1 Reply Latest reply on Dec 7, 2011 2:13 PM by rusboy

    How does jboss work on process level?

    rusboy

      Hello,

       

      I have the following situation..

       

      My configuration:

      2 servers with same H/W configuration (1GB, 60Gb HDD but different process)

      There are linux server.

      Server A is working on kernel with PEA

      Server B is working on kernel without PEA feature.

       

      Both servers have jboss.

       

      I started the jboss on both servers and when it was up I checked size of process (using top or ps command).

      I see that size of jboss process on Server A is 500Mb but on server B the size is 700Mb.

       

      Then, I got dump of both process (process A and process B accordingly).

       

      Next, I compared heap of both process. it looks like same (difference about 5-10 Mb..).

      Ok.. Then I got data from /proc/Jboss-ID/maps file..

      As you know this file contains:

      heap,

      stack

      and list of deployed jar..

       

      And I found the list of the jars is different between the process...

       

      It looks like jboss uploads the jar to process dinamically.. Not all at once..

      But I've not found proves for this teory...

       

      Anybody met same question?

      Could you share your expertise?

       

      Short comment:

      Size of process is became same after some time about 2-3 days (them's size is about 900Mb).

       

      Regards,

      Ruslan