Invoices Error out via BI but not all the time.

sheeba

Reputable Poster
Our invoices fail to print when routed thru BI Publisher - It is not a consistent issue - BI version prints invoices sometimes and it fails sometimes - we have not been able to identify the cause for BI version erroring out - however it works correctly when a non-BI version is run.


The jde.log always ends in this when the BI version errors out

5100/2068 MAIN_THREAD Wed Jan 21 20:10:31.413000 ntexhnd.cpp101
Exception code: C0000005 ACCESS_VIOLATION

5100/2068 MAIN_THREAD Wed Jan 21 20:10:31.944000 ntexhnd.cpp151
====> Exception occurred with call stack:

prt_AllocatePrintInfo+87C C:\JDEdwards\E812\DDP\system\bin32\jdekrnl.dll

INFO: Done setting IPC Handle State structures to abandoned. Process exiting. iParam: 0000004296


Has anyone seen this? How did you fix the problem. I have an Oracle call open on this. Will let you know if I here back from them.

However, we think this happens when its a large data set - but I believe at one time it worked with 10100+ lines. So really want know when it fails .

thanks
sheeba
 
Hi sheeba,

I've the same problem.
From my opinion this issue is not closely related to
large selection but is related to action like fetch of attachment..

I've alse an open call with oracle but they told
me that they can't replicate the problem in
any way.
 
I tried another thing.

Took off the BI object from P95620 and let the invoice run. It ran successfully - RD with status D. It created the xml data also but when I tried to republish it gave me a system error on the RD jobs view. Saved the xml off to a folder and tried to preview using word - it errored with Java memory error. the XML output was 34MB big. I was able to republish the 8MB xml output thou.
frown.gif


yep - I am waiting for an answer from Oracle . Will update you as soon as I hear from them

thanks
sheeba
 
Are you doing any data sorting in your template? I've been told that that can cause blowups.... if so, do the sorting with data sequencing at the UBE level and don't do it in your template.
 
Hy ... Since my problem is with the p95620 i go ask one doubt in this post! when i submit a report it stays always in D status , dont pass to status P ... I dont know why? do u know any configuration that mades this work fine? im using version 9 of jde. I tryed to find solutions in metalink but i cant see nothing near to my problem!
 
no data sorting on template - all is done in JDE. again this is not a consistent issue - it seems like it depends on the input dataset - it crashes if its too huge. waiting on oracle for a definite response on this case.
 
Hi sheeba,

I would like to update you about this issue.
This kind of error seems to be related to wrong varariable initialization and it has
no relation with large data selection.

In one case I've found some null value in a database field (due to manual upload I think);
in other case the memory violation error was caused by RV not initialized in our UBE.
The last example that I found is a "Get MO attachment" performed with a blank key.

In all cases I've no problem with PDF process but if I choose XML processing I had the "C0000005 Access Violation".

So in my opinios, the XML engine seems to be more
sensitive that standard UBE processing.

I have another cases reported but I'm still trying to understand what are the problem for those ones. No garbage data in that cases...
 
Curious if you ever found out the culprit? I am getting the same error message when running a custom report (not using XML Publisher). The errors occur when the data set is large, but is not consistent. I haven't found much on the KG or elsewhere.
 
Have also experienced this and learned you need to change the version(s) to run in a single threaded queue to resolve. The error occurs because multiple versions attempt to run the same template at the same time.
 
Back
Top