UBE EventRule memory violation

Hi List,

we upgraded our application server from NT to W2K3 3 weeks ago.
Last night I've got an error within our customized invoice print. It stopped with the error message printed within the PDF:

--
ERROR:
EventRule : The ER line [743] for Event [13] in Section [On Payment Terms] has caused a memory violation.
--

We run this report approx. 300 times a day. We never got this kind of message and are running that since years.

How to find a reason for that? Also, line 743 is a comment line. Last full package is about 3 weeks old and there was no update on that report. Full package has been created in DEP7333 on deployment server.

Please give me a tip were to look.

Thanks, Michel

Did you build a new full package after you migrated to Win2k3. I suggest you do.

Have been building 2 for PD7333 for testing and a new one on the weekend of the golive after importing oracle.
Have build an update pack for the invoice print yesterday as well.

The main changes in our new data center are:
-using Oracle RAC for Production instance (B7333PRD)
-using two NICs within each server. So each server is wihtin our network and within the ASP-Admin network (admin, backups, monitoring).
-every machine is running W2K3. Within the old install we had Oracle and a second application server running on W2K3. Citrix n W2K and JDE main app server and deployment server on NT.

First you can narrow down the type of problem by determining if ER line [743] for Event [13] in Section [On Payment Terms] is a business function or not.

1. If it is a business function call, you need to debug the business function to see where it is blowing up. The good news is that since the problem is repeatable, this should be relatively straightforward. Once you see the problem, then either you will know how to fix it or you can post the code to the forum so we can review the code issue.
2. If the line of ER is not a business function, then this would tend to point to either a Service Pack issue with Win2k3 or perhaps a spec issue (such as an existing Data Dictionary item problem that only now shows up.)

In either case, different operating systems can react differently to memory violations. WINNT may have been covering up the memory violation all along...

Cheers,

Chris

Hi Bruce,

the strange thing is that line 743 is a comment line within that section. The issue has been a one-off so far, so I cannot replicate/debug. Hope it does not come up again.

There is still one issue I cannot explain. 20 of 2000 UBEs per day are set to status E. The strange thing is that runbatch is running fine and updates the job status to Done and then the jdequeue process that spawned the runbatch updates the status to Error. Found that out by enabling jdedebug.log on the application server.

So what the hell is the reason for jdequeue to update the status to E?
Perhaps there is a connection between these issues?