Clustered PO Abend with memory protection violation

[Jo]

hi

we have 8 clustered po's, and one of them did not run properly. they run in protected memory.
the message on the server console looks like follow "po2 addressspace removed because memory protection violation"
the po screen hung and nothing happen. no failover, nothing.
maybe someone have the same probs, or a solution.

regards
joerg

[Tim Heywood (NSC Sysop)]

Joerg,

Two quick questions, does this PO do this no matter which node the service is
run? IE if you manually load this PO in a different node from normal does it
still behave in the same manner? Secondly, which GW version and Which NCS
version are you using?

Tim

No Direct e-Mail Please!
--
****************
Tim Heywood
Independent Consultant
Scotland
(God's Country)
Novell Support Connection SYSOP
****************

In theory, practice and theory are the same,
In practice they are different!

[Jo]

Tim

#does this PO do this no matter which node the service is
#run?

yes, it happen on every node i tried (3 different nodes)
the po runs witout any probs when i load it not in protectet memory.

#Secondly, which GW version and Which NCS
#version are you using?

GW 6.5 sp2 beta (but it happen with 6.5 sp1 as well)
NW6 sp3 NCS 1.6

thanks
joerg

>>> Tim Heywood (NSC Sysop)<t...@no.spam.please.iqx.co.uk> 25.01.2004 21:58:42 >>>

[Tim Heywood (NSC Sysop)]

I assume that you have run GWCheck in stand alone mode to confirm that there are
no structural problems etc.

Have you applied the memory protection patch?
http://support.novell.com/servlet/filedownload/uns/ftf/nw56up3.exe/

Tim

[Jo]

Tim

Thanks for your input. i will try soon as possible.
this was helpfully:)

regards
joerg

>>> Tim Heywood (NSC Sysop)<t...@no.spam.please.iqx.co.uk> 27.01.2004 22:10:26 >>>

[Jo]

Tim

I install this new server.exe but without success. it is crashing again.
i don't know what next i can do.

thanks anyway
joerg

>>> Tim Heywood (NSC Sysop)<t...@no.spam.please.iqx.co.uk> 27.01.2004 22:10:26 >>>

[Tim Heywood (NSC Sysop)]

This is one of those little B* problems that you will have to chip away at :-/

If you can, over the W/E you want to take that PO office (but leave the volume
mounted. Delete all of the quickfinder indexes (POSTOFFICE\OFUSER\INDEX ) Then
run the latest version of the stand alone gwcheck against the post office for a
full structural check, with indexes and fix errors. Then run gwcheck a second
time with a contents check and fix errors.

I think that you will find that you have one exceedingly large attachment in
this PO that the quickfinder index program is choking on. To find this file;

"Steps taken to remove the corrupted messages:
1- Load the POA with /debug to enable diagnostic logging then run QF
Index.
2- Read the POA log file, the last user's database listed as being
indexed is the problem database.
3- Log into the user's account.
4- Archive messages in the Inbox, sent items folder, and cabinet
folders.
5- Identify the messages which report a 8E03 error or cause the hang
the GW client when archiving.
6- Delete the messages and empty the GroupWise trash folder.
7- Run QF Index and repeat steps 1-6 if the POA abends or hangs on a
different user's database. "

Tim

On Fri, 30 Jan 2004 14:33:26 GMT, Jo <te...@test.com> wrote:

>Tim
>

>I install this new server.exe but without success. it is crashing again.
>i don't know what next i can do.
>
>thanks anyway
>joerg
>

No Direct e-Mail Please!

[Jo]

tim

i've open an incident, may they can solve my probs.

thanks anyway
joerg

>>> Tim Heywood (NSC Sysop)<t...@no.spam.please.iqx.co.uk> 30.01.2004 21:53:33 >>>

[Tim Heywood (NSC Sysop)]

Thanks for letting us know Joerg, Please do come back and let us know what it
takes to resolve the problem.

Tim