[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference csc32::consolemanager

Title:POLYCENTER Console Manager
Notice:Kits, Scans, Docs on CSC32:: as PCM$KITS:,PCM$DOCS:, PCM$SCANS:
Moderator:CSC32::BUTTERWORTH
Created:Thu Aug 06 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1541
Total number of notes:6564

624.0. "DCR 1.5a/OSF V3.2 => trap: invalid memory read ac" by GENIE::LUETHIR (Roland L�thi, NIS @EBO (DTN: 7-760-4753)) Thu Feb 23 1995 12:10

Hi

Hi

I'm trying to run PCM V1.5a under OSF/1 V3.2 but PCM crashes at
startup time:

# ./consoled start
POLYCENTER Console Manager starting.
Starting the ENS Daemon
Starting the Console Controller Daemon

        trap: invalid memory read access from kernel mode

        ...

        panic (cpu 0) kernel memory fault

I've done an upgrade from OSF/1 V3.0 to V3.2 and reinstalled PCM.
Is PCM V1.5a not supported on OSF/1 V3.2.

Thanks for any help
Cheers .../Roland
T.RTitleUserPersonal
Name
DateLines
624.1Same experienceHGOVC::JOSEPHKWOKFri Feb 24 1995 07:427
Same experience as .1, just after installupdating V3.0 (Rev 358) to
V3.2 (Rev 214), my system will crash just when staring console manager.
I was using Console Manager V1.5 and I have to deinstall it so that
I may start the system.

-Joseph
624.2OPG::SIMONFri Feb 24 1995 08:565
I shall look into this as a matter of urgency. I am hampered by not having 
received my V3.2 kits as yet. I think the UK must be at the end of the camel 
train route from the USA.

Cheers Simon...
624.3check kernel parameter for maxusers & maxuprcHGOVC::JOSEPHKWOKFri Feb 24 1995 09:1922
Oops!

I make a big mistake.  Not my fault, it was installupdate!
It erase all my kernel configuration information!!

As state in Installation Guide, we should define

maxusers 256
maxuprc 100

But after installupdate, all are gone, I need to rebuild kernel to
get back my originally configuration.  After that, I can start
console manager.

By the way, I still have problem with IOS* (I18N savesets) being
installed in V3.2 with Console Manager V1.5a.  I still can't edit
a database, refer to note 500.*

When this can be fixed?

-Joseph
624.4RE.3 is wrongHGOVC::JOSEPHKWOKFri Feb 24 1995 10:1811
Ooop!

I am wrong again!  I thought I was right but I am not!
The problem is really there!  I can't start console manager
after system startup.  I have tried first on single user mode
with only license reseted and IP started which seems fine.

Console Manager V1.5a is really can't run on my DEC 7000/610 system.

-Joseph
624.5installed OSF/1 V3.2 from scratchGENIE::LUETHIRRoland L�thi, NIS @EBO (DTN: 7-760-4753)Fri Feb 24 1995 15:509
Hi

I've installed OSF V3.2 from scratch. Now it seems DCR V1.5a is working
fine :-)

Does anybody where's the difference is between an UG from V3.0 to V3.2
and an installation of V3.2 from scratch. 

Cheers .../Roland
624.6same problemKYOSS1::GREENFri Feb 24 1995 20:533
    	We are seeing same problem here with upgrade installation.
    		dick
    
624.7de-install, re-install?KYOSS1::GREENFri Feb 24 1995 22:214
    	I wander if you de-install before 3.2 upgrade then re-install after
    would resolve the problem??
    		dick green
    
624.8OPG::SIMONWed Mar 01 1995 16:007
Hi,
    the people who have seen this problem. Are you using LAT as the connection
method? If so can you disable your LAT managed systems and just leave any others.

Do you still get the crash?

Cheers Simon....
624.9using telnetKYOSS1::GREENWed Mar 01 1995 16:397
    	In our environment we connect to console port via telnet.
    	The person responsible for consolemanager restore OSF 3.0.
    	I had suggested de-installing console manager then 3.2 upgrade,
    re-install console manager, but unless I could provide proof that
    this was documented way of dealing with upgrade, they won't do it.
    	So we are back at OSF 3.0 and 1.5a.
    		dick
624.10OPG::PHILIPAnd through the square window...Wed Mar 01 1995 17:1710
Dick,

  Could you mail an export of your PCM database to elbrig::sdj, we can't
  reproduce a system panic here when using telnet, so there could be other
  factors involved. However, we can cause a system panic using LAT, hence
  Simon's question.

Cheers,
Phil

624.11re -1KYOSS1::GREENWed Mar 01 1995 20:133
    	Just found out, they are using LAT and TELNET.
    		dick
    
624.12ProgressOPG::SIMONThu Mar 02 1995 08:4816
Hi all,
       although I have not got the SSB V3.2 I put a T3.2 kit up yesterday and
reproduced the crash. It certainly looks like a bug in LAT on V3.2 in the way we
use it. I wrote a small program (based on the code we use to open LAT ports) 
and this also crashed the machine. I have been in contact with LAT eng and they 
reproduced the problem and admit that there is a bug in LAT, i.e. opening the
port in the way we do should not cause a system panic. 

So what is happening is:

1) I am QARing this with OSF/1 LAT eng as a show stopper bug in V3.2.

2) I am looking at modifying the way we do LAT opens for OSF/1. I will need to  
   ensure this does not produce other knock on effects, so watch this space.

Cheers Simon....
624.13OPG::SIMONSimon Jackson @reo 830 X3879Tue Mar 07 1995 10:155
Hi,
   OSF engineering have created a patch for the problem, it is currently being
tested and will be made available shortly.

Cheers Simon...
624.14lat patch for this?CSC32::PITTMon Mar 13 1995 19:1737
    
    
    I got his mail from a customer today in ref to this crash problem.....
    
    
    I have additional information for you from the DEC OSF/1 LAT
    Engineering
    group:
    --------------------8<--------------------
    Randy,
    
    Yes, you hit a bug in the LAT device driver.  The Polycenter Console
    Manager
    folks alerted me to this about two weeks ago and I have provided a
    patch
    to our support organization.
    Have you ever obtained patches for OSF/1?  If you haven't, and have no
    idea of how to go about it, let me know and I'll have them contact you.
    Otherwise, use your normal channels and ask for the V3.2 LAT patch.
    (it's
    the only V3.2-specific patch that we've generated to date)
    
    Please don't hesitate to contact me if you have any further questions.
    
    
    Regards,
    
    Scott
    
    
    
    ----------------------------------
    
    SO who is Scott, and where can I get the v3.2 LAT patch?
    
    thanks.
    Cathy
624.15OPG::SIMONSimon Jackson @reo 830 X3879Tue Mar 14 1995 08:316
Cathy,
      rathewr than give this guy's name in the notes conference I shall mail
you off line. The patch is currently being genereated by the LAT folks and so
is available on demand from the eng group rather than through TIMA. 

Cheers Simon....
624.16LAT patch now availableRUSURE::SPINNEYno good deed goes unpunishedFri Mar 31 1995 01:4137
The official V3.2 LAT patch for this problem is now available.
The patch ID is OSFV320-017-1.

Roger Spinney
UNIX Support Engineering Group

===============================================================================

/usr/sys/BINARY/lattioc.o	(STLQ10114, HPXQ30CD8, QAR 30769)
CHECKSUM: 58233    11
---------------------

Patch ID: OSFV320-017-1

This patch fixes two problems:

1. On a DEC OSF/1 V3.2 system, when logging in via LAT, sometimes the
line feed and carriage return characters are not echoed to the screen
after entering the username. While the problem is cosmetic, when using
C2 security, explanatory text about generated passwords does not get
displayed.

2. System panics with a kernel memory fault in lats_param().
The beginning of the stack trace looks like this:

0 boot
1 panic(s = 0xfffffc000057da30 = "kernel memory fault")
2 trap()
3 _XentMM
4 lats_param"../../../../src/kernel/lat/streams/lattioc.c":259
5 lats_tioc"../../../../src/kernel/lat/streams/lattioc.c":183
6 latsuwput"../../../../src/kernel/lat/streams/latuwput.c":306
7 csq_turnover"../../../../src/kernel/streams/str_synch.c":934

A kernel rebuild is required.
===============================================================================
624.17the patch ID has changedRUSURE::SPINNEYTue Apr 04 1995 17:3037
For some inexplicable reason, the patch ID has changed. It is now
OSF320-043. The patch itself has not changed.

Roger Spinney
UNIX Support Engineering Group

===============================================================================

/usr/sys/BINARY/lattioc.o	(STLQ10114, HPXQ30CD8, QAR 30769)
CHECKSUM: 58233    11
---------------------

Patch ID: OSF320-043

This patch fixes two problems:

1. On a DEC OSF/1 V3.2 system, when logging in via LAT, sometimes the
line feed and carriage return characters are not echoed to the screen
after entering the username. While the problem is cosmetic, when using
C2 security, explanatory text about generated passwords does not get
displayed.

2. System panics with a kernel memory fault in lats_param().
The beginning of the stack trace looks like this:

0 boot
1 panic(s = 0xfffffc000057da30 = "kernel memory fault")
2 trap()
3 _XentMM
4 lats_param"../../../../src/kernel/lat/streams/lattioc.c":259
5 lats_tioc"../../../../src/kernel/lat/streams/lattioc.c":183
6 latsuwput"../../../../src/kernel/lat/streams/latuwput.c":306
7 csq_turnover"../../../../src/kernel/streams/str_synch.c":934

A kernel rebuild is required.
===============================================================================