RE: UV Crash on W2K3
I raised a case with IBM when I tracked down our crash to a file not converting properly when we moved from AIX to W2K3. This file had a trigger, which we had removed, but even so it did not convert completely. The result was an unpredictable crash of the session which took a long time to track down. I am advised by IBM support that currently there is not a way to move UniVerse files, with Triggers, to another Machine Class as fnuxi does not support SQL tables. They do not regard this as being a bug. The application of a Trigger to a standard UniVerse file turns this into a pseudo SQL Table with a SICA block within the file's header, to store the trigger information. The CREATE/DROP TRIGGER commands are updating the uv/SQL/CATALOG/UV_TABLES and UV_USERS files. They have provided me with a workaround to clear the sica pointer from the file header, which will allow me to make the changes when moving between machine classes. I am going to request an enhancement to make this available to other users without them needing to write additional hacks to get around this situation. This is the second time we have had to move between machine classes within 3 years. It happens - we need to be able to handle it simply without the aggravation it has caused us. At the very minimum IBM need to cover this in their documentation. If they have I didn't find it. UniVerse 10.1 on Windows 2003 Sara Burns Sara Burns (SEB) Development Team Leader Public Trust Phone: +64 (04) 474-3841 (DDI) Mobile: 027 457 5974 mailto:[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Information contained in this communication is confidential. If you are not the intended recipient the information should not be used, disclosed, copied or commercialised. The information is not necessarily the views nor the official communication of Public Trust. No guarantee or representation is made that the communication is free of errors, virus or interference. -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
RE: UV Crash on W2K3
Sara, Did you try a RESIZE on the file? And if so, what was the result? Andrew Mack Deputy Sys Admin NZDF HRMIS 04 2371 914 From: Sara Burns [EMAIL PROTECTED] Reply-To: U2 Users Discussion List [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED] Subject: UV Crash on W2K3 Date: Tue, 20 Apr 2004 12:23:10 +1200 After extensive effort, creating an example to send to IBM we believe we have found the problem. One of our files had a trigger. Due to this the big/little endian conversion, using CONVERT.FORMAT did not seem to completely work. However this was not obvious. We thought we had reconverted all files with triggers, but maybe we missed this one. The following commands worked without obvious problems. uvbackup COUNT BAD.FILE CT BAD.FILE * LIST BAD.FILE SORT BAD.FILE However FILE.STAT BAD.FILE failed with a message, but did not terminate the session. uvfixfile did not appear to do anything. No error, just nothing. ACCOUNT.FILE.STATS worked without obvious error but the output in STAT.FILE showed incomplete information. The result of this was that basic programs failed intermittently which made this very, very difficult to track down. I will be raising this with IBM as I believe it is a bug. Sara Burns (SEB) Development Team Leader Public Trust Phone: +64 (04) 474-3841 (DDI) Mobile: 027 457 5974 mailto:[EMAIL PROTECTED] Information contained in this communication is confidential. If you are not the intended recipient the information should not be used, disclosed, copied or commercialised. The information is not necessarily the views nor the official communication of Public Trust. No guarantee or representation is made that the communication is free of errors, virus or interference. -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users _ Listen to music online with the Xtra Broadband Channel http://xtra.co.nz/broadband -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
Re: UV Crash on W2K3
In a message dated 4/16/2004 3:38:16 PM Pacific Daylight Time, [EMAIL PROTECTED] writes: 1.is this a problem with PROCS - does anyone else run a UniVerse system on NT which relies on PROCS to launch several layers of programs? Sara can you explain this more? Perhaps with an example of exactly what you mean? Thanks Will -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
RE: UV Crash on W2K3
Hi all,
I have encountered a similar error when attempting to resize a dynamic file
on our system. We are running UV 10.0.15 on W2K3 and when, as Administrator
with Full Control, I try to RESIZE {filename} * * * a dynamic file I get
this message both on screen and in the Application event log: error 1004
UniVerse error: Unhandled Exception raised at address 0x00405CF7 : Access
violation. Attempted to read from address 0x10309044. Binary data is
processor CONTEXT structure..
I suspect that these events are related as I believe RESIZE attempts to
logto the account you are in and create a temp file. I have tried
different accounts, files and server setups. I have adjusted permissions and
programs, to no avail.
Any help would be appreciated as I have run out of ideas.
Cheers
Andrew Mack
Systems Admin
HRMIS
New Zealand Defence Force
ph 64 4 2371 914
cell 64 021 626 446
From: Sara Burns [EMAIL PROTECTED]
Reply-To: U2 Users Discussion List [EMAIL PROTECTED]
To: '[EMAIL PROTECTED]' [EMAIL PROTECTED]
Subject: UV Crash on W2K3
Date: Fri, 16 Apr 2004 12:02:37 +1200
We are starting testing our application on W2K3 and getting unexplained
problems. It is proving difficult to isolate the exact cause but we are
starting to see some trends. The result is the session crashing with the
following general message in the Event Log
Universe error; Unhandled exception raised at address 0x77bd4528
Access violation Attempted to read from address 0x00d8
Binary data is processor CONTEXT structure
Sometimes the address to be read is 0x
So far we have found that it is very much less likely to occur if we do a
LOGTO DEV rather than logging on directly to DEV.
Can anyone help with information about the difference between a direct
login
and a LOGTO which may have some bearing on this.
This situation does appear to be associated with running programs from PROC
and may be influenced by the size of the program. We ran a very simple 3
line PROC and 3 line program all night without a failure.
The general mechanism used as the framework for our application involves
launching programs via PROC and using PROCREAD and PROCWRITE to pass data.
This is an example of the LAYER.STACK
There are currently 1 uniVerse sessions; 1 interactive, 0 phantom
Pid.. User name.. Port name. Last command
processed
.
252 ned telnet:252 telnet:252 runt [ get.input
@
0x2F4 ]
Layer type... Program name Address...
BASIC run machine get.input 0x02F4
BASIC run machine MAINTLNK 0x4160
BASIC run machine runt 0x0B06
Verb
Proc
Proc
Proc
Command Language
Execute
BASIC run machine CALL.ALL.FUNCTION 0x00DC
Verb
Proc
Proc
Proc
Command Language
Running only basic programs, outside this framework, does not appear to
cause a problem.
We have not managed to get any useful information from DrWatson as this
does
not seem to work for a service.
Can anyone give hints as to how to get better information for a UniVerse
session which crashes so we can pass this to IBM. We tried chaning the
MALLOCTRACING flag to 1 but cannot find any output.
We also see at times the @ACCOUNT = 0, @LOGNAME = 0 and @PATH = 00 These
have changed from the original normal values.
We have set up a completely isolated network for W2K3 so everything must be
moved to these machines the old way - floppy disk or CD-ROM. Our network
consists of a W2K3 PDC, a W2K3 member server with UniVerse 10.1 and a
XP-Pro
client. All are running on Compaq desktops 1.8GHz, 248Mb RAM
Initially we tried the Personal Edition 10.0.10 on a W2K3 PDC.
We then received the 10.1 version from IBM and tried on the W2K3 PDC
Finally we tried the 10.1 version on a W2K3 member server.
All gave the same results.
I am also posting this on the U2UG so I can view any replies over the
weekend as it is now Friday midday.
Sara Burns
Sara Burns (SEB)
Development Team Leader
Public Trust
Phone: +64 (04) 474-3841 (DDI)
Mobile: 027 457 5974
mailto:[EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]
Information contained in this communication is confidential. If you are not
the intended recipient the information should not be used, disclosed,
copied
or commercialised. The information is not necessarily the views nor the
official communication of Public Trust. No guarantee or representation is
made that the communication is free of errors, virus or interference.
--
u2-users mailing list
[EMAIL PROTECTED]
http://www.oliver.com/mailman/listinfo/u2-users
_
Listen to music online with the Xtra Broadband Channel
http://xtra.co.nz/broadband
--
u2-users mailing list
[EMAIL PROTECTED]
http://www.oliver.com/mailman/listinfo/u2-users
RE: UV Crash on W2K3
Hi Sarah What is the application currently running on and what version. (ie Win2000 Uv9.4?) and what flavour. In the past Loging in went through a login paragraph LOGIn. I think around release 1, the login paragraph went by the account name not LOGIN. Also with direct Login and Logto one would go to LOGIN paragraph and the other would goto the Account name Paragraph. If you are using the lOGIN paragraph to open files or set environment parameters, this may not have happened causing the fatal. Regards David Jordan Managing Consultant [EMAIL PROTECTED] Dacono Holdings Pty Ltd Business Technology Consulting PO Box 909 Lane Cove NSW 2066 Australia Ph 61 2 9418 8329 Fax 61 2 9427 2371 www.dacono.com.au -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sara Burns Sent: Friday, 16 April 2004 10:03 AM To: '[EMAIL PROTECTED]' Subject: UV Crash on W2K3 We are starting testing our application on W2K3 and getting unexplained problems. It is proving difficult to isolate the exact cause but we are starting to see some trends. The result is the session crashing with the following general message in the Event Log Universe error; Unhandled exception raised at address 0x77bd4528 Access violation Attempted to read from address 0x00d8 Binary data is processor CONTEXT structure Sometimes the address to be read is 0x So far we have found that it is very much less likely to occur if we do a LOGTO DEV rather than logging on directly to DEV. Can anyone help with information about the difference between a direct login and a LOGTO which may have some bearing on this. This situation does appear to be associated with running programs from PROC and may be influenced by the size of the program. We ran a very simple 3 line PROC and 3 line program all night without a failure. The general mechanism used as the framework for our application involves launching programs via PROC and using PROCREAD and PROCWRITE to pass data. This is an example of the LAYER.STACK There are currently 1 uniVerse sessions; 1 interactive, 0 phantom Pid.. User name.. Port name. Last command processed . 252 ned telnet:252 telnet:252 runt [ get.input @ 0x2F4 ] Layer type... Program name Address... BASIC run machine get.input 0x02F4 BASIC run machine MAINTLNK 0x4160 BASIC run machine runt 0x0B06 Verb Proc Proc Proc Command Language Execute BASIC run machine CALL.ALL.FUNCTION 0x00DC Verb Proc Proc Proc Command Language Running only basic programs, outside this framework, does not appear to cause a problem. We have not managed to get any useful information from DrWatson as this does not seem to work for a service. Can anyone give hints as to how to get better information for a UniVerse session which crashes so we can pass this to IBM. We tried chaning the MALLOCTRACING flag to 1 but cannot find any output. We also see at times the @ACCOUNT = 0, @LOGNAME = 0 and @PATH = 00 These have changed from the original normal values. We have set up a completely isolated network for W2K3 so everything must be moved to these machines the old way - floppy disk or CD-ROM. Our network consists of a W2K3 PDC, a W2K3 member server with UniVerse 10.1 and a XP-Pro client. All are running on Compaq desktops 1.8GHz, 248Mb RAM Initially we tried the Personal Edition 10.0.10 on a W2K3 PDC. We then received the 10.1 version from IBM and tried on the W2K3 PDC Finally we tried the 10.1 version on a W2K3 member server. All gave the same results. I am also posting this on the U2UG so I can view any replies over the weekend as it is now Friday midday. Sara Burns Sara Burns (SEB) Development Team Leader Public Trust Phone: +64 (04) 474-3841 (DDI) Mobile: 027 457 5974 mailto:[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] Information contained in this communication is confidential. If you are not the intended recipient the information should not be used, disclosed, copied or commercialised. The information is not necessarily the views nor the official communication of Public Trust. No guarantee or representation is made that the communication is free of errors, virus or interference. -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users -- u2-users mailing list [EMAIL PROTECTED] http://www.oliver.com/mailman/listinfo/u2-users
