cancel
Showing results for 
Search instead for 
Did you mean: 

Escape Character (?') causes High-CPU thread-lock in TPM:B2B-Addon SP3 - Upgrade to SP4

JuanDK78
Participant
0 Kudos


I wanted to share an issue I recently saw which might impact other system using the B2B add-on, not updated to SP4, and using the Trading Partner Manager to manage the conversion and acknowledgment of messages.

Symptoms:

- FTP channel was not picking up the file.

- EDI messages stayed in status "Delivering" in the message monitor

- System was very slow, CPU usage reached 100%

The issue was seen with some bulk and single EDI HANCOM orders.


Initially when the FTP channel was not picking up the file, we though is an an encoding error FTP authorization. But the system eventually turned very slow, the java processes in the OS reached 100% of CPU utilization. We could only recover the resources by restarting the instance.


The communication channel was using the module EDISearchParametersModule, which was disabled to continue the debug process.


The files were picked up but then stayed in status “delivering” when they reached the EDISeparator channel; and trigger again high-cpu usage in the server.

In  the transaction had to be changed manually from status DLNG to NDLV by DB modification of the transactions status, before restarting.

After a lot of debugging I isolated the issue in two B2B Add-on Modules, which we used as part of our TPM solution.

EDISearchParametersModule

TPMcontentAccessModule

What triggered the issue was  the address in the message contained  the character “ ‘ “ (apostrophe), which is the end of segment qualifier in EANCOM HANCOM.

And the character escape character “?” was managed properly in the listed modules.

example from the message:

NAD+DP+579000123456::9++FØTEX RO?'S TORV+RO?'S TORV 41+ROSKILDE++4000+DK'

NAD+BY+579000123456::9++FØTEX RO?'S TORV+RO?'S TORV 41+ROSKILDE++4000+DK'

When the two modules are removed from the CC channels and the standard EdifactConverterModule was used, the message flowed without problems.

I moved the solution to our sandbox system which has a  B2B SP4, to debug further but the message worked properly.

We moved to an immediate upgrade of the customer system to SP4 to solve the issue.


I hope this can help some of you and prevent possible severe issue in productive system.


Note: the SP4 upgrade was done in all 3 system within 10 hours with less than 15 minutes of down time in each system.

Accepted Solutions (0)

Answers (1)

Answers (1)

Dimitri
Active Contributor
0 Kudos

Hi Juan,

Thank you for sharing...

Perhaps you should consider writing a document or blog in the future ( ).

Kind regards,

Dimitri