We recently just upgraded our environment to the Hyperic 4.1.2 release. I noticed that when we install a new agent the provider never gets configured when run using the Java wrapper process (custom configuration). I noticed this is the result of the following change made to the org.hyperic.hq.bizapp.agent.AgentClient class:
// Only force a setup if we are not running the agent in Java Service Wrapper mode
if(providerInfo == null && !WrapperManager.isControlledByNativeWrapper()){
SYSTEM_OUT.println();
return FORCE_SETUP;
} else {
redirectOutputs(bootProps); //win32
return 0;
This change was introduced in the trunk in SVN r9381 and updated in SVN r9956. I don't understand the purpose of this change since it breaks compatibility with previous agent releases and the agent release doesn't keep the provider setup from running anways because the setup is forced via the external bin/hq-agent.sh script.
start() {
...
${SETUP_IF_NO_PROVIDER_CMD}
}
I'm requesting a fix for this to allow the provider setup to be reverted back to the original behavior. I've attached a patch to fix this.
// Only force a setup if we are not running the agent in Java Service Wrapper mode
if(providerInfo == null && !WrapperManager.isControlledByNativeWrapper()){
SYSTEM_OUT.println();
return FORCE_SETUP;
} else {
redirectOutputs(bootProps); //win32
return 0;
This change was introduced in the trunk in SVN r9381 and updated in SVN r9956. I don't understand the purpose of this change since it breaks compatibility with previous agent releases and the agent release doesn't keep the provider setup from running anways because the setup is forced via the external bin/hq-agent.sh script.
start() {
...
${SETUP_IF_NO_PROVIDER_CMD}
}
I'm requesting a fix for this to allow the provider setup to be reverted back to the original behavior. I've attached a patch to fix this.