Get rid of JTA warning message for WebLogic

This is the second time I ran into the JTA exceptions. Basically, you will get this message every one minute when running your EJB on any version of WebLogic server:

<Warning> <JTA> <BEA-110486> <Transaction BEA1-05576B5644FBD4F5B49F cannot complete commit processing because resource [weblogic.jdbc.wrapper.JTSXAResourceImpl] is unavailable. The transaction will be abandoned after 67,910 seconds unless all resources acknowledge the commit decision.>

Last I did google it, get rid of it, and eventually forget about it at all. Now it's a good chance for me to pick it up and record it down. Hopefully next time I can handle it in no time without spending half an hour wondering around. I tried different ways like restart server, uninstall/reinstall application, etc. No luck at all.

This warning message is because something is wrong in your code, specifically of transaction. There must be some of the transaction still keep opening after some exception happens or similar alike. As a result, the server complain that there is a transaction still running which no one seems to be using it at all.

In order to get rid of this ugly message, just go to your domain: <your-domain>/servers/AdminServer/data/store/default, there should be a file like "_WLS_ADMINSERVER000000.DAT" there. The number will random. If you delete this file, and restart the server, problem solved!

GAE/J deployment transaction conflict error 409

Today when I tried to deploy my application onto the App Engine, a weird error prompt  up:

Unable to update app: Error posting to URL: https://appengine.google.com/api/appversion/create?app_id=metacdn&version=1&
409 Conflict
Another transaction by user featheast.lee is already in progress for this app and major version. That user can undo the transaction with appcfg.py's "rollback" command.
See the deployment console for more details

I haven't seen error before, and have no clue how it comes out. After several google page, I found a way to solve this problem.

1. Open your terminal and get into the directory of your project. ( For windows users, please do it in CMD accordingly).

2. Execute the sh script of appcfg.sh under with the parameter of rollback and war. You should prefix the path of your appcfg.sh, which is usually under the directory of eclipse plugin. ( Windows guys, continue to use cmd to replace the sh script)

3. Now after successfully rollback the deployment, you are free to do anything you want to now.

I guess the same story should be applied to Python as well, just simply use the appcfg.py instead of the script and things should work out.

PS: I got a ZipExeption during the rollback process with a warning that Could not find API version from .svn. Though it still solves my deploy issue.