Advise?

[vmsserbo]

One of our systems crashed lastnight. It is on a windows2003 server running Charon OPenvms 7.3.

WHen I rebooted this is what IS happening? Please advise?

Nightly backups and accounting did finish successfully lastnight. There was an error in the crash dump but it was from July 29th.

See below:



OpenVMS (TM) VAX System dump analyzer



Dump taken on 29-JUL-2006 09:02:12.10

CPUSANITY, CPU sanity timer expired



Also, I noticed that all the addm0100.log from the previous nightly runs, seemed to have vanished and there is one log with a date of 30-NOV-2006??

See below:



SEATTL> dir addm0100.log



Directory DUA1:[ADMIN]



ADDM0100.LOG;1826 108/112 8-AUG-2006 23:45:00.06

ADDM0100.LOG;1825 4/8 30-NOV-2006 11:36:02.59





Also, in the operator.log there were opcom messages dated 30-NOV-2006 as well.

See below:



%%%%%%%%%%% OPCOM 30-NOV-2006 11:36:44.24 %%%%%%%%%%%

Message from user SYSTEM on SEATTL

%MULTINET_NTPD-I-EXIT, NTPD shutting down by NETCONTROL request.



%%%%%%%%%%% OPCOM 30-NOV-2006 11:36:47.00 %%%%%%%%%%%

Message from user SYSTEM on SEATTL

%MULTINET_NTPDATE-E-UNEXPCTD, Can't adjust the time of day: error 0



%%%%%%%%%%% OPCOM 30-NOV-2006 11:36:47.76 %%%%%%%%%%%

Message from user SYSTEM on SEATTL

%MULTINET_NTPD-I-STARTUP, NTP release v4.1.1 (rev 1.0).



%%%%%%%%%%% OPCOM 8-AUG-2006 17:14:11.10 %%%%%%%%%%%

Message from user SYSTEM on SEATTL

STARTUP, *** SEATTL reboot has completed. You may log in.



Total of 2 files, 112/120 blocks.

[Robert Gezelter · Accepted solution]

Miles,

There is very little information in your posting.

For a start, since you are running in a simulated (virtual) environment (Charon-VAX) on a Windows platform, I would want to know if any disks were restored as part of the Windows restore. The "disks" in the Charon environment are frequently only Windows container files, and can be backed up (and restored) from Windows without a trace. Also, perhaps your simulated environment was pointed to different files when you were running before the Windows crash.

I would check (and check with colleagues) if anything else is amiss.

- Bob Gezelter, http://www.rlgsc.com

[Wim Van den Wyngaert]

There is a system parameter called SMP_SANITY_CNT that is per default on 30 * 10ms.

May be VMS didn't get his hands on the CPU during that interval (windows did).
May be the parameter is too low on your site.

Fwiw

Wim without VAX and Sharon (prefer the Stone version).

[Volker Halle]

Miles,

may I repeat: PLEASE use meaningful titles for your questions. In this case, it could have been: 'CHARON-VAX crashed'

Did the Windows system crash as well or did just the CHARON-VAX emulator itself crash ?

Did you check the CHARON.LOG file in the CHARON-VAX\VAX directory ? Which version of CHARON-VAX are you running ? Was there a TOY related error message in CHARON.LOG

The CPUSANITY crash (29-JUL-2006) indicates, that you are running a SMP system (CHARON-VAX 66x0 ?). A CPUSANITY crash is taken, when one CPU in a SMP system suddenly stops working. What is the value of the SYSGEN parameter SMP_SANITY_CNT ?

$ MC SYSGEN SHOW SMP_SANITY_CNT

The fact, that you see timestamps from the future may indicaite, that the CHARON-VAX emulator had a problem reading the system time from the CHARON.DAT file. This can happen especially, if the CHARON-VAX emulator has not been cleanly stopped. Problems regarding this have been recently solved in V3.1.62 for CHARON-VAX/XM/XL/XK.

Check your current system time ( $SHOW TIME)

Volker.

[vmsserbo]

Thanks! For all the advise!

[Volker Halle]

Miles,

the CHARON-VAX 66x0 release notes state, that you should increase the SMP_SANITY_CNT to 3000 - did you do this ?

ftp://ftp.sri-gva.ch/Documentation/Release%20notes/30-17-046.html

Volker.

[vmsserbo]

SEATTL> MC SYSGEN SHOW SMP_SANITY_CNT
Parameter Name Current Default Min. Max. Unit Dynamic
-------------- ------- ------- ------- ------- ---- -------
SMP_SANITY_CNT 300 300 1 -1 10ms.

[Volker Halle]

Miles,

the link to the CHARON-VAX 66x0 Release Notes posted above does not work. Here is a better one:

http://www.softresint.com/charon-vax/app_notes.htm#rn

The release note explain, that you should set SMP_SANITY_CNT 3000 to prevent CPUSANITY crashes.

Volker.

[Volker Halle]

Miles,

the link to the CHARON-VAX 66x0 Release Notes posted above does not work. Here is a better one:

http://www.softresint.com/charon-vax/app_notes.htm#rn

The release note explains, that you should set SMP_SANITY_CNT 3000 to prevent CPUSANITY crashes.

Volker.

[vmsserbo]

Event Type: Information
Event Source: DrWatson
Event Category: None
Event ID: 4097
Date: 8/8/2006
Time: 4:35:12 PM
User: N/A
Computer: CHARON-SEATTL
Description:
The application, C:\Program Files\CHARON-VAX\6630\vx6k630.exe, generated an application error The error occurred on 08/08/2006 @ 16:35:12.218 The exception generated was c0000005 at address 7C3429C1 (MSVCR71!memmove)

[vmsserbo]

My previous email I sent was with the error we got yesterday

Do I need to increase it to 3000? What is the command to do that and do I need to rebbot the server??

[vmsserbo]

Can you email me the release notes about setting it to 3000? I can't open it up

[Duncan Morris]

Miles,

here are the relevant lines from the release notes:

Setting SMP_SANITY_CNT:

Configure VMS to accept larger CPUSANITY timeouts as follows:

>>> B/R5:1 DU0

- Initializing system

....

- SYSBOOT>SHOW SMP_SANITY_CNT

Parameter Name Current Default Min. Max.
SMP_SANITY_CNT 300 300 1 -1 10ms.

- SYSBOOT>SET SMP_SANITY_CNT 3000

- SYSBOOT>CONT

- ...

*OR*

Modify the modparams.dat to extend the SMP_SANITY_CNT, autogen the system and reboot.


Regards,

Duncan

[Volker Halle]

Miles,

increasing SMP_SANITY_CNT to 3000 should prevent the CPUSANITY crash you've seen on 29-JUL-2006 09:02:12.10.

Please use the following to correctly change this parameter:

$ EDIT SYS$SYSTEM:MODPARAMS.DAT

add line (at the end of that file):

SMP_SANITY_CNT = 3000 ! Prevent CPUSANITY crash on CHARON-VAX

Run AUTOGEN to make this change permanent:

$ @SYS$UPDATE:AUTOGEN SAVPARAMS SETPARAMS FEEDBACK

or to manually change the parameter:

$ MC SYSGEN
SYSGEN> USE CURRENT
SYSGEN> SET SMP_SANITY_CNT 3000
SYSGEN> WRITE CURRENT
SYSGEN> EXIT

To activate the changed parameter, you need to reboot the system.

Regarding the Windows Exception generated by vx6k630.exe, please log a call with your CHARON-VAX reseller or SRI, if you have a support contract.

Volker.