Posts

Showing posts from January, 2009

Findbugs XSLT style Error java.lang.reflect.InvocationTargetException

Tried running out-of-the-box Findbugs 1.3 with Ant 1.6 and Ant 1.7, on JDK 1.4 and JDK 1.5 It throws an error for the task findbugs.report which is a simple XSLT transform using a stylesheet. Peculiarly this fails only on Windows but works fine on Unix. build.xml contains: <style basedir="${build.report.temp}" destdir="${build.report}/findbugs" includes="findbugs-post.xml" style="${script.tools}/findbugs/tools/xsl/report.xsl"> <param name="messages.xml" expression="${script.tools}/findbugs/tools/xsl/messages.xml"/> <param name="output.dir" expression="${build.report}/findbugs"/> </style> The error is [style] Processing D:\Construction\reports\ebilling\tmp\findbugs-post.xml to D:\Construction\reports\ebilling\findbugs\findbugs-post.html [style] Loading stylesheet D:\Construction\buildtools\bin\findbugs\tools\xsl\report.xsl [style] : Error! j...

Tag Clouds Jan 2009

Image
www.wordle.net/create makes pretty tag clouds out of any URL. Here is a comparison of my Twitter stream against my Blogs. Should do this once in a while

App Server High CPU or Low Threads - Support Team Actions

Image
Occasionally Weblogic/JEE app servers run into High CPU (>70% or as per your app-defined thresholds) , then below is a list of actions and investigations that the support teams need to do. Check prstat –L –p Capture into a file if required. Output kill -3 into a file i.e. take thread dump. On Weblogic 10,you can do this from the Admin console as well Output pstack into a file Repeat this 4 times at 5 second intervals. At the end, you will have 5 sets of prstat, TD and pstack files. Run the TDs through Samurai or TDA to identify deadlocks. These tools show a graphical view of what each thread is doing at each point of time (when the thread dump was taken). The image below from Samurai shows RED blocks which are Stuck, and Green with < in them. Both of these need to be analyzed by clicking on the link in blue and seeing what line the thread is at. The Green ones with < might not be stuck, but they were at the same line as in previous thread dump and that...