[jira] [Created] (BROOKLYN-555) Tomcat entity stop / start fails due to jmx.service.url sensor still has the IP from original VM

["Duncan Grant (JIRA)" <ji...@apache.org>]

Duncan Grant created BROOKLYN-555:
-------------------------------------

             Summary: Tomcat entity stop / start fails due to jmx.service.url sensor still has the IP from original VM
                 Key: BROOKLYN-555
                 URL: https://issues.apache.org/jira/browse/BROOKLYN-555
             Project: Brooklyn
          Issue Type: Bug
    Affects Versions: 1.0.0
            Reporter: Duncan Grant


Quoting https://issues.apache.org/jira/browse/BROOKLYN-552?focusedCommentId=16246734&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-16246734
"I tried reproducing this in AWS, with Tomcat. When I called stop and then start (i.e. such that it terminated the VM and spun up a new one), it went on-fire.
This is because it uses JMX for its health check, but the jmx.service.url sensor still has the IP from the first VM:
jmx.service.url: service:jmx:jmxmp://ec2-34-252-168-149.eu-west-1.compute.amazonaws.com:31001
webapp.url: http://ec2-52-49-28-164.eu-west-1.compute.amazonaws.com:8080/
This is presumably because things like JmxHelper will only recalculate the url and set the sensor UsesJmx.JMX_URL if that sensor is null.
However, I tried setting that sensor to null manually and then calling restart() effector - but it did not reset the jmx.service.url so there may be more to fixing it.
I didn't see the entity id or the run.dir change at all, or the security groups / keypairs mess up in any way."



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)