[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
Title: | DECmcc user notes file. Does not replace IPMT. |
Notice: | Use IPMT for problems. Newsletter location in note 6187 |
Moderator: | TAEC::BEROUD |
|
Created: | Mon Aug 21 1989 |
Last Modified: | Wed Jun 04 1997 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 6497 |
Total number of notes: | 27359 |
1904.0. "ACC VIO on GETEVENT verb" by COL01::LUNT () Tue Dec 10 1991 06:03
Hello all,
After searching through all notes on events and access violations I
still have no anser to my problem which is the following:
I'm trying to learn how to sink events on my MCC node and then set
alarm rules to correspond to them. So I got out the Phase IV manual and began
to read. First I set up my local sink(node=BYLV21) to be of type monitor and
have the name mcc_dna4_evl. I made sure that the mcc_dna4_evl.exe was installed
as a shareable image. I checked that the file mcc_dna4_evl.com and the
directory containing it had world read,executable protection. I then choose a
local node to act as the victim(node=BYLU08) for this test. I created an
outbound stream to my mcc node of type monitor for the event 0.9(line counters
zeroed). I then created an event and used reply/enable to ensure that my mcc
node was actually receiving the events. I then issued a getevent node4
victim-node line *, line counters zeroed command and received and access
violation.
Have I missed something simple in the setup?
The following is a sethost log to show the setup:
(Please excuse the Carriage and Form feeds, it is easier to read if printed
out.)
Username: SYSTEM
Password:
Welcome to VAX/VMS V5.4-2
Last interactive login on Tuesday, 10-DEC-1991 10:29
Last non-interactive login on Tuesday, 10-DEC-1991 10:47
BYLV21 SYS > mcr ncp show logging monitor
Logging Volatile Summary as of 10-DEC-1991 11:36:54
Logging sink type = monitor
Sink Node Source Events State Name
on
1.82 (BYLV21) (All sources) 0.0-9 mcc_dna4_evl
(All sources) 2.0-1
(All sources) 4.2-13 15-16
18-19
(All sources) 5.0-18
(All sources) 128.0-4
(All sources) 352.0-31
(All sources) 353.1-4
>
BYLV21 SYS > manage/enter
DECmcc (V1.1.0)
)0[4l
=
>
)0[4l
=
>
MCC> show node4 bylu08 outbound stream bylv21 all attributes
Node4 1.396 Outbound Stream 1.82
AT 10-DEC-1991 11:37:54 All Attributes
Address = 1.82
Name = BYLV21
Node could not open required file: Permanent Database.
File = Permanent Database
MCC>
MCC> show node4 bylu08 outbound stream bylv21 all attributes, by user system,by password superman1
Node4 1.396 Outbound Stream 1.82
AT 10-DEC-1991 11:38:20 All Attributes
Address = 1.82
Name = BYLV21
Initial Address = 1.82
Initial Name = BYLV21
MCC>
MCC> show node4 bylu08 outbound stream bylv21 remote sink * all attributes, by user system, by password superman1
Node4 1.396 Outbound Stream 1.82 Remote Sink Monitor
AT 10-DEC-1991 11:39:34 All Attributes
Type = Monitor
EventFilter = { ( Choice = C5,
Source_Entity = All Sources,
Class = 0,
Event_Type = { 9})}
No information was returned for the entity requested.
MCC> show node4 bylv21 local sink * all attributes
Node4 1.82 Local Sink Monitor
AT 10-DEC-1991 11:40:12 All Attributes
Type = Monitor
State = On
Name = "mcc_dna4_evl"
Initial State = On
Initial Type = Monitor
MCC> getevent node4 bylu08 line *, line counters zeroed
Type ? for a list of valid event IDs
Event: line counters zeroed
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=0000000C, PC=000F0B80, PSL=03C00004
>
>
%SYSTEM-E-ACCVIO, access violation, reason mask=00, virtual address=0000000A, PC=0000000A, PSL=0000000F
BYLV21 SYS > manage/enter
DECmcc (V1.1.0)
)0[4l
=
>
)0[4l
=
>
MCC> getevent node4 bylu08 line *
Type ? for a list of valid event IDs
Event: ?
Automatic Line Counters
Line Counters Zeroed
Automatic Counters
Counters Zeroed
Initialization failure
Send Failed
Receive Failed
Collision Detect Check Failed
Retransmit Maximum Exceeded
FRMR Received
Illegal Frame Received
Data Set Ready Transition
Ring Indicator Transition
Unexpected Carrier Transition
Memory Access Error
Communications Interface Error
Performance Error
Any Events
Event: line counters zeroed
%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=0000000C, PC=000F0B80, PSL=03C00004
>
>
%SYSTEM-E-ACCVIO, access violation, reason mask=00, virtual address=0000000A, PC=0000000A, PSL=0000000F
BYLV21 SYS > reply/enable
%%%%%%%%%%% OPCOM 10-DEC-1991 11:42:04.46 %%%%%%%%%%%
Operator _BYLV21$RTA2: has been enabled, username SYSTEM
%%%%%%%%%%% OPCOM 10-DEC-1991 11:42:04.48 %%%%%%%%%%%
Operator status for operator _BYLV21$RTA2:
CENTRAL, PRINTER, TAPES, DISKS, DEVICES, CARDS, NETWORK,
CLUSTER, SECURITY,
LICENSE, OPER1, OPER2, OPER3,
OPER4, OPER5, OPER6, OPER7, OPER8, OPER9, OPE
%%%%%%%%%%% OPCOM 10-DEC-1991 11:42:17.62 %%%%%%%%%%%
Message from user DECNET on BYLV21
DECnet event 0.9, counters zeroed
From node 1.396 (BYLU08), 10-DEC-1991 11:42:44.14
Line SVA-0
BYLV21 SYS > log
SYSTEM logged out at 10-DEC-1991 11:42:21.42
THis is a copy of the mcc_dna4_evl.log:
�$! This command procedure is always run when anybody on the entire system
�$! logs in. It is equivalent to LOGIN.COM except that the instructions
�$! contained herein are executed everytime anyone on the VMS system
�$! logs in to their account.
�$!
�$! For interactive processes, turn on Control T, and set the terminal type
�$!
�$ IF (F$MODE() .EQS. "BATCH") THEN GOTO EXIT
�$ IF (F$MODE() .EQS. "NETWORK") THEN GOTO EXIT
�$ IF (F$MODE() .EQS. "INTERACTIVE") THEN SET CONTROL=T
�$ IF (F$MODE() .EQS. "INTERACTIVE") THEN SET TERMINAL/INQUIRE
�$!
�$!---------------------------------------
�$! disable Control Y
�$!---------------------------------------
�$ set nocontrol=Y
�$!
�$!---------------------------------------
�$! redefine DTR symbol
�$!---------------------------------------
�$ XX = ""
�$ IF XX .EQS. ""
�$ THEN
�$ IF f$search("SYS$SYSTEM:DTR32.EXE") .nes. ""
�$ THEN
�$ DTR == "$DTR32"
�$ ELSE
�$ ENDIF
�$ ENDIF
�$!---------------------------------------
�$! end dtr definition
�$!---------------------------------------
�$!
�$!
�$! Avoid customer the use of /Nocommand at login
�$!
�$!
�$ aa= F$GETJPI(0,"ACCOUNT")
�$ if aa .eqs. "CUSTOMER"
�$ else
�$ set control=Y
�$ endif
�$!
�$!
�$!
�$! For MicroVAX systems only, use the command INSTALL/COMMAND
�$! and MOUNT/NOASSIST.
�$!
�$ IF (.NOT. F$TRNLNM("SYS$MICROVAX")) THEN GOTO SKIP_MICROVAX_COMMANDS
�$SKIP_MICROVAX_COMMANDS:
�$!
�$! Place your site-specific LOGIN commands below
�$!
�$ kr :== set def sys$sysdevice:[customer.kretschmer]
�$ ALARM :== SET DEF DISK$DATA:[MCC_ALARMS]
�$ DW :== SET DEF DISK$DATA:[DECWRITE]
�$EXIT:
�$EXIT
�$!
�$! This command procedure is a framework for building an individual
�$! login command procedure. It is executed every time that you log
�$! into your account.
�$!
�$ IF F$MODE() .NES. "INTERACTIVE" THEN GOTO ENDINTER
�$ENDINTER:
�$ IF F$MODE() .NES. "BATCH" THEN GOTO ENDBATCH
�$ENDBATCH:
�$!
�$! Put commands that you want executed in any mode of login below.
�$!
�$!
�$ manage/enter/presen=mcc_dna4_evl
�Network object MCC_DNA4_EVL is declared, Status = 52854793
�Waiting for the event message from EVL.....
�%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=0000000C, PC=000F0B80, PSL=03C00004
�Failed to establish connection with EVL, status = 52877226
�%SYSTEM-E-ACCVIO, access violation, reason mask=00, virtual address=0000000A, PC=0000000A, PSL=00219200
� SYSTEM job terminated at 10-DEC-1991 10:47:13.75
�
Accounting information:
� Buffered I/O count: 68 Peak working set size: 2103
� Direct I/O count: 34 Peak page file size: 7088
� Page faults: 1629 Mounted volumes: 0
� Charged CPU time: 0 00:00:02.13 Elapsed time: 0 00:00:10.20
T.R | Title | User | Personal Name | Date | Lines |
---|
1904.1 | event pool corrupted? but why? | TOOK::JEAN_LEE | | Tue Dec 10 1991 17:02 | 22 |
|
Yes, it is a bizarre symptom. However, I suppose it's possible that
the event pool got corrupted.
Questions:
Is it reproducible?
Have you been running dna4_evl sink without problem until now?
What type of systems are bylu08 and bylu21 (I can't see it)?
How busy are those events generated?
What is your process quota?
In the meantime, I am trying to find a V11 MCC system so I can try to
reproduce this problem.
Thanks for the input.
Jean
|
1904.2 | answers to questions | COL01::LUNT | | Wed Dec 11 1991 02:52 | 20 |
| Hello Jean,
Thanks for the quick answer. The two machines are both VMS machines.
The BYLU08 is standalone 3100. The BYLV21 is a 3800 with one sattelite, a
3100, as cluster member. The BYLV21 is runnin VMS 5.4-3. BYLU08 VMS version
is minimum 5.4, but I'm not sure exactly which version as I am writing this
from Digital and the machine is at a customer site.
I only started using the mcc event logger on Monday, the 9th of
December and the problem is very reproducable, in that I have never
succeeded in having a GETEVENT directive work with out MCC giving me an
access violation. As to the number of events, it is 1. I create the event
myself by zeroing the line counters on BYLU08 and see with Reply/enable exactly
one event. There are however the general events occuring on BYLV21 such
as adjaceny down, or Downline Load failures, but they occur about
2 per minute, not very frequently.
The process quota on the BYLV21 is, I believe, 20.
Julie Ann
|
1904.3 | have I been forgotten??? | COL01::LUNT | | Wed Jan 08 1992 10:25 | 8 |
| Hello,
Has my problem been forgotten? I realize with the field test and
the volume of notes that you people are pretty busy, but I need to
resolve my problem also.
Thanks
Julie Ann
|
1904.4 | event pool corrupted | TOOK::JEAN_LEE | | Fri Jan 10 1992 15:15 | 22 |
| Julie Ann,
No, your problem is not forgotten. Like you said, the tight schedule
to release V1.2 and holidays really distracted our attention somewhat.
I finally found a V1.1 to sink events (not easy, we all use V1.2 now ;-)
). I tried and couldn't reproduce the acc-violation problem at all.
Your log showed acc-viols occurred on both event getter (in getevent)
and event putter (in sink) ends. I have consulted this behavior with Kernel
team regarding the possibility of such acc-violations. The answer is that
the event pool may have been corrupted. For V1.1, this is possible and
unpredicatable. The advice I got is that you should exit and kill ALL MCC
processes on the entire system. Restart MCC again.
Do you have access to V1.2 kit? V1.2 MCC Event Manager is much
improved and more reliable. It may be a good idea to upgrade your mcc to
V1.2. If you decide to do so, let me know. You will need to have some
new images to run event sink.
Jean
|
1904.5 | it works now under V1.2 | COL01::LUNT | | Thu Jan 30 1992 10:55 | 5 |
| Hi Jean,
Well, it works under V1.2.
Thanks, Julie Ann
|