The reason that I think this may have something to do with Camel is because, if you look at the call tree in the screen shoot, quite a number of org.apache.camel DefaultErrorHandler and RedeliveryErrorHandler methods are called before the final sun.misc.Unsafe.park. Any pointer would be really appreciated.
________________________________ From: Pham Ngoc Hai <pngoc...@yahoo.com> To: "users@camel.apache.org" <users@camel.apache.org>; Pham Ngoc Hai <pngoc...@yahoo.com> Sent: Friday, April 20, 2012 4:34 PM Subject: Re: sun.misc.Unsafe.park CPU usage We are using Java HotSpot(TM) 64-Bit Server VM version 17.1-b03 on CentOS 5.5 64bits ________________________________ From: Pham Ngoc Hai <pngoc...@yahoo.com> To: "users@camel.apache.org" <users@camel.apache.org> Sent: Friday, April 20, 2012 4:27 PM Subject: sun.misc.Unsafe.park CPU usage Hi, We are using Fuse servicemix 4.2.0-fuse-02-00 In this SMX, we use multiple camel components like cxf, ativemq, file, ftp, smtp, xslt... Some time JVM CPU usage jumps up to 250% and stays there for a few hours and gets back to normal. Once in a while when the system becomes irresponsive we have to kill the JVM process. We use Solaris Studtio to do profiling and the see that JVM spends 40% of CPU time at sun.misc.Unsafe.park link: http://www.phamngochai.net/myfiles/sun.misc.Unsafe.park.jpg Is this a JVM bug or is this camel? Thanks.
