We have an application that has caused intermittent problems on Windows 98.
Occasionally the R6 client on Win98 will crash while editing a document in a particular application we have. If the workstation is rebooted and the user tries to open the document that was being edited, R6 will crash again. However, if instead of rebooting, the ntaskldr Windows task is killed and Notes is restarted, the document can once again be accessed.
My first question is - does anyone know why Notes behaves differently when you kill the ntaskldr task and re-start Notes vs. restart Windows?
My second question - has anyone else experienced similar problems?
Some other info about this problem:
-
We have not been able to succesfully duplicate the initial crash/corruption. It is very intermittent.
-
This is only a problem with R6 not R5. We have had the problem on 6.0.3, 6.0.2 CF2, and 6.0.1 CF2
-
All other documents in the database can be accessed without a problem
-
the form that causes the crash has an embedded view
-
we have experienced problem on both local and server replicas in multiple locations
-
if you replicate the document that causes the crash to another database replica and try to access it will crash the workstation
-
if you copy and paste the document and try to access the copy of the document, it will not crash the workstation (seems to be tied to the doc’s UNID?)
-
we can right-click on the “corrupt doc” in the view and look at its properties (nothing looks out of the ordinary)
-
if another Win98 accesses the document it too crashes until the ntaskldr task is killed - this won’t happen if the document is accessed by XP or Win 2000
-
we have tried runing fixup/compact on the database + client workspace with no luck
-
we only have had this problem with this application
-
if we run the LS debugger on the document while it is being opened it doesn’t get past the Global Initialize (which has no code) before crashing
Very strange!
Any thoughts/ideas would be greatly appreciated!
Thanks!
-Adam