and data have been updated for Ver-
sion V9.0, Release 36. The Revision IPRs (Internal Processor Reg-
isters) including byte 2 of the SID (the System ID register) *WILL
CHANGE* as you upgrade to Console Version V9.0, Release 36. Soft-
ware that is incorrectly keyed off the entire SID *LONGWORD*
(rather than the serial number field contained in the low 2 bytes)
may need to be corrected. Please contact your software vendors if
you need assistance.
New commands:
1. The >>>ABORT command is used to terminate command procedures.
2. The >>>GOTO command is used in command procedures to trans-
fer control to a line other than the next line in the flow.
3. The >>>POWER TRIP_BREAKER command will cause a "TOTAL OFF"
condition, where CB1 on the 876 power control will be trip-
ped and all power removed from the system, *INCLUDING* the
Console subsystem. Manual intervention is required to reset
the breaker.
4. The >>>SET { } {/[NO]LOGGING,/[NO]PRINTING} command allows
you to enable or disable logging or printing for any stream
or OPA.
5. The Micromonitor's >>>SHOW FLAGS command provides a count-
erpart to the Micromonitor's >>>SET FLAGS command. The
>>>SHOW FLAGS command can only be issued while in Micro-
monitor (MIC>) mode.
6. The >>>SHOW LOGGING/SHOW PRINTING commands display which
streams and opas are enabled for logging and printing.
7. The >>>SHOW OPA command displays logging, printing, and
XON/XOFF state information for the specified opa.
8. The >>>SHOW STREAM command has been added as a synonym for
>>>SHOW TERMINAL.
9. The >>>TYPE command allows you to type out a file that is
stored on the console's fixed disk without the need to exit
from the Console application.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 5[m
10. The >>>WRITE command will issue a message from within .COM
files without the need to destroy the setting of the VERIFY
switch.
11. The >>>XON and >>>XOFF commands can explicitly start or stop
the transmission of characters from the VAX to the various
console terminals or OPA ports.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 6[m
Modified commands:
1. The Micromonitor's >>>CLEAR FLAGS command now allows you
to specify 'ALL' in place of specific flag names. ('ALL'
flags has always been the default but now this can be ex-
plicitly specified.) The >>>CLEAR FLAGS command can now
only be issued while in Micromonitor (MIC>) mode.
2. The >>>CONTINUE command is now valid when the VAX CPU is
running. It will simply return you to PROGRAM mode. This
was done to make this Console implementation more compatible
with the Consoles of other VAX processors.
3. The >>>EXAMINE command can finally use the mnemonics CBER,
EBER, and IBER to access the respective processor IPRs.
4. The >>>HALT command will wait for the cpu to halt and will
time out if the cpu doesn't halt. Therefore the >>>WAIT
command is no longer necessary and no longer exists.
5. The >>>HELP command has been enhanced and the helpfile text
has been improved. Helpfile text is now page-oriented and
the screen displayed by simply typing >>>HELP has been vast-
ly simplified.
Various new HELP topics are available.
User-written topics may now be included by simply creating
files (in the [CONSOLE], [8800], [8700], [8550], or [8500]
directories) with the extension .HLP. A specific help menu
topic, "USERTOPICS" has been created to allow you to list
user-written topics.
6. The >>>IF command now allows you to test for "BOUNDED" (a
narrow cabinet, e.g, a VAX 8530 or VAX 8550) and "DUAL"
(a VAX 8800 or 8820-N).
7. >>>INITIALIZE/CREATE will now delete all old versions of
MEMCONFIG.DAT, displaying a confirmational message with
each version deleted.
It will then go on to perform all of the functions of
>>>INITIALIZE/MEMORY including the creation of a new
MEMCONFIG.DAT file and the installation of that data
into the MCL Decode RAM.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 7[m
8. Changes to both >>>INITIALIZE/CREATE and INITIALIZE/MEMORY:
- Various memory misconfigurations will now be reported
explicitly. Initialization will continue if the error
was non-fatal. Previously, these errors simply caused
the command to fail without citing an explicit reason.
- The format of MEMCONFIG.DAT has been improved somewhat.
In addition, MEMCONFIG.DAT will now be checked for cor-
rect syntax and applicability prior to use. If MEMCON-
FIG.DAT contains an error or is not applicable to the
current array configuration, it will be recreated.
- All of memory will now be cleared any time MEMCON-
FIG.DAT is created.
9. The Micromonitor's >>>SET FLAGS command can now only be
issued while in Micromonitor (MIC>) mode.
10. The >>>SET MARGINS command can now be issued with no para-
meters. This causes all regulators to be returned to NOR-
MAL margins. In addition, the keyword 'ALL' can now be
used in place of specific regulator names. ('ALL' regu-
lators has always been the default but now this can be
explicitly specified.)
11. The >>>SET NEXT_PRIMARY command will now accept the full
range of CPU specifiers (e.g., LEFT, RIGHT, NEXT_SECONDARY,
etc.). Previously, it was limited to either LEFT, RIGHT,
0, or 1. The command will now also print a confirmational
message displaying whether the LEFT or RIGHT CPU was se-
lected.
12. The >>>SHOW CPU command and the CPU portion of the >>>SHOW
STATUS command now includes much more detailed information
as to the exact STATE of each CPU.
13. The >>>SHOW LOGFILE command now operates much faster at the
LOCAL terminal. New, EDT-like keypad commands allow easy
access to the beginning or end of the LOGFILE. Improved
help and status information is available while viewing the
logfile.
14. The >>>SHOW REVISION HISTORY command output has been re-
formatted to increase clarity.
15. The >>>SHOW REVISION SENSED command output has been re-for-
matted to increase clarity and to include information about
module presence and correctness. This command now indicates
the revisions of each individual VAXBI channel. Also, it
now always shows all four possible VAXBI channels, no matter
what CPU model the command is executed upon.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 8[m
16. The >>>SHOW TERMINAL command and the terminal portion of the
>>>SHOW STATUS command has been enhanced to display the
<XON>/<XOFF> status of the OPA port in use by that terminal.
>>>SHOW [ALL] TERMINALS is now allowed.
17. The >>>START (macrocode) command now affects only one cpu.
18. The >>>START/CONTROL_STORE (microcode) command still can
affect both cpus but now requires you to >>>SET CLOCK OFF
first. This feature was added to reduce the risk of ac-
cidentally (re-)starting microcode thereby crashing the CPU.
19. The >>>TEST command (and its variants) will now prompt for
confirmation if it is issued while one or more CPUs are
believed to be running.
20. The >>>VERIFY REVISION command output has been re-formatted
for increased clarity. In addition, the algorith that de-
termines "top rev" has been simplified and improved, with
more explicit error reporting of invalid module combinations.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 9[m
Obsolete commands:
1. >>>SENSE REVISION CONSOLE -- This is now done automatically
at the time that the Console Application is invoked.
2. >>>SENSE REVISION CPU -- This is now done automatically
at the time that power is applied to the CPUs.
3. >>>VERIFY MODULE PLACEMENT -- This is now done automatically
at the time that power is applied to the CPUs.
4. >>>WAIT -- >>>HALT and >>>START/CONTROL now automatically
wait for the CPU to actually halt and give an error message
if it doesn't.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 10[m
Bug fixes:
1. Reliability of recovery from "brownouts" has been improved.
2. Various programmed delays during power-on and power-off
have been lengthened to allow a greater range of regulator
turn-on and turn-off times. This should eliminate the spo-
radic issuance of the "Power not set to desired state"
error message.
3. The V9.0 Console corrects a problem where previous Consoles
would fail to detect subsequent air flow faults after an
initial air flow fault. In addition, the clearing of air
flow faults is now detected correctly. Previouly, an
already-cleared air flow fault might still result in the
shutdown of the CPU.
4. The V9.0 Console corrects a problem where the V8.x Console
would occasionally fail to display the initial typeout from
a VAX program after the >>>START or >>>CONTINUE commands
were issued. (All the output *WAS* correctly captured at
the PRINTER and in the LOGFILE.)
5. Previously, a Micromonitor >>>CLEAR FLAGS/>>SET FLAGS com-
mand which contained a syntax error could still change the
state of the Micromonitor flags. All of the occurences of
this behavior have been eliminated in V9.0. (The Micromon-
itor >>>SELECT command still manifests this bug -- This will
be corrected in a future release of the Console.)
6. Previously, a Micromonitor Single-Bit Error threshold great-
er than 32767 would be incorrectly displayed as a negative
number. This has been corrected.
7. The V9.0 Console corrects a problem whereby previous con-
soles would allow a >>>DEPOSIT/WCS or >>>DEPOSIT/SDF when
the machine was running or in an otherwise incorrect state.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 11[m
Other New Features:
1. There are two new sets of command files: PRIHALTxx.COM and
SECHALTxx.COM, where xx is the hex halt code. They are in-
voked instead of RESTAR.COM and LOADNBOOT.COM. All the
PRIHALTxx files contain @RESTAR.COM and all the SECHALTxx
files contain @LOADNBOOT.COM.
These files exist so that people (e.g., Field Service) can
modify them to collect specific information about specific
VAX HALT conditions (e.g., Double Error Halts).
2. The functions previously performed by three different task
images (CONSOLE1.TSK, CONSOLE2.TSK, and CONTROL.TSK) have
been merged. Only one task image, CONSOLE.TSK, is now re-
quired and this one task image supports all CPU configura-
tions. This also means less work to set up a STANDALONE
(diskette-based) configuration.
3. Similarly, the Console Command files have been merged so
there is no longer a set of command files unique to single
processor systems and a disjoint set of command files unique
to dual processor systems. All command files are now stored
in the [CONSOLE] directory.
4. Similarly, the naming scheme for the EMM data files and re-
vision history data files has been changed to allow all of
these files to reside in the [CONSOLE] directory.
5. The EMM data files are now optional. In the absence of any
of the data files, the Console Application will supply reas-
onable default data. Feature patches exist to modify this
default data to suit your particular requirements.
6. The EMM and revision history files now allow comments any-
where in the data file and contain a standardized header
similar to other Console ASCII files.
7. The revision history files are now stored in ASCII for
greater ease of maintenance. The data in the files has
been updated. >>>VERIFY REVISION xxx has been re-included
in SYSINIT.COM.
8. Console revisions will now be automatically sensed at the
time that the Console Application is invoked. The >>>SENSE
REVISION CONSOLE command is no longer required or allowed.
9. The presence, correctness, and revision of the various CPU
modules will now be automatically sensed at the time that
system power is applied. The revision data is now coupled
with the module presence data, resulting in better report-
ing. The >>>SENSE REVISION CPU and >>>VERIFY MODULE PLACE-
MENT commands are no longer required or allowed.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 12[m
10. Console power status display has been improved, in the form
of both the >>>SHOW POWER command and more informative mes-
sages output during power state transitions.
11. The default PRO/DCL editor has been changed from PROSE to
EDT. As has been true since Version V6, Release 29, a fea-
ture patch exists to allow the selection of either as your
default.
12. The software modules of the Console Application have been
reorganized among the various segments. This, in turn, al-
lowed the inclusion of patch space in all segments of the
Console task image, allowing more likelihood that it will
be possible to correct bugs via patches.
13. Every Console Application now includes, in human-readable
form, the exact date and time that the application was built
by Digital. This will aid in identifying sub-versions of
the Console Application.
14. Additional synonyms have been added to a variety of commands
to allow easier use of the Console.
15. The V9.0 Console now has a single, unified method of dis-
playing "Bugchecks" when certain internal consistency checks
fail within the Console Application. Previously, these
failures often went un-reported. Error checking within the
Console has been substantially increased as well.
16. The Console "automatic screen clear and terminal reset"
(first incorporated into V8.0) has been made more aware of
the actual state of the terminal, resulting in fewer auto-
matic screen clears.
17. The very long escape sequence used to clear the screen and
reset the terminal now supports hard-copy terminals with a
page length of 66 lines. A feature patch allows for the
support of other pages sizes (such as the various "A4"
sizes used outside of the United States).
18. A "logged-out" Console port now displays the prompt
"LOGIN>>>" rather than the standard Console prompt of ">>>".
19. Much as VAX/VMS DCL command lines can begin with a "$",
Console command lines can now begin with ">>>". Digital
recommends this style for use in command files.
20. VAX CPU Operational Microcode Version 12 has been provided.
21. The VMB from VAX/VMS Software V5.2, FT has been added.
[4m25-Apr-1989 CONSOLE RELEASE NOTES Section I, Page 13[m
20. If all of memory contains Single-Bit Errors, >>>FIND/MEMORY
will sucessfully find a 64K bank of memory but a warning
message will report the presence of single-bit errors.
21. For the REMOTE port, the MODEM will now default to DISABLED.
Previously, the default was ENABLED. With the increased
popularity of systems such as the VAXcluster Console System
(VCS), this change of default is appropriate.
22. The CONSOL task (CONSOLE.TSK) can now receive and display
unsolicited messages from any task (e.g., the LBS -- Logical
Block Server). Today, this capability is unused but a fu-
ture release of the LBS may use it for improved error re-
porting and software diagnosability.
23. The Console's use of RMS-11 I/O in general has been improv-
ed. The Console will now display an RMS error message for
any unexpected file I/O failure. The status message in-
cludes the operation attempted (e.g., '$CREATE', '$OPEN',
'$GET', etc.) and the RMS completion code returned.
24. Logfile creation has been improved:
- Free space is now measured before the creation of the
Logfile. Insufficient free space will abort Logfile
creation, and a warning message will be displayed if
a shorter-than-normal Logfile is being created.
- The Logfile is now created at the correct size. This
approximately halves the time required to initialize
the Logfile. The Logfile initialization progress mes-
sages have been improved as well.
- The Version V9.0, Release 36 Console is capable of "up-
grading" existing Logfiles in the V8.0 format (from any
Version V8.x Console) to the new V9.0 format, making it
unnecessary to create a new logfile when upgrading. No
tool exists to "fall-back", however.
- All newly created or upgraded Logfiles will have the
DESIRED_POWER_STATE set to "OFF". This is done to en-
sure that a full power-up must be performed after cre-
ating or upgrading the Logfile.
|
| 34.2 | CIBCA-BA Wrong uCode | KERNEL::SOWTON | Diagnosis does it down the phone.. | Mon May 15 1989 09:32 | 251 |
|
From: COMICS::SMITHERS "Glenn Smithers UK CSC - UVO 833 3372 - All things are difficult before they are easy 12-May-1989 2006" 12-MAY-1989 20:09:28.44
To: @NAUTI
CC: SMITHERS
Subj: Wrong version uCode for CIBCA-BA in diags 35
4.21 TITLE: WRONG UCODE FOR CIBCA-BA IN RELEASE 35
DEVICE: 85X0, 8700,8800,8810,8820-N,897X (Added: 10-MAY-1989)
CLD #: * PRISM #:
PROBLEM/DESCRIPTION:
VAX Diagnostic Release 35 saw the inadvertent inclusion of an incor-
rect version of microcode for the CIBCA-BA CI adapter in the various
Console software kits for PC38N VAX Console Subsystem. This microcode
will not allow the bootstrapping of VAX/VMS Software V5.0 or later via
the CIBCA-BA adapter. The first attachment to this details what
diskettes and software kits are affected.
STATUS:
A procedure is attached that will allow you to verify whether or not
you have the correct microcode. A third attachment details another
way to obtain the corrected microcode as well as how to correct other
affected diskettes.
SOLUTION/WORK-AROUND:
We can replace the Microcode file on your Console or Upgrade your console
to V9. V9 has been submitted to SDC for release 17-Jul-1989.
Replacing the Microcode file:
Console diskette NCON5 is available in an "Image" BACKUP at;
VOLKS::NAUTILUS$CONSOLE:CSAX_NCON5_V35.BCK. - Image Copy
or
Upgrading to Console Version 9:
Console Release 36, Console Version V9, and Diagnostic Release 36 will
be available at;
VOLKS::NAUTILUS$CONSOLE - CONSOLE VERSION V9 - Available Now
VOLKS::NAUTILUS$DIAGNOSTICS - DIAGNOSTICS RELEASE 36 - Available 15-May-89
/Attachments
- Attachment 1 -
Diskettes and Software Kits Affected
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
The incorrect file is CIBCB.BIN, internal version 4005.4003, and it
appears on the following diskettes:
o BL-FI48K-ME "RX98 CI VMB SYSTEM CODE", volume label "NCON5"
o BL-FK18D-ME "STANDALONE2 FOR DUAL"
o BL-FK19D-ME "STANDALONE2 FOR SINGLE"
The software kits affected include:
o 00-ZM920-C3.35 "VAX 8800/8700/8550/8530 DIAG"
o 00-ZM920-H3.35 "8800/8700/8550/8530 UPD DIAG"
o BT-Z5ACY-C3.35 "VAX CIBCA MICROCODE"
o BT-ZMAAD-C3.35 "VAX 8800/8700/8550/8500 CNSL"
VAX processors affected by this error include:
o VAX 8530
o VAX 8550
o VAX 8700 / 8810 / 8974 / 8978
o VAX 8800 / 8820-N
The corrected version of CIBCB.BIN (internal version 4005.4002) ap-
pears on the BL-FL02A-ME diskette. Use this diskette wherever the
installation procedure calls for the BL-FI48K-ME "NCON5" diskette.
The corrected file may also be copied to your existing BL-FK18D-ME
"STANDALONE2 FOR DUAL" and BL-FK19D-ME "STANDALONE2 FOR SINGLE"
diskettes.
- Attachment 2 -
Determining if you have the correct Microcode
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
The following procedure can be conducted with software running in the
VAX processor, however, you may choose to do it while the VAX proces-
sor is shut down. If you do this with VAX/VMS software running in
the VAX, please login to the VAX/VMS software system prior to exe-
cuting this procedure.
1. Press <^P> to get to CONSOLE mode.
2. At the Console prompt, issue the "EXIT" command.
3a. If you are determining the version of microcode on a diskette,
place the diskette in the upper (or leftmost) diskette drive
and close the door. At the PRO/DCL prompt, issue the follow-
ing command. A dump similar to that shown on the next page
will be printed.
DUMP DZ1:CIBCB.BIN /OUT=LP: /ASCII /BL:1:1
3a. If you are determining the version of microcode on the fixed
drive, then at the PRO/DCL prompt, issue the following command.
A dump similar to that shown on the next page will be printed.
DUMP CIBCB.BIN /OUT=LP: /ASCII /BL:1:1
4. The major version is contained at the beginning of the line
starting at address 000400. Both the correct and incorrect
versions are major version "4005".
5. In the line starting at address 000000, look for the four-
digit number after "CIBCB.BIN ". This is the minor version.
"4002" is correct. "4003" is incorrect.
6. At the PRO/DCL prompt, issue the "RUN CONTROL" command.
7. If software is running in the VAX, then, at the Console prompt,
issue the "SET TERMINAL PROGRAM" command to return to PROGRAM
mode.
Typical dump of the incorrect microcode
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Dump of LB0:[CONSOLE]CIBCB.BIN;1 - File ID 705,1,0
Virtual Block 0,000001 - Size 512. bytes
000000 C I B C B . B I N 4 0 0 3 C ^I
000020 X Y ^A ^@ ^@ ^@ _ ^_ ^@ ^@ ^C @ ^@ ^@ ^@ ^@
000040 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000060 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000100 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000120 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000140 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000160 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000200 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000220 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000240 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000260 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000300 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000320 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000340 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000360 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000400 4 0 0 5 ; h [ ^A ^Q ^@ ^@ ^@ ^@ 0 ^@ ^@
000420 ^E @ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000440 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000460 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000500 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000520 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000540 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000560 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000600 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000620 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000640 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000660 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000700 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000720 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000740 ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@ ^@
000760 ^@ ^@ ^@ ^@ ^@ ^@ C I B C B . B I N
- Attachment 3 -
Performing the upgrade using the CIBCB.BIN file
from one of the Licensed Diagnostic Kits
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Prior to performing the Console upgrade, prepare a diskette that
contains a correct copy of CIBCB.BIN. This procedure assumes that
VAX/VMS software is running in the VAX. SYSPRV and VOLPRO may
be required in order to execute this procedure. You must be
executing from an account that has a UIC which is less than [377,
377] octal. Most system accounts will suffice.
1. Copy the file(s) from VOLKS::NAUTILUS$CONSOLE.
2. Obtain a blank or scratch diskette and ensure that it does not
have a write-protect tab installed. Mount the blank diskette
in the ether diskette drive and close the diskette drive door.
3. Issue the following commands. These will create a copy of
the NCON5 diskette and mount it for further access.
MOUNT /NOASSIST /FOREIGN CSAx:
BACKUP /IMAGE /VERIFY filename.BCK/SAV CSAx
DISMOUNT /NOASSIST CSAx:
MOUNT /NOASSIST CSAx: NCON5
Note: x = Diskette Drive number (1 = Top/Left, 2 = Bottom/Right)
4. Changing the CI Microcode on the Console system Disk use the
following commands:
a. Exit USER Mode - Enter CONSOLE Mode
$ ^P
b. Exit CONSOLE Mode - Enter PRODCL Mode
>>> EXIT(ret)
c. Mount the new NCON5 diskette into either diskette drive.
d. Delete the old CI Microcode
PRODCL> DEL LB0:[USERFILES]CIBCB.BIN
PRODCL> DEL LB0:[USERFILES]CIBCB.BIN
e. Copy the new CI Microcode
PRODCL> COP/CONT DZx:[USERFILES]CIBCB.BIN LB0:[USERFILES]CIBCB.BIN
PRODCL> COP/CONT DZx:[CONSOLE]CIBCB.BIN LB0:[CONSOLE]CIBCB.BIN
f. Remove the diskette from the diskette drive.
g. Enter CONSOLE Mode
PRODCL> RUN [CONSOLE]CONTROL
h. Enter USER Mode
>>> SET TERM USER
5. The same correction can also be applied to both of the STAND-
ALONE2 diskettes, however, the microcode should only be copied
to the [USERFILES] directory, not to [CONSOLE].
5a. To change the CI Microcode on your STANDALONE2 diskettes issue
the following commands;
MOUNT /NOASSIST CSA1: STANDALONE2 - The old STANDALONE2 diskette
MOUNT /NOASSIST CSA2: NCON5 - New diskette with good Microcode
DELETE /VERIFY/LOG CSA1:[USERFILES]CIBCB.BIN
COPY CSA2:[USERFILES]CIBCB.BIN CSA1:[USERFILES]CIBCB.BIN
DISMOUNT /NOASSIST CSA1:
DISMOUNT /NOASSIST CSA2:
|
| 34.3 | NAUTILUS SID BITS. | KERNEL::MOUNTFORD | | Thu Jul 13 1989 17:04 | 57 |
| For ALL nautilus processors the cpu type field in the SID will be 06.
At this time this covers 8500, 8530, 8550, 8700, & 8800, all of which use
essentially the same cpu module set.
SID IPR 3E(H) System ID Register.
3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1
1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|0 0 0 0 0 1 1 0| | | system serial number (0-65535)|
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
^ ^ ^
| | `----- CPU Top revision (1-126.)
| | 127. (all 1's) = undefined, 0 = unavailable.
| |
| `---- Left/Right Processor bit (0= right; 1 = left)
| ---------------------------------------------
` CPU type (6 = Nautilus)
For the Nautilus family, more information about the CPU type can be
obtained with the lexical command `f$getsyi("XSID")', this returns a
longword with 2 bytes containing significant data. Remember that this
command is only meaningful for Nautilus processors, so make sure to check
the SID first.
The HIGH byte (bits 24-31) is coded by VMS as follows:-
00 = 8800, 01 = 8700, 06 = 8550, 07 = 8500/8530.
The LOW byte (bits 0-7) is supplied by the console as in the following tables.
on a DUAL PROCESSOR system, BIT 3 is ZERO
1 0
Bit 0: Left cpu available. available unavailable
Bit 1: Right cpu available. available unavailable
Bit 2: Secondary enable enabled disabled
Bit 3: Single/Dual processor -------- <ALWAYS FALSE>
Bit 4: DIAG key key ok key not ok
Bit 5: VMS Key key ok key not ok
Bit 6: Primary identity left cpu right cpu
Bit 7: Slow/fast "Slow" (8800) "Fast" (Unused)
on a SINGLE PROCESSOR system BIT 3 is ONE
1 0
Bit 0: Bounded 8500,8550 (bounded) 8700 (expandable)
Bit 1: Not used ------- <ALWAYS FALSE>
Bit 2: Not used ------- <ALWAYS FALSE>
Bit 3: Single/Dual processor <ALWAYS TRUE> --------
Bit 4: DIAG key key ok key not ok
Bit 5: VMS Key key ok key not ok
Bit 6: Not used ------- <ALWAYS TRUE>
Bit 7: Slow/fast "Slow" (8500,8700) "Fast" (8550)
First check bit 3 and use the appropriate table, bits 0 and 7 define the
system type, bits 4 and 5 are currently defined but not used, and bits 1 and 2
are only relevant on a dual processor system.
John Travell.
|
| 34.4 | | KERNEL::SOWTON | Diagnosis does it down the phone.. | Fri Aug 25 1989 01:28 | 253 |
| Gentlemen,
I hope this might be of some use to you all.
Regards
Glenn.
From: VOLKS::HARRIGAN "Jerry with a J 07-Feb-1989 1329" 7-FEB-1989 18:40
To: CXO_GROUP
Subj: I added some new part number include the new gold clean wipes.
I just updated the parts list and though that you may like the update. I have
the new part number for the gold cleans and backplane cleaner pads.
Jerry Harrigan
Support Tools
KEY, ZIF LOCK part number 7432127-02
(For backplane replacement, a special tool is required to open the ZIF
connectors. Replacement of the Main (CPU) backplane should only be done
with Support involvement.)
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
VAXBI Tool Kit - CD kit number A2-M1094-10
(which consists of part # 29-25608-00, the special tool assembly (this is
the torque wrench + bits)
The Hex bit is included in the VAXBI Tool Kit, part number A2-M1094-10.
This kit includes the following parts:
29-17381-00 - Torque Screwdriver 1-3
29-25609-00 - DIOFF PH SL SCRW DR
29-25610-00 - DIBIT, SL W/NOTCH .046"
29-25611-00 - DIBIT, PH 1 SL. .048" **** USE THIS BIT
29-25612-00 - DIEXT. 6" (TORQUE SCRW DR)
29-25613-00 - DIADAPTER - HEX -> 1/4" SCR
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
The part number for the T-20 Torx bit for use on the Nautilus backplane is
29-27457-01. I checked the part number on APPIX and it's being stocked in
SR17.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
MCA Extraction Tool - (1) 29-25915-00 EXTRACTION TOOL - or EQ-01480-02
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Backplane Pin Insertion Tool:
Includes the following: 29-26229-00 = Winchester Electric 107-42054 handle
29-26230-00 = " " 107-42528 tip
Backplane Pins: 12-18518-03 - 22 pins in a package
- 25 pieces equals 1 whole box
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Connector cleaning kit, 4 segment ZIF; consisting of:
Cleaner Extractor handle - 47-00116-02
Cleaning pad, chemically impregnated - 12-26321-06 - NEW Part - USE
Cleaning pad, dry (inactive) - 12-26321-04
Connector cleaning kit, 5 segment ZIF; consisting of:
Cleaner extractor handle - 47-00116-01
Cleaning pad, chemically impregnated - 12-26321-05 - NEW PART - USE
Cleaning pad, dry (inactive) - 12-26321-02
Gloves, Solvex - 29-26403-00
Gold Clean Wipes - 49-01013-00 Replaced Gold Wipes
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Backplanes and Card Guides
CPU:
8530 CPU Backplane - 70-23573-01
8550 CPU Backplane - 70-23098-01
8800 CPU Backplane - 70-21696-02
Top Card Guide - 70-21619-01
Top Card Guide - 70-22918-01 - New style
Bottom Card Guide - 70-21619-02
Bottom Card Guide - 70-22918-02 - New style
Memory:
5 Slot
MAR4 Memory Backplane - 70-23099-01
MAR16 memory backplane - 70-24626-01
Top Card Guide - 74-34023-01
Bottom Card Guide - 70-23950-01
Install the new memory - EK-OMS88-IN
8 Slot
Backplane - 70-21701-01
Top Card Guide - 70-22468-01
Bottom Card Guide - 70-22468-01
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Polarstar Parts
Console (UVAX) - 6300PS-A2 - Complete Console for Pstar
QC module - M7553 - Uvax interface to the big VAX
CIM - F1032 - Big VAX interface to the UVAX
Cable - Console Bus - 17-01545-01
Polarstar console media
BT-ZL009-C5.6.0 VAX 8820/8830/8840 Console
Includes: Console only release notes (aa-fj78d-te)
Console User's guide (aa-fj95a-te)
Console software TK50 (aq-fj80d-me)
Cnsl diags = microVAX diags plus QC diags (aq-fk23b-me)
The info. I have says that the console/diags notes is also part
of this kit...they really shouldn't be and I'll get it changed for
the next release.
00-ZA900-HW.6.0 - VAX 88XX F.S. DIAG CMPLT
00-ZA900-WW.6.0 - VAX 88XX F.S. DIAG UPDATE
Includes: Console only release notes (aa-fj78d-te)
Console/diagnostics release notes (aa-fk01d-te)
Console/diagnostics TK50 (aq-fj79d-de)
Console software TK50 (aq-fj80d-me)
Cnsl diags = microVAX diags plus QC (aq-fk23b-me)
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
82x0/83x0
Label, Words - 36-23595-01 - (Power Switch - On/Off, i. e.)
Label, Symbols - 36-23595-02 - " "
8250 Logo
8200 Style cabinet - 74-30800-04 -
Antares Style cabinet - 74-33826-03 -
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
BI Parts
Near End Terminator - 20-24486-01
Far End Terminator - 20-24487-01
Flexible Backplane Ext.- 17-01038-01
BI Cable, bottom 3.5' - 70-21692-01
BI Cable, top 3.5' - 70-21693-01
BI Cable, bottom 15' - 70-21694-01
BI Cable, top 15' - 70-21695-01
Card Cage - H9400-Ax
Guide Post - 12-25003-02(ECO'ed)
BI Load Board - 70-23417-02
Mounting hardware
Tinnerman Nuts (4) - 90-07786-00
Standoffs (4) - 90-00001-21
Screws (PTH) (4) - 90-09700-00
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Miscallaneours Parts
Logo # "5" - 70-31904-06
KLESI-B Cables:
5 feet - 17-01098-05
8 feet - 17-01098-04
15 feet - 17-01098-06
Nautilus(8800) Power Connector - 12-26148-02
Console Outlet, Bulkhead ass'y - 70-22532-01 - single outlet
Console Outlet, Bulkhead ass'y - 70-22532-02 - double outlet
Jumps, 02POS, Push on - 12-23353-01
Door Latch grip range - 12-24095-01
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
85xx BBU info
8530/8550 System Battery Backup Installation Guide, part number EK-85BBU-MA
The EK-85BBU-MA manual also has a update pkg associated with it. I would assume
that one would receive this update pkg with each order of a EK-85BBU-MA, but
just in case the update pkg p/n is as follows : EK-85BBU-MA-001CN (as extracted
from first pg). It contains a large number of corrections to the original
manual.
Extract from pg 1-2 of EK-85BBU-MA (table 1-1)
BBU-Ready System Components :
Component Part Number Mininum Rev
MPS Backplane 54-16709-01 C
SPC backplane 54-16787-01 B
SPC Backplane
shield 74-32844-01 B
Memory Backplane 70-24626-01 A
+5V/gnd Bus bar 17-01507-01 B
H405 UCI H405-A (60hz) C
H405-B (50Hz) B
Power Controller H7176-A D
Cpu voltage/sense
cable 17-01210-01 E
power interface
cable 17-01536-01 (60hz) B
17-01537-01 (50hz) A
o If the parts in the system do not match above table, then the
system is NOT BBU READY.
Option kit part numbers :
BBU-ready system (60 or 50 hz) H7232-AA
NOT BBU-ready system (60 Hz) H7232-UA
NOT BBU-ready system (50 Hz) H7232-UB
Came across some additional info with regards to Battery Backup on
Nautilus systems ...
1) On 85XX systems, there has been a recent FCO to the H7176-A to
correct a problem where this regulator was missing a FSE signal
required for battery backup. The H7176 should be a mininum rev
of "D2". FCO kit is EQ-01508-01.
Reference Speed Bulletin #523 under A2-W1154-10-I001 (CD kit FCO)
2) There is also another fco for Nautilus systems on the DWBUA.
It must be a mininum rev of "H1" or the the system "may" get
a fatal bugcheck during a warm restart. T1010 should be mininum
of "H1". FCO kit is EQ-01509-01.
Reference Speed Bulletin #528 under DWBUA-I-002.
|
| 34.5 | | KERNEL::SOWTON | Diagnosis does it down the phone.. | Fri Aug 25 1989 01:29 | 54 |
| Gentlemen,
As you are all aware by know Nautilus Rev 9 console is with us along
with Diag Rev 36.
However, this is not such a major release and consequently, I would
not recommend upgrading to rev 9 if you are already at rev 8 or greater, as
there is very little to be gained.
For those of you that do wish to upgrade, the Backup Savesets are
on COMICS:: in SYS$PUBLIC as listed below.
Before you all rush off and start making generating new floppys,
I should like to pass on some info from Brent and Graham in the City who
have already been through this exercise.
Firstly NDIAG9 now contains an additional diagnostic EZKBE.WC1/BIN
and consequently makes this floppy very full along with NDIAG8.
So if you have any bad blocks on your floppys then the chances are
that these will not fit. Also you must make sure that you perform the
Backup using the /IMAGE qualifier, as any floppy formatted and initialised
on a PRO380 ends up with less blocks than are required; but the /IMAGE
will overcome this for you.
Also I must point out a new "Feature" of rev 9 which Kelvin has
found. This only effects a system that has powered failed and either did
not have Battery Backup or has exceeded its support cycle. This will then
cause the console to invoke a "Cold Boot", the only problem is, that on
successful completion the Console does not go into Prog IO Mode but instead
stays in Console Mode. This does not impair VMS coming up and running, but
of course could cause problems with OPCOM messages etc. This is currently being
worked on and you can follow its progress in the Notes file, entry 946.
Good luck,
Regards
Glenn.
Directory DISK$USERS2:[PUBLIC.NAUTILUS_CURR_CONSOLE]
CONSOLEDCL1.BCK;1 CONSOLEDCL2.BCK;1 CSAX_NCON5_V35.BCK;1
KITCONT.36;1 NCON1.BCK;2 NCON1A.BCK;1 NCON2.BCK;1
NCON3.BCK;1 NCON4.BCK;1 NCON5.BCK;1 NDIAG4.BCK;1
READ_ME.TXT;1 RELEASE.36;1 RELEASE.INFO;2 SANOTES.36;1
STANDALONE1.BCK;1 STANDALONE2.BCK;1
Total of 17 files.
Directory DISK$USERS2:[PUBLIC.NAUTI_DIAGS]
CLKTESTA.TSK;1 FLOPPY_DIRS_V36.LIS;2 LISTINGS.DIR;1
NDIAG1.BCK;1 NDIAG10.BCK;1 NDIAG11.BCK;2 NDIAG12.BCK;1
NDIAG13.BCK;1 NDIAG14.BCK;1 NDIAG15.BCK;1 NDIAG16.BCK;1
NDIAG17.BCK;2 NDIAG18.BCK;1 NDIAG19.BCK;2 NDIAG2.BCK;1
NDIAG20.BCK;2 NDIAG21.BCK;1 NDIAG22.BCK;2 NDIAG23.BCK;1
NDIAG3.BCK;1 NDIAG4.BCK;1 NDIAG5.BCK;1 NDIAG6.BCK;1
NDIAG7.BCK;1 NDIAG8.BCK;1 NDIAG9.BCK;1 RPR1.BCK;3
RPR2.BCK;2 RPR3.BCK;2 V35.DIR;1
Total of 30 files.
|
| 34.6 | | KERNEL::SOWTON | Diagnosis does it down the phone.. | Fri Aug 25 1989 01:30 | 40 |
|
Subj: 85XBA-F010/871BA-F005/882BA-F007 FCOs ON HOLD!
Please find attached notification from CSSE of above FCOs that have
been put on hold.
V36 of console microcode should not be installed.
Best regards,
Vita
********************************************************************************
TO: Distribution DATE: 23 AUG 1989
FROM: Bonnie Gayton
DEPT: MID-VAX CSSE
LIFE CYCLE PLANNING
EXT: 240-6487
M/S: AET1-1/6
NET: VOLKS::Gayton
SUBJ: V36 Console FCOs on hold
85XBA-F010
871BA-F005
882BA-F007
Due to problems that have occurred recently with V36 Console Code V9.0,
we have decided to place these FCOs on hold. The corrected code, V9.0D
will be available on 01-SEP over the ENET on VOLKS. Release 38,
scheduled for 15-JAN-90, from the Software Distribution Center (SDC)
via the Automatic Release Cycle will contain V9.1 of the console code.
Therefore, for the above FCOs it is recommended that you install either
V9.0D or V9.1 and not use V9.0 as stated in the CIP and FA documents.
|
| 34.7 | MEMCHECK | KERNEL::BLAND | toward 2000 ... | Sat Oct 21 1989 12:01 | 292 |
| ***** The following information has been extracted from a
***** Saveset (MEMCHECK001.A), on tape # 2 of VAXPAX release
***** # 37. Note this VAXPAX release has not yet been officially
***** released (very soon), however some of us have accesss.
** This is the HELP file **
1 MEMCHECK
A utility used to determine what percentage of memory is interleaved
given a particular memory configuration for the 85XX/87XX/88XX systems.
Format:
MEMCHECK mar-size-spec [,...]
Parameters:
mar-size-spec specifies the sizes of the memory arrays to be
interleaved. MEMCHECK accepts as input a comma separated list
of up to eight parameters. The valid values for the memory arrays
are [64,32,16,8,4]MB. NOTE: DEC does not sell 32MB or 8MB memory
arrays for the 85XX/87XX/88XX systems.
2 Command_Qualifiers
/FULL
Requests a full mapping of the results of the interleaving algorithm
be displayed. The default for the utility is /NOFULL.
/NOFULL
Requests a brief form of the output of the utility. The default for
the utility is /NOFULL.
2 Examples
1. $ MEMCHECK 16,4,4,4
85% interleaved; 24 MB interleaved of 28 MB total
4 MB of board A (16 MB) not interleaved
The MEMCHECK command uses the default /NOFULL format. The full mapping
of the results of the interleaving algorithm is not displayed.
2. $ MEMCHECK/FULL 16,4,4,4
85% interleaved; 24 MB interleaved of 28 MB total
4 MB of board A (16 MB) not interleaved
Board Chunk # Interleaves with Board Chunk #
C 1 A 1
A 2 C 2
D 1 A 3
A 6 D 2
B 1 A 7
A 8 B 2
The MEMCHECK command with the /FULL qualifier displays The full
mapping of the results of the interleaving algorithm.
2 Hardware Requirements
MEMCHECK is a non-privileged, user-mode software utility that
runs on any VAX/VMS processor. The installation of the specified
configuration or the use of a 85XX/87XX/88XX system is not required.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
MEMCHECK V1.0 Release Notes
Memory interleaving checking utility
1 Introduction
MEMCHECK is a program used to determine how much memory
is interleaved given a particular memory configuration, for
the 85XX/87XX/88XX systems. The higher the percentage of
interleaved memory, the greater the degree of performance the
system will exhibit. It is therefore a goal to try and configure
a system in such a way to achieve 100% interleaving of memory.
This utility is intended for Sales Support personnel, Field Service
representatives, CSC, and RDC organizations. The utility will
run in user mode on any VAX processor. The use of a 85XX/87XX/88XX
system is not required.
This program is an adaptation from the 88XX console code.
2 Installation
2.1 Installation Requirements
The installation should take approximately 3 to 5 minutes,
depending on the type of media and your system configuration.
Before installing MEMCHECK, you must have the following privileges
and resources:
o Minimum privileges for software installations.
o A minimum of 1000 blocks of free disk space during
installation; 100 blocks are used after the installation.
o The VAX/VMS Linker must reside on the system.
o Any version of VAX/VMS V4.0 or higher.
2.2 Installation Procedure
The installation is accomplished by logging into the SYSTEM
account and invoking VMSINSTAL. No other commands are required
to establish the environment. The VMSINSTAL procedure is invoked
as follows:
@SYS$UPDATE:VMSINSTAL MEMCHECK001 [device]
During the installation the MEMCHECK command is added to the DCL
command table. The MEMCHECK.OBJ and MEMCHECK_MSG.OBJ are linked
to create the MEMCHECK.EXE and this file is placed in the
SYS$SYSROOT:[SYSEXE] directory. The MEMCHECK.HLP file is inserted
into the DCL help library.
2.2.1 Side Effects
The installation of this utility will cause the directory
SYS$SYSROOT:[SYSTEST.MEMCHECK] to be created. The following files
will reside in this directory.
MEMCHECK_MSG.MSG the message source file
MEMCHECK_MSG.OBJ the message object file
MEMCHECK.OBJ the utility object file
MEMCHECK_IVP.COM the Installation Verification Procedure
2.3 The Installation Verification Procedure (IVP)
During the IVP process a temporary file MEMCHECK_IVP.LOG is created.
If the installation is successful this file will be deleted. However
if the IVP process fails the MEMCHECK_IVP.LOG will be typed to the screen
along with an error message. In the case of a failure, the MEMCHECK_IVP.LOG
will not be deleted and will reside in SYS$SYSROOT:[SYSTEST.MEMCHECK].
The following is an example of a successful MEMCHECK_IVP.LOG
Example of a successful MEMCHECK_IVP.LOG
$!
$ SET NOON
$ OK_COUNT = 0
$!
$! This test on $STATUS should be successful
$ MEMCHECK 64,64
100% interleaved; 128 MB interleaved of 128 MB total
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$! This test on $STATUS should fail
$ MEMCHECK 4,5
Invalid memory array size \5\
Re-enter with 64,32,16,8,4
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$! This test on $STATUS should be successful
$ MEMCHECK 4,16,16
100% interleaved; 36 MB interleaved of 36 MB total
** Configuration Summary **
The following configuration violations were noted:
- Hardware installation requires the boards to be in descending order.
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$! This test on $SEVERITY should be a warning
$ MEMCHECK/FULL 4,16
40% interleaved; 8 MB interleaved of 20 MB total
12 MB of board B (16 MB) not interleaved
** Configuration Summary **
The following configuration violations were noted:
- Hardware installation requires the boards to be in descending order.
Board Chunk # Interleaves with Board Chunk #
A 1 B 1
B 8 A 2
$ IF ($SEVERITY .EQ. 0) THEN OK_COUNT = OK_COUNT + 1
$!
$! This test on $STATUS should fail
$ MEMCHECK 4,4,4,74293
Argument [4] too large or negative \74293\
Re-enter with 64,32,16,8,4
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$! This test on $STATUS should be successful
$ MEMCHECK 4,64,64,16,16
100% interleaved; 164 MB interleaved of 164 MB total
** Configuration Summary **
The following configuration violations were noted:
- Hardware installation requires the boards to be in descending order.
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$!
$! This test on $STATUS should be successful
$ MEMCHECK 16,0,16,0,32,0,32
100% interleaved; 96 MB interleaved of 96 MB total
** Configuration Summary **
The following configuration violations were noted:
- Configuration is valid only for VAX 87XX/88XX systems.
- DEC Does not sell 8MB or 32MB memory for VAX 85XX/87XX/88XX systems.
- Hardware installation requires the boards to be in descending order.
- Systems with "narrow space" backplanes should be "expanded".
Do not skip slots.
$ IF $STATUS THEN OK_COUNT = OK_COUNT + 1
$!
$ VFY_FLAG = F$VERIFY(VFY_FLAG)
3 MEMCHECK COMMAND
Format:
MEMCHECK mar-size-spec [,...]
Parameters:
mar-size-spec specifies the sizes of the memory arrays to be
interleaved. MEMCHECK accepts as input a comma separated list
of up to eight parameters. The acceptable values for the memory
arrays are [64, 32, 16, 8, 4]MB. The memory arrays must be entered
in descending order. NOTE: DEC does not sell 32MB or 8MB memory
arrays for the 85XX/87XX/88XX systems.
3.1 Command Qualifiers
/FULL
Requests a full mapping of the results of the interleaving algorithm
be displayed. The default for the utility is /NOFULL.
/NOFULL
Requests a brief form of the output of the utility. The default for
the utility is /NOFULL.
4 Changing the Language of The Messages
4.1 Editing the MEMCHECK_MSG.MSG
The MEMCHECK utility uses $GETMSG, $FAO, and LIB$PUT_OUTPUT for
all output. Therefore, the MEMCHECK utility executeable must be linked
with the MEMCHECK_MSG message file in order to define the message codes.
It may be desirable to change the native language of the output text
for the utility. This is accomplished by editing the MEMCHECK_MSG.MSG
source file and translating the text messages from one language to
another. Keep in mind the order of the parameters for the FAO directives
are defined by the source code. The next step would be to generate the
new MEMCHECK_MSG.OBJ file by compiling it using the Message Utility.
Ex. MESSAGE MEMCHECK_MSG.MSG
4.2 Linking MEMCHECK.EXE
The next step would be to link the newly created MEMCHECK_MSG.OBJ
file with the MEMCHECK.OBJ and replacing the old SYS$SYSROOT:[SYSEXE]
MEMCHECK.EXE with the the new language version.
5 Known Problems and Restrictions
No known problems or restrictions.
Regards, Norman Bland
|
| 34.8 | EZKBC failure | KERNEL::WRIGHTON | Pass me a +L-14005 | Mon Nov 13 1989 18:36 | 52 |
|
Chaps ,
Just a quicky to warn you about an error callout from
one of the NAUTILUS microdiagnostics .
In theory , the micros work in such a way as to test a chunk of
system and then use that bit to test another bit and so on and
so on . In thoery no untested hardware is used to test another
part of the system . Most of the time this works ok ( the call-
outs are pretty accurate ) , unfortunately there are a few tests
that don't conform to these rules , one of them is EZKBC . The
following example failure was actually caused by a bad NBIB
( T1020 ) that has not been verified at the point at which this
test runs .
What the diagnostic does is set bit 19 in the NBIA . This bit
is called the "FLIP 29/22 bit" . This bit is used by diagnostics
to make the NBIA initiate reads and writes to/from memory .
Any address directed at the I/O must have bit 29 asserted .
When the FLIP bit is set , the NBIA responds to the address
and then swaps bits 29 and 22 over ( bit 22 is set to ensure
good parity ) . This new address is passed to the NBIB which
sends it out over the BI , but of course nothing can respond
so the address comes back to the NBIB which recognises it as
a memory address , hence initiating the memory cycle . In this
particular test , memory has been set up to cause a timeout
but the diagnostic actually fails before even sending the address
to memory . At this point the only test that the NBIB has done
is its selftest ( not very extensive , I think it just
makes sure it can turn the "pass" light on ) . There are only
two callouts in the diag listings , the NBIA and the SEQ modules .
In the case below the problem was fixed by the NBIB but could
due to virtually any BI related problem ( terminators/cables
etc etc )
TEST 0A - NBI No Access - No Response Timeout
11-NOV-89 05:02:06, DIAGNOSTIC DETECTED ERROR
*** CPU FAILED *** EZKBC - REV 07.00 *** TEST 0A - ERROR 01, PASS 01 ***
FAILING HW: NBIA0
Dave ( I_love_fiche ) W
|
| 34.9 | Nautilus console 9.4 | KERNEL::WRIGHTON | Pass me a +L-14005 | Thu Jan 18 1990 20:26 | 209 |
|
This is Field Test only ( so far ) , might be of interest ....
================================================================================
Note 1068.0 Announcing Nautilus Console V9.4 Field Test 1 reply
PROXY::SCHMIDT "Thinking globally, acting locally!" 198 lines 17-JAN-1990 17:02
--------------------------------------------------------------------------------
17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 5
================================================================================
FEATURES and BUG FIXES for VERSION V9.4 / VAX RELEASE 40
--------------------------------------------------------------------------------
Bug fixes:
1. Previous consoles contained an error which, under certain
circumstances, allowed the LOCAL stream proportionally more
computing time than either the REMOTE or REBOOT streams.
This error has been corrected.
2. Previous consoles could crash if certain errors arose
while writing the Logfile. In addition, these errors
could corrupt the Logfile, making it unusable. These
errors are now corrected. In addition, many instances
of Logfile corruption will now be explicitly detected
when the Console is first started.
3. On dual-processor systems, loading of the Decode RAM or
Cache Control Store of one CPU corrupted the state of
the other CPU. This was due to an error in the RTDRV.TSK
driver task. This error is corrected with the Version 10
RTDRV.TSK shipped as part of this release.
4. Previous consoles contained a minor error in the Logfile
initialization routine. This error caused the pattern
initially written into the Logfile (at the time of the
file's creation) to be slightly different than expected.
This error is now corrected, and Logfiles newly created
with using Version 9.4 / Release 40 will have the correct
initial data pattern. This bug is entirely harmless --
there is *NO NEED* to recreate the Logfile merely to
correct this initial data pattern.
5. If previous versions of the Console were unable to read
the EMM Firmware Revision, they incorrectly reported:
"Can't read RTI driver revision". The Console now cor-
rectly reports "Can't read EMM firmware revision".
6. Several different problems dealing with Air-Flow Faults
have been corrected.
7. The timer used during >>>POWER OFF has been extended again,
to 10 seconds. This should eliminate any marginal situations
where systems with very lightly loaded memory power supply
regulators were reporting "Power not set to desired state".
8. REVR1 (a processor internal register) was being loaded
with incorrect data. This is now corrected.
9. The >>>SHOW LOGFILE command will now allow <XOFF> and <XON>
as input characters. Previously, slow terminals could cause
an infinite loop of "Invalid command..." error messages as the
terminals attempted to use <XOFF> and <XON> for flow control.
[4m17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 6[m
10. If the PC38N recovered from a power failure (or brownout con-
dition) before the main cabinet of the CPU Complex, previous
versions of the Console would occasionally fail to locate
the address of the (not-yet-powered) EMM and thereby failed
to ever establish a connection with the EMM. Correcting
this condition required exitting and re-entering the Console
application. This version of the Console now uses its saved
EMM address if the EMM can't be found at startup time.
[4m17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 7[m
Modified Commands and Features:
1. The >>>POWER ON command will now check for and immediately
report failures to establish correct flow of cooling air,
followed by an automatic reversion to >>>POWER OFF.
2. In order to provide a more certain Remote Diagnosis en-
vironment, more commands are disabled when the >>>DISABLE
LOCAL CONSOLE command is issued from the REMOT port. Now,
any command that can change the "global" Console environment
is disabled.
3. >>>SHOW LOGFILE is substantially modified. In brief:
- Simple escape sequences captured in the Logfile (used
for bold, underline, etc.) will now frequently be acted
upon. Previously, all escape characters were converted
to '$' making it impossible to act upon the escape se-
quences.
- A 132-column mode is now available. Press <^T> to
toggle between the 80-column and 132-column displays.
- The command is able to make a much better determination
of when to display the "diamond" that indicates that a
line has been truncated at the right margin.
- The command does a better job of diplaying the [START
OF LOGFILE] and [END OF LOGFILE] tags, as well as the
most recent data logged by active Consoles.
- Lines longer than 512 to 1023 bytes in length are now
no longer displayed. (Previously, the first 80 or so
bytes were displayed.) [ %%% LINE TOO LONG TO DISPLAY
%%% ] is substituted for each line that is too long.
- Imbedded Carriage-Returns (<CR>s) are now displayed as
the reverse-question mark or "splat" character. Pre-
viously, these were hidden. The <CR> that frequently
appears at the end of a line is still hidden so as not
to clutter up the display.
4. Field-Test versions of the Console now contain an expiration
date and a simple enforcement mechanism. Various >>>SHOW xxx
commands have been modified to display the expiration date if
any expiration date is specified.
5. The >>>VERIFY EMM command is now able to locate the address
of an EMM. It will also report any change of EMM address.
6. During initialization, the Console will now report any
change of EMM address from whatever address was previously
noted.
[4m17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 8[m
7. The automatic power subsystem will now explicitly report and
handle failures to re-establish full Console control over the
power system after a momentary power outage switches the system
to dc STANDBY ("battery") mode and then to ac STANDBY mode.
8. The automatic power subsystem will now take less than 10
seconds to react to and report the clearing of Air Flow
Faults. Previously, this could take as long as 5 minutes.
9. The RMS "Record too big" error is now explicitly reported,
rather than just displaying a numerical error code. The
new explicit report is:
RMS$_RTB, Record too big
[4m17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 9[m
New Features:
1. The >>>CONVERT LOGFILE command allows the conversion of
the Logfile (or portions thereof) to a normal RMS11
sequential file. The full syntax of the command is:
>>>CONVERT [LOGFILE] [/BLOCKS:n]
If the /BLOCKS:n qualifier is ommitted, the Console will
attempt to convert the entire logfile. If the /BLOCKS:n
qualifier is included, only the most recent n blocks will
be converted. Please note that the Console's throughput
is substantially reduced while the conversion is taking
place.
2. The >>>SET PRINTER 80 [COLUMNS] and >>>SET PRINTER 132
[COLUMNS] commands have been included. If the printer
is available and enabled, these send the appropriate
escape sequences to switch the printer between 10 char-
acter-per-inch and 16.5 character-per-inch printing.
The result of these commands is *ONLY MAINTAINED BY THE
PRINTER* and *MAY NOT BE MAINTAINED* across multiple
invocations of the Console, power failures, other use
of the printer, etc.
[4m17-Jan-1990 CONSOLE RELEASE NOTES Section I, Page 10[m
Restrictions:
1. All RMS11 files read by the Console can not contain lines
longer than 120 bytes. Files read by the Console using
RMS11 include:
- *.COM indirect command files,
- *.HLP help files,
- *.DAT EMM and memory configuration data files, and
- Any file displayed by the >>>TYPE command
This restriction has existed for all versions of the Console
and will continue to exist for the foreseeable future.
|
| 34.10 | PMCL FCO | KERNEL::MOUNTFORD | | Fri Jan 19 1990 10:34 | 38 |
| <<< KERNEL::DISK$APD1:[NOTES$LIBRARY]CSGUK_SYSTEMS.NOTE;1 >>>
-< CSGUK_SYSTEMS >-
================================================================================
Note 14.3 VAXCLUSTER FCO's. 3 of 4
KERNEL::MOUNTFORD 30 lines 18-JAN-1990 12:37
-< PMCL MODULES >-
--------------------------------------------------------------------------------
From: KERNEL::WIBREW "18-Jan-1990 1159" 18-JAN-1990 12:16:15.84
To: @POST:SYSTEMS
CC:
Subj: Nautilus PMCL modules
There is a world shotage of PMCL modules mainly due to the fact
that engineers are installing them in place of the MCL module in
Nautilus systems.
The upgrade from MCL to PMCL is an FCO.
Logistics stock the PMCL module for maintenance repair not for
upgrades or FCO's
If a PMCL module is required for the FCO then it should only be
ordered by the FCO number and NOT the module part number. This
should help improve the availabilty of the module when it is required
to replace a faulty one.
If the FCO is required, then order the following:-
for 8530/8550 85XBA-I008 EQ-01518-01
8700 871BA-I004 EQ-01518-01
8800 882BA-I006 EQ-01518-01
Regards,
Steve W.
|
| 34.11 | PMCL FCO | KERNEL::BLAND | toward 2000 ... | Fri Jan 19 1990 18:57 | 8 |
| Ref .10
The FCO states that the PMCL FCO should only be done on systems
with at least one 64Mb memory module (MS88-DA). The shortage may
have come about because the FCO is being done on systems with 16Mb
modules only.
NB: The SMU callout is correct for both 16 & 64Mb modules with a
PMCL module installed.
|
| 34.12 | NAUTILUS new config rules | KERNEL::BLAND | toward 2000 ... | Tue Feb 06 1990 03:21 | 174 |
| ================================================================================
Note 9.0 TT Nautilus BI I/O Config Rules No replies
CSC32::J_MACGOWN "PS-HW>BOOT/R5:10 Colorado_spring" 169 lines 5-JAN-1990 03:31
--------------------------------------------------------------------------------
Subj: VAX BI I/O device configuration rules
TITLE: NEW BI CONFIGURATION RULES DATE: 15-Dec-1989
AUTHOR: Jerry Harrigan TD #: 000156
DTN: 240 6494
ENET: VOLKS::HARRIGAN CROSS REFERENCE #'s:
DEPARTMENT: CSSE MID-RANGE SUPPORT (PRISM/TIME/CLD#'s)
INTENDED AUDIENCE: ALL PRIORITY LEVEL: 2
(U.S./EUROPE/GIA) (1=TIME CRITICAL,
2=NON-TIME CRITICAL)
DEVICES: 85x0, 8700, 8800, 8810, 8820-N, 897x
=====================================================================
VAX_BI I/O DEVICE CONFIGURATION RULES
=====================================
*** SUPERCEDES ALL OTHER 85/87/88xx CONFIG. RULES !! ****
PROBLEM:
On NBI-based systems (88X0, 85/87/8800), I/O device
interlocked-memory operation timeouts (OR severe latencies) can
occur under heavy contention from CPU-to-BI CSR-accesses
resident on same BI. This is due to NBI-adapter design
quirk allowing CPU priority over all DMA devices, although
allowing single DMA (by potential DMA-HOG) between CPU
transactions to BI.
DEBNA DATA-OVERRUNS have been observed on 1 8700 site, with
single BI containing DMB32, CIXXX, and DEBNA !!
RESOLUTION/WORKAROUND:
BI-DEVICE CATEGORIES
====================
BCR: BI Category "RETRY" (to NBIB) Devices = DWBUA
BCC: BI Category "CSRhog" Devices = DRB32, DMB32, DHB32, KLESI-B,
MS820......CPU CSR-access intensive
BCI: BI Category "Interlock" devices: DEBNx, TBK50, KA/MS800....
Memory-interlock intensive, but few CSR-accesses
BCB: BI Category "Both" (BCI & BCC) devices = CIBCI, CIBCA-A/B
BCS: BI Category "Safe" devices = KDB50, KA82x....Few interlocks
or CPU-CSR accesses.
***** DIGITAL Internal Use Only *****
"SOLID-GOLD" CONFIGURATION RULE -- INTERLOCK LATENCIES AVOIDED --
===============================
1. NBIB always at NODE-ID "0".
2. Only one (1) "BCR" device (DWBUA) per BI-BUS, NODE-ID = 2.
3. Only one (1) "BCB" or one (1) "BCI" device per BI-BUS, with
NODE-ID higher than NBIB.
4. Any quantity of "BCS" devices on any BI-BUS.
"FIRM-BRASS" CONFIGURATION RULE -- MINIMUM INTERLOCK RISKS --
===============================
NOTE: These rules are less restrictive, but with
greater risks.
1./2./4. RULES 1, 2, & 4 SAME AS "SOLID GOLD", ABOVE.
3A. ALLOWED MIX OF "BCI" & "BCB" DEVICES ON SINGLE BI-BUS:
+ 1-BCI & 1-BCB, WITH BCI NODE-ID < BCB NODE-ID, NO-BCC;
+ 3-BCI & NO-BCB & NO-BCC;
+ 2-BCB & NO-BCI & NO-BCC.
3B. ALLOWED MIX OF "BCC" WITH "BCI or BCB" DEVICES ON SINGLE BI:
+ NO-BCC WITH >1-BCI/BCB (TOTAL, ANY MIX);
+ ANY-QTY-BCC WITH 1-BCI OR 1-BCB OR NO-BCI/BCB.
NOTE: Special rules for "VAX RTA" KA/MS800/82x Systems: Refer
to SASE::NAUTILUS NOTE # 596.2 FOR FURTHER INFO.
ADDITIONAL COMMENTS:
HIGHLIGHTS/IMPACT OF NEW BI CONFIGURATION RULES
===============================================
- "XCON" implemented in Fall-1988;
- VAX-62/63/64xx Does not have similar interlock problem,
since DWMBA uses fair CPU/BI-ROUND-ROBIN Arbitration Rules;
- VAX-82/83xx avoids interlock problems of INTER-BUS Adapters;
- More than 2-BCI/BCB per BI, or mix of BCC with BCI/BCB per BI
increases risk of VMS/CPU crashes due to Interlock Timeouts:
62/63/64xx are only safe CPU'S for such violations.
- 85/87/88xx, in addition to Timeout crashes, can experience
degraded SMP-CPU and INTERLOCKING-I/O device performance
with heavy interlock traffic due to "ONLY ONE PRIMARY
MEMORY INTERLOCK" PER-SYSTEM.
- DWBUA NODE-ID # irrelevant if only BI-DEVICE on NBIB;
But if mixed with other BI-DEVICES, DWBUA must be node #0 !
- DWBUA mixed with other BI-DEVICES may see SSYN-TIMEOUT,
RETRY-TIMEOUT (RTO), HUNG UNIBUS, OR CSR-NXM.
- 85/87/88xx Systems without errors with mixed DWBUA & other
BI-DEVICES on one (1) NBIB should not be reconfigured.
wide variance of UNIBUS-TO-MEMORY Latencies, Applications
Workloads, and device Latency-Tolerance allow many configs
***** DIGITAL Internal Use Only *****
- 85/87/88XX with mixed DWBUA & BI-DEVICES, UNIBUS Cautions:
+ AVOID TU80, TU81, & UDA50;
+ BE WARY OF DHU11, DMF32, & DMZ32;
+ USE DR11-W WITH 45 usec. NPR DELAY;
+ TRUST DUP11 & LP11.
- 85/87/88XX with mixed DWBUA & BI-DEVICES, BI-CONFIG Cautions:
+ "BCS" TYPE BI-DEVICES ARE SAFEST WITH DWBUA (BCR);
+ "BCC" ARE 2ND-BEST WITH DWBUA;
+ AVOID "BCI/BCB" WITH DWBUA;
+ NEVER USE "BCR" (OR 2ND DWBUA) WITH DWBUA.
***** DIGITAL Internal Use Only *****
|
| 34.13 | Subj: A small mistake in the memo on cleaning materials | COMICS::ROBB | | Wed Feb 07 1990 20:39 | 49 |
|
Subject: VAX 6000 Backplane & Module Cleaning (7-feb-1990)
Last year there were some changes made to the cleaning practices of
VAXBI, NMI and XMI modules and backplanes. Just to clear up any
confusion that still seems to exist in the field, this memo details
the part numbers of the NEW cleaning materials. Any of the OLD
cleaning materials should never be used and should be safely
disposed of.
The old materials contained a lubricant like oil, which allowed
contaminants to stick to any newly cleaned contact surface areas,
possibly making the original problem, for which the modules were
cleaned, worse.
The NEW part numbers for the cleaning materials are:-
Gold Wipes 49-01603-02
5 segment paddle wipes 12-26321-05 (BI/XMI Backplanes)
4 Segment paddle wipes 12-26321-06 (NMI Backplanes)
Existing parts, still valid are:-
5 segment wipe handle 47-00116-01
4 segment wipe handle 47-00116-01
Solvex Gloves 29-26403-00
Old Parts that must no longer be used:-
5 Segment wipe 12-26321-03
4 segment wipe 12-26321-01
Please note that the new gold wipe part number has sometime been
quoted incorrectly as 49-01013-00. It should be 49-01603-02.
I hope this may help clear up any confusion, if not please let me
know.
regards
Pete Beddall
|
| 34.14 | New console command with V9 | KERNEL::BLAND | toward 2000 ... | Sun Mar 18 1990 01:38 | 70 |
| >>>SHOW IO !** This is new to me and could be very useful **
VAXBI Node Summary:
VAXBI Bus 0
Node 1 Rev=4302 DevType=0108 CIBCA-A, CI Adapter
Node 2 Rev=FFF1 DevType=0106 DB88, NMI System Adapter (NBIB)
Node 4 Rev=2108 DevType=0109 DMB32, Comm Adapter
VAXBI Bus 1
Node 2 Rev=FFF3 DevType=0106 DB88, NMI System Adapter (NBIB)
Node 7 Rev=FFFF DevType=FFFF ?????, Uninitialized BIIC
Node 15 Rev=0F1C DevType=010E KDB50, Disk Adapter
VAXBI Bus 2
(Does not exist in this machine)
VAXBI Bus 3
(Does not exist in this machine)
>>>SHOW REVS ! ** Included this as it was part of same note **
Component | LEFT CPU | RIGHT CPU | Comments
--------------+---------------+---------------+-------------------------
| |
KERNEL Top | ------- CLOSEST TO 07 ------- |
| |
CPU Top Rev | REVISION 04 | REVISION 04 |
F1009 (WCS) | A | A |
F1008 (SEQ) | A | A |
F1007 (DEC) | A | A |
F1006 (CCS) | A | A |
F1005 (ADP) | B | B |
F1004 (SLC0) | A | A |
F1003 (SLC1) | A | A |
F1002 (SHR) | A | A |
F1002 (CLK) | ------- A ------- |
| |
Microcode Top| REVISION 14 | REVISION 14 |
WCS microcode| 014 | 014 |
SDF microcode| 004 | 004 |
IBD microcode| 002 | 002 |
CCS microcode| 001 | 001 |
UWCS ucode | Not sensed | Not sensed |
| |
MCL Top Rev | ------- REVISION 06 ------- |
MCL hardware | ------- F ------- |
| |
NBI Top Rev | ------- CLOSEST TO 02 ------- |
VAXBI_0 | ------- REVISION 02 ------- |
NBIA0 | ------- A ------- |
NBIB0 | ------- B ------- |
VAXBI_1 | ------- CLOSEST TO 02 ------- |
NBIA0 | ------- A ------- |
NBIB1 | ------- D ------- |
VAXBI_2 | ------- Uncalculated ------- |
NBIA1 | ------- No module ------- |
NBIB2 | ------- No module ------- |
VAXBI_3 | ------- Uncalculated ------- |
NBIA1 | ------- No module ------- |
NBIB3 | ------- No module ------- |
| |
Console Top | ------- CLOSEST TO 09 ------- |
Console prog | ------- 009 ------- |
RTI driver | ------- 011 ------- |
XK driver | ------- 000 ------- |
P/OS software| ------- 032 ------- |
EMM firmware | ------- 068 ------- |
LBS firmware | ------- 008 ------- |
| |
|
| 34.15 | Problem with the PMCL module (F1021) | KERNEL::BLAND | toward 2000 ... | Mon Mar 26 1990 23:20 | 53 |
| <<< SASE::WRKD:[NOTES$LIBRARY]NAUTILUS.NOTE;4 >>>
-< VAX 85xx, 8700, and 88xx Product Information >-
================================================================================
Note 1110.0 PMCL Error Call Out Problem No replies
VOLKS::HARRIGAN 45 lines 23-MAR-1990 16:56
--------------------------------------------------------------------------------
*****************
* D I G I T A L * I N T E R O F F I C E M E M O
*****************
TO: Distribution DATE: 8-Mar-1990
FROM: Jerry Harrigan
DEPT: CSSE Mid-Range VAX Support
EXT: 240 6494
M/S: AET1-1/6
NET: VOLKS::HARRIGAN
SUBJECT: Problem with the PMCL (F1021)
Problem:
A problem has been discovered with the PMCL/F1021 relating to the logging of
correctable memory errors. If a high number of CRD errors occur, some of
the errors will *incorrectly* report the indicted SMU. However, the
majority of errors will be reported correctly and accurate diagnosis is
possibly by focusing on these 'majority' errors.
When the problem occurs, the symptom is that the controller will incorrectly
latch the wordcount field (bits 31:30) of MEMCSR1. Note that this symptom,
although similar to the MCL word count problem, is *not* identical in that
the majority of logged errors will, by far, be accurate. In fact, in many
cases where only a small number of CRDs occur, the latched word count will be
accurate for all errors.
Solution:
The following steps should be taken to isolate memory errors with the PMCL
module.
1. Run the Errorlog with the following command;
$ ANALYZE/ERROR_LOG/SUMMARY=MEMORY/INCLUDE=MEMORY/BRIEF
NOTE: If you want to direct the output to a file use the
/OUTPUT=filename.ext option.
2. Review the Error log output and determine the SMU with the highest number
of errors. You will usually see that one SMU would have 95% of the errors
while the other three SMU has 5% of the errors.
|
| 34.16 | S/A backup..withouot system disk. | KERNEL::MOUNTFORD | | Wed Apr 25 1990 18:07 | 237 |
| How to boot standalone on a Nautilus when system disk fails. Brief logfile
extract.
>>>@csboo.com
SET VERIFY
! CSBOO.COM
! REV 1.0
!
! Standalone backup boot command file.
! Boot from CS1:
! This file boots a VAX SINGLE CPU system
!
SET TERMINAL OPA0 ! Set up logging
SET DEFAULT HEXADECIMAL,PHYSICAL,LONGWORD
IF NOT HALTED THEN HALT ! Halt it and avoid unnecessary warning messages
WAIT ! Wait for processor to halt
EXAMINE/SDF 0CA ! Retrieve halt code
M 000000CA 00000003
DEPOSIT AP @ ! Move halt code to AP
INITIALIZE ! Initialize CPU
UNJAM ! Clear BI errors, NMI and BI node state info
DEPOSIT/L/P 3E000000 020F00F0 ! Clear interlock timeout in mem CSR0
EXAMINE/L/P 3E000010 ! Read mem CSR4 (clears interrupts)
P 3E000010 020F00F0
EXAMINE/L/P 3E000014 ! Read mem CSR5 (clears NMI fault)
P 3E000014 020F00F0
DEPOSIT/L/P 3E000008 F0000000 ! Clear RDS error bits in mem CSR0
INITIALIZE ! Initialize CPU
DEPOSIT R0 40 ! Device type (console storage device).
DEPOSIT R1 0 ! Clear unused register
DEPOSIT R2 0 ! Clear unused register
DEPOSIT R3 1 ! Unit
DEPOSIT R4 0 ! Clear unused register
DEPOSIT R5 4000 ! Boot flags
FIND/MEM ! Find 64kb of working memory, set cold
! start bit
IF NOT $STATUS THEN @EXIT ! Boot if successful
EXAMINE SP ! Show address of working memory + 512
G 0000000E 00000200
LOAD/MAINMEMORY/START=@ VMB.EXE ! Load bootstrap...
LOAD done, 72 blocks loaded starting at physical address = 00000200
START @ ! and start it.
[Entering PROGRAM mode on LEFT OPA0]
Please remove the volume "SYSTEM_1" from the console device.
Insert the next standalone system volume and enter "YES" when ready: YES
Resuming load operation on volume "SYSTEM_2", please stand by . . .
VAX/VMS Version V5.1 Major version id = 1 Minor version id = 0
Exiting TRANS mode [protocol on]...
REM> disc
MOUNTFORD disconnected at 13-APR-1990 15:18:06.15
MOUNTFORD Connected at 13-APR-1990 15:20:21.51
System id :870067409 Cpu type :8530 Console :MDS01
Speed :1200 Keyboard :TXC4 Phone :0533513136
tion established
>>>
*******************************************************************
*** Connection Established by RDC Basingstoke. ***
*******************************************************************
Message "RSDS$MGR:RSDSMSG.EN" sent successfully.
REM> tr
Entering TRANS mode [protocol off] (^A to exit)...
>>>login
LOGIN
! 13-APR-90 15:21:41 *** INFO, Now logged in.
>>>set t p
[Entering PROGRAM mode on LEFT OPA0]
Insert the next standalone system volume and enter "YES" when ready: YES
Resuming load operation on volume "SYSTEM_3", please stand by . . .
Please remove the volume "SYSTEM_3" from the console device.
Insert the next standalone system volume and enter "YES" when ready: YES
Resuming load operation on volume "SYSTEM_4", please stand by . . .
PLEASE ENTER DATE AND TIME (DD-MMM-YYYY HH:MM) 13-APR-1990 15:25
Available device CSA1: device type CRX50
Available device DUA0: device type RA81
Available device DUA1: device type RA82
Available device DUA2: device type RA81
Available device MUA0: device type TU81
Please remove the volume "SYSTEM_4" from the console device.
Insert the standalone application volume and enter "YES" when ready: yes
Resuming load operation on volume "BACKUP", please stand by . . .
%BACKUP-I-IDENT, Stand-alone BACKUP V5.0; the date is 13-APR-1990 15:29:45.17
$ backup/image mua0:system.bak/save dua0:
Exiting TRANS mode [protocol on]...
REM> tr
Entering TRANS mode [protocol off] (^A to exit)...
Exiting TRANS mode [protocol on]...
REM>
%BACKUP-I-RESUME, resuming operation on volume 2
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready:
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready:
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready:
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready:
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready:
REM> inf
Press <RETURN> to continue ? EX
REM> tr
Entering TRANS mode [protocol off] (^A to exit)...
%BACKUP-I-READYREAD, mount volume 2 on _SABKUP$MUA0: for reading
Enter "YES" when ready: yes
Exiting TRANS mode [protocol on]...
REM>
%BACKUP-I-PROCDONE, operation completed. Processing finished at 13-APR-1990 16:
07:22.27
If you do not want to perform another standalone BACKUP operation,
use the console to halt the system.
If you do want to perform another standalone BACKUP operation,
ensure the standalone application volume is online and ready.
Enter "YES" to continue:
REM> tr
Entering TRANS mode [protocol off] (^A to exit)...
If you do not want to perform another standalone BACKUP operation,
use the console to halt the system.
If you do want to perform another standalone BACKUP operation,
ensure the standalone application volume is online and ready.
Enter "YES" to continue: \<[?38l[4i[7i[?4i[!p[0q(B)0*<
}[0m[?0m>[20l[?1l[?3l[?4l[?5l[?6l[?7l[?8h[?9l[?25h[1;24r[1;sP1pS(I0,E)\[2J
>>>h
>>>
?00 CPU halted
PC = 8016D087
>>>@defboo
SET VERIFY
! DEFBOO.COM
! Rev Num 1.0
!
! This command procedure will be executed automatically during system
! initialization or following a power failure or an error halt if you have
! enabled the AUTO_BOOT keyswitch. It will also be executed when the command
! BOOT is entered.
!
! Replace DDDnnn with the name of the boot command procedure that you
! wish to execute.
! DDD = device name: BCI, BDA, UDA
! nnn = boot device unit number
!
boot bda000/r5:00000080
SET VERIFY
! BDABOO.COM
! REV 1.0
!
! COMMAND PROCEDURE TO BOOT VMS FROM A BDA DISK.
!
! This file boots a VAX SINGLE CPU system
!
! The following register deposits must be done before executing this
! command procedure or must be edited to correspond to the hardware
! configuration:
!
! R1 - Bus address information
! R2 - <31:24> = optional controller letter specifier, remaining bits
! depend on device type in R0
! R3 - device unit number
!
SET TERMINAL OPA0 ! Set up logging
SET DEFAULT HEXADECIMAL,PHYSICAL,LONGWORD
IF NOT HALTED THEN HALT ! Halt it and avoid unnecessary warning messages
WAIT ! Wait for processor to halt
EXAMINE/SDF 0CA ! Retrieve halt code
M 000000CA 00000000
DEPOSIT AP @ ! Move halt code to AP
INITIALIZE ! Initialize CPU
UNJAM ! Clear BI errors and nmi and BI node state info
DEPOSIT/L/P 3E000000 020F00F0 ! Clear interlock timeout in mem CSR0
EXAMINE/L/P 3E000010 ! Read mem CSR4 (clears interrupts)
P 3E000010 020F00F0
EXAMINE/L/P 3E000014 ! Read mem CSR5 (clears nmi fault)
P 3E000014 020F00F0
DEPOSIT/L/P 3E000008 F0000000 ! Clear RDS error bits in mem CSR0
INITIALIZE ! Initialize CPU
DEPOSIT R0 21 ! BDA boot device type codE
DEPOSIT R1 20 ! <3:0>=BI node #, <5:4>=BI #
DEPOSIT R2 0 ! <31:24>=optional controller letter specifier
DEPOSIT R3 0 ! Unit # of drive, decimal radix
SET DEFAULT HEXADECIMAL ! Reset radix
DEPOSIT R4 0 ! Not applicable
!EPOSIT R5 0 ! Software boot control flags and
! root directory # in <31:28>. Defaults
! to sys0 or unrooted directory
FIND/MEM ! Find 64kb of working memory; set cold
! start bit
IF NOT $STATUS THEN @EXIT ! Boot if find was successful
EXAMINE SP ! Show address of working memory + %X200
G 0000000E 00000200
LOAD/MAINMEMORY/START=@ VMB.EXE ! Load VMB into good mem + %X200
LOAD done, 72 blocks loaded starting at physical address = 00000200
START @ ! Start executing VMB
[Entering PROGRAM mode on LEFT OPA0]
Exiting TRANS mode [protocol on]...
REM> disc
MOUNTFORD disconnected at 13-APR-1990 16:09:07.77
|
| 34.17 | modules overheat - filter problem | KERNEL::BLAND | toward 2000 ... | Fri Oct 19 1990 04:29 | 95 |
| Author : RODNEY BOYLE
User type : USER
Location : CSSE
Vaxmail address : CSSE::BOYLE
+---------------------------+TM
| | | | | | | |
| d | i | g | i | t | a | l | TIME DEPENDENT CASE
| | | | | | | | (blitz)
+---------------------------+
TITLE: Potential air flow problem on Nautilus and Polarstar systems.
AUTHOR: Michael Person DATE: 17-OCTOBER-1990
DTN: 240-6489 TD #; 000433
ENET: VOLKS::PERSON CROSS REFERENCE #'s: PR06019
DEPARTMENT: CSSE VAX SYSTEM SUPPORT (PRISM/TIME/CLD#'s)
INTENDED AUDIENCE: ALL PRIORITY LEVEL: 2
(U.S./EUROPE/GIA) (1=TIME CRITICAL,
2=NON-TIME CRITICAL)
=====================================================================
PROBLEM:
A problem was discovered with the NOISE BAFFLE FOAM located
inside both rear doors on the Nautilus and Polarstar cabinets.
The foam becomes loose, falls into the plenum, and ends up
wedged in the bottom of the plenum, blocking the airflow with
the potential to cause thermal damage to modules. On 9/26/86
an ECO was implemented changing the configuration of the foam
pieces so that they locked each other in, eliminating the risk
of the foam dislodging. Any system that shipped prior to this
date will need to be checked. It is impossible to track down
which systems have the ECO. As a result, every Nautilus and
Polarstar system should be checked to see if it is at risk for
the foam to dislodge.
RESOLUTION/WORKAROUND:
On the systems that are potentially a problem, the top piece is
stuffed or wedged in the top. The only force that is holding
this piece in comes from the adhesive along the top. The
diagram below shows this configuration:
top foam ---- ---- adhesive
| |
V V
_______________________________
| | | |
| | | |
| |_______________| |
side foam --> | | | | <--- side foam
| | | |
| | | |
On the systems that incorporate the ECO dated 9/26/86, the top
foam piece is held captive by the side pieces. The side foam
overlaps the top by at least 3 1/2 inches on either side,
locking in the top piece as shown in the diagram below:
top foam ---- ---- adhesive
| |
V V
_______________________________
| |
| |
|_______________________________|
| | | |
| | | |
side foam --> | | | | <--- side foam
| | | |
| | | |
The systems at risk do pose the threat of blocked airflow to
the CPU and as a result, potential damage to the modules. To
fix these systems, it is recommended that the top piece of foam
be checked to see if there is still integrity in the adhesive.
If the adhesive shows any signs of deterioration, the piece of
foam should be removed and reinserted with new adhesive, part
number 12-25345-01. All systems that are at risk should be
checked periodically to insure that the foam pieces are still
secure.
ADDITIONAL COMMENTS: NONE
*** DIGITAL INTERNAL USE ONLY ***
|