| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 454.1 | OPA2 errors vms6.2-1h3 | CSC32::HUTMACHER | | Thu Jan 30 1997 17:33 | 28 |
| Hi i cant seem to recreate this on our asrv4100 but we have S3
vga card in the backplane. setting console to serial and running
vms 6.2-1h3 from there is not generating errors on opa2 with
keyboard installed or not
so have questions
what srm console version? ours is 3.0-10
do you have any keyboard on mouse connected to distribution panel
where you connect vt terminal to com1? if you do remove them
does >>>show console show serial for sure?
there is a sysgen> param Window_system that should be set to 0 if no
vga adapter present
sysgen>set Window_system 0
sysgen>write current
sysgen>write active
reboot system
dont know if this will help but only thing i can think of checking
so far.
jim hutmacher mvhs colorado csc 800-354-9000 ext 25561
|
| 454.2 | WINDOW_SYSTEM is set to "1" | KAOFS::M_NAKAGAWA | | Thu Jan 30 1997 17:55 | 34 |
| Hi Jim, Thanks for such a quick reply.
>>what srm console version? ours is 3.0-10
YES! V3.0-10
>>do you have any keyboard on mouse connected to distribution panel?
NO, nothing is connected.
We connected a keyboard with live system but it did not make any diff.
VT terminal is connected to COM1.
>>does >>>show console show serial for sure?
YES!
>>there is a sysgen> param Window_system that should be set to 0...?
It is set to "1" !!!!!
SYSGEN> SHOW WINDOW_SYSTEM
Parameter Name Current Default Min. Max. Unit Dynamic
-------------- ------- ------- ---- ---- ---- -------
WINDOW_SYSTEM 1 0 0 2 Pure-numbe D
We will try setting WINDOW_SYSTEM to "0".
Will update this entry with the result.
Thanks again,
CRDC/Mitz
|
| 454.3 | Window_system = 0 didn't help | KAOFS::M_NAKAGAWA | | Thu Jan 30 1997 18:11 | 18 |
| Set sysgen parameter "Window_system" to "0".
It did not help.
Since its a Dynamic parameter we did not reboot system but did
SYSGEN> WRITE CURRENT, WRITE ACTIVE, USE CURRENT.
(This is a critical system and can't reboot it now)
Customer still has S3 VGA card on site but we do not have empty PCI
slot to put it backin for troubleshooting.
Jim, any chance for you to remove S3 VGA card from your system to
see whether we can reproduce the symptom?
Thanks,
Mitz
|
| 454.4 | opa2 errors on vms6.2-1h3 | CSC32::HUTMACHER | | Thu Jan 30 1997 18:37 | 31 |
| another thought i had was maybe system is running ECU for WNT
not VMS/UNIX
might be worth going into alphabios - ultilities - run ecu and
load the ECU floppy for VMS/UNIX just in case
what version ECU are loading?
i again tried to break our machine by loading WNT ECU and running
vms and no opa2 errors??
i cant get access to our machine right now again but will try
pulling s3 vga tommorrow morning
might want to boot vms in min. conf see if it errors then
b -fl 0,1 dka0
sysboot>set startup_p1 "min"
sysboot>cont
vms should come up see of opa2 errors
to set back to full vms boot
sysboot>set startup_p1 ""
sysboot>cont
if doesn't fail is there some software trying to use opa2??
did system always do this?
do the errors start right away or wait awhile?
it's possible that the saddle b3050-aa module or RIO module 54-24674-01
have gone bad ??
jim hutmacher mvhs colorado csc 800-354-9000 ext 25561
|
| 454.5 | | KAOFS::M_NAKAGAWA | | Thu Jan 30 1997 22:10 | 45 |
| Jim,
Thanks for the tips and your assistance, will pass it to our local
engineer here.
>>did system always do this?
>>do the errors start right away or wait awhile?
The engineer who installed this system did not notice it.
He installed VMS by himself for UETP purpose and did not have any
problems or at least did not notice problems with OPA2.
(...but VGA card was still in the system that time)
Then he had to reconfigure the PCI bus to add options, S3 VGA card
was removed during PCI bus reconfiguration to create PCI space.
(I suspect the problem started after this)
He handed system to the customer and customer reinstalled VMS and
noticed OPA2 errors.
>> i cant get access to our machine right now again but will try
>> pulling s3 vga tommorrow morning
Greatly appreciated if you could try it for us.
This is our system configuration just in case you wondered:
$ mc sysman io show bus
%SYSMAN-I-OUTPUT, command execution on node CDW023
_Bus__________Node_TR#__Name____________Base CSR__________
PCI 8 3 PCI/SCSI FFFFFFFFEF59E000
PCI 512 4 PCI/SCSI FFFFFFFFEF5A2000
PCI 24 3 PCI/FDDI FFFFFFFFEF5A4000
PCI 32 3 PCI/NI FFFFFFFFEF5A6000
PCI 40 3 PCI/NI FFFFFFFFEF5A8000
XBUS 0 7 System Board FFFFFFFFEF5B6000
XBUS 1 7 Floppy FFFFFFFFEF5B6000
XBUS 2 7 Parallel Port FFFFFFFFEF5B6000
XBUS 3 7 Serial Port FFFFFFFFEF5B6000
PCI 16 5 PCI/DSSI FFFFFFFFEF7DC000
PCI 40 5 PCI/CI FFFFFFFFEF7DE000
Unknown 1 2 Unknown device 0000000000000000 <-- ???
Thanks again,
Mitz
|
| 454.6 | opa2 errors on asrv4100 vms6.2-1h3 | CSC32::HUTMACHER | | Fri Jan 31 1997 10:02 | 46 |
| Hi Mitz
was able to get time on our asrv4100 again here at colorado csc
pulled the S3 vga card, >>>set console serial, pulled off the
mouse/keyboard and booted vms 6.2-1h3
no errors being generated on OPA2 keyboard port in normal use.
so tried to create some errors by
$reply/terminal=opa2 "hello this is test"/bell
this command does create 1 error on every attempt
could this customer have some kind of .com procedure or program
poking the opa2 port?
could he have peek/watch program thats spying on user key strokes
on opa2 port with no keyboard attached?
does $sho dev opa2/full or $sho terminal opa2 show any process
allocated to it?
boot min might be good idea see if that kind of boot gets opa2 errors
b -fl 0,1 dka0
sysboot>set startup_p1 "min"
sysboot>cont vms should come up and see of opa2 errors occur
to set back to a full vms boot
sysboot>set startup_p1 ""
sysboot>cont
if you do go with the hardware when replacing the B3050 sadlle module pci
backplane than need to be real careful because it easy to damage it.
Note 93.29 describes this issue.
basically when you remove the RCM/RIO module it exposes two metal
standoffs that that can contact the B3050 module while its being
removed/installed. these killer standoffs usually destroy the etching
to two large asic chips located in the lower right hand corner of the
b3050 module. it recommended that you lean a piece of cardboard or
a plastic Visa card with a real "High Credit Limit" against the
standoffs to keep any contact from being fatal.
jim hutmacher mvhs colorado csc 800-354-9000 ext 25561
|
| 454.7 | OPCOM directed to OPA2 ??? | KAOFS::M_NAKAGAWA | | Fri Jan 31 1997 15:00 | 24 |
| Jim, thanks for your time and effort.
I was kinda hoping that OPA2 and OPA3 to disappear if you remove the
VGA card and set console to serial, something like forllowing:
[With VGA] [Without VGA]
VGA = OPA0 COM1 = OPA0
Keyboard = OPA2
Mouse = OPA3
But its not the case I guess.
As I entered in .0 base note, OPA2 is owned by system,
OPA2 Unknown UCB address: 82586080
Owner UIC [000001,000004] Operation count 25139 ORB address 8262AD80
PID 0001000B Error count 25127 DDB address 8E724878
The VT serial system console is not getting any OPCOM massages as per
customer, is VMS sending all OPCOM messages to OPA2?????
Mitz
|
| 454.8 | opa2 errors on vms 6.2-1h2 | CSC32::HUTMACHER | | Fri Jan 31 1997 15:43 | 15 |
| Mitz you could try this to make sure opa2 not getting opcom
and opa0 will when vt console is logged out.
on any terminal with privs
$assign/user opa0: sys$command this says next command for opa0 command
$reply/enable enables opa0 opcom
$assign/user opa2: sys$command this says next command for opa2 command
$reply/disable disables opa2 opcom
if this fixes it cant really explain how we got into this. a reboot
with window_system = 0 should bring it up where opa0 gets opcom
by default not opa2
jim hutmacher mvhs colorado csc 800-354-9000 ext 25561
|
| 454.9 | Someone changed startup command file?? | KAOFS::M_NAKAGAWA | | Fri Jan 31 1997 15:58 | 28 |
| Jim,
I just logged into customer's system to see what SYSTARTUP_VMS.COM
has to say about OPA2:
Directory SYS$COMMON:[SYSMGR]
SYSTARTUP_VMS.COM;7 17-JAN-1997 10:24:46.48
^^^^^^^^^^^
$! SYS$MANAGER:SYSTARTUP_VMS.COM
$! 17-Jan-1996 GBU Added x25$startup,
$! reply/enable to opa2
:
:
$! OPA2: is one of the terminal devices.
$!
$ define sys$command opa2:
$ reply/enable
$ deassign sys$command
Wouldn't this explain our problem!!!
Mitz
|
| 454.10 | opa2 errors on vms 6.2-1h3 asrv4100 | CSC32::HUTMACHER | | Fri Jan 31 1997 16:14 | 10 |
| yes i think so, you may want to try 454.8 procedure to shut it
down see if errors stop
if so will need to comment those out of startup so it does not
trying using opa2 as terminal port
looks like somebody installed x25 startup pointing to opa2?
jim hutmacher mvhs colorado csc 800-354-9000 ext 25561
|
| 454.11 | Big T H A N K S to Jim | KAOFS::M_NAKAGAWA | | Fri Jan 31 1997 16:40 | 14 |
| Jim,
.8 suggestion is done. BINGO! It worked.
OPA2 error count has stopped increasing.
I asked customer to check VT terminal to see whether OPCOM messages
are showing up there.
I hope this will clearup console being slow symptom as well.
Thanks again,
Mitz
|