AppD agent does not send data to AppD

[CommunityUser]

Hi,

I have got this wired problem. I can see my boxes which have AppD agent running ([Machine Agent v3.8.2.0 GA Build Date 2014-05-20 17:59:18], [Server Agent v3.8.2.0 GA #2014-05-20_17-53-50 r9ff4de878ca8b9fdc48c9a2a50a93fc1504e2f93 1
7-3.8.2.next]) on AppD. But no data is ever sent to AppD.

On the nodes tab, all information on column "App Agent Stauts", "App Agent Version" and "JVM / CLR" are blank.

In both machine agent and app server agent logs, I am seeing some warnings but do not know how to interpret them.

Here I attached those log files in /opt/appd/AppServerAgent/logs and /opt/appd/MachineAgent/logs. Can someone give me some clue what I may have missed.

Thanks.

[Arun_Dasetty]

Hi,

Can you qucikly do the following changes and see how it goes?

1) Sop both appserver agent and machine agent jvm for agent node and restart the appserver agent first with following -Djvm arg in place:
-Dappdynamics.agent.uniqueHostId=stg-use1c-pr-20-registr-01x00x67-0005.prv-openclass.com

Add above in addition to -javaagent entry in the jvm startup scripts of server and save the changes before restart

2) Now check appserver agent node shows up fine in UI and then start machine agent as follows:
shell> java -Dappdynamics.agent.uniqueHostId=stg-use1c-pr-20-registr-01x00x67-0005.prv-openclass.com -jar machineagent.jar

We suggested above changes as we see both appserver and machineagent host are resolved to different values but trying to register with same node name and restarting with same uniqueHostId arg should fix the issue:
[Thread-0] 20 Nov 2014 23:14:45,640 INFO XMLConfigManager -
Default Host Identifier Resolver using host name for unique host identifier [stg-use1c-pr-20-registr-01x00x67-0005.prv-openclass.com]
.. [Agent-Scheduler-1] 20 Nov 2014 23:13:31,144 INFO RegistrationChannel - setting unique host information Host Name [stg-use1c-pr-20-registr-01x00x67-0005]

Which is evident from below agent logs:

[AD Thread Pool-Global1] 20 Nov 2014 23:15:47,214  WARN ConfigurationChannel - ResponseReadException creating Response Wrapper [xb], : com.singularity.ee.rest.c: Error in controller in processing binary request AppAgent Config Data  - Application Component Node with name [stg-use1c-pr-20-registr-01x00x67-0005] already exists for Component with name [GRID: Registrar Service]. The existing Application Component Node has id [52877]. If you specified the JVM property [appdynamics.agent.uniqueHostId] for the Application Agent, you MUST set this JVM property to the same value for the Machine Agent that is associated with this Application

let us know if that information helps.

Regards,

Arun

[CommunityUser]

Arun,

I followed your suggestion and things work now. Even though I am seeing the following error occuring a couple of times in the machine agent log:
com.singularity.ee.rest.ResponseReadException: Error in controller in processing binary request Event Data - null
at com.singularity.ee.rest.controller.request.ABinaryControllerRequest.<init>(ABinaryControllerRequest.java:71)
at com.singularity.ee.rest.controller.request.EventDataRequest.<init>(EventDataRequest.java:38)
at com.singularity.ee.agent.systemagent.components.monitormanager.managed.ManagedMonitorDelegate$EventSender.publish(ManagedMonitorDelegate.java:438)
at com.singularity.ee.agent.commonservices.eventgeneration.EventGenerationService$EventPublisher.run(EventGenerationService.java:387)
at com.singularity.ee.util.javaspecific.scheduler.AgentScheduledExecutorServiceImpl$SafeRunnable.run(AgentScheduledExecutorServiceImpl.java:118)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:471)
at com.singularity.ee.util.javaspecific.scheduler.ADFutureTask$Sync.innerRunAndReset(ADFutureTask.java:315)
at com.singularity.ee.util.javaspecific.scheduler.ADFutureTask.runAndReset(ADFutureTask.java:150)
at com.singularity.ee.util.javaspecific.scheduler.ADScheduledThreadPoolExecutor$ScheduledFutureTask.access$101(ADScheduledThreadPoolExecutor.java:123)
at com.singularity.ee.util.javaspecific.scheduler.ADScheduledThreadPoolExecutor$ScheduledFutureTask.runPeriodic(ADScheduledThreadPoolExecutor.java:208)
at com.singularity.ee.util.javaspecific.scheduler.ADScheduledThreadPoolExecutor$ScheduledFutureTask.run(ADScheduledThreadPoolExecutor.java:238)
at com.singularity.ee.util.javaspecific.scheduler.ADThreadPoolExecutor$Worker.runTask(ADThreadPoolExecutor.java:683)
at com.singularity.ee.util.javaspecific.scheduler.ADThreadPoolExecutor$Worker.run(ADThreadPoolExecutor.java:715)
at java.lang.Thread.run(Thread.java:745)

My other concern is that other teams in Pearson did not have to do what you said and their instance worked. Our instances are all in EC2 and we do recycle IPs. Do you think it is possible that AppD gets confused by reused IP and that is why it "seems" that it works for other teams even though they did not follow what you suggested regarding unique host ID?

Thanks a lot.

George

[Arun_Dasetty]

Hi George,

Yes, your understanding is right it depends various factors how agents are associated in the context of any machine agents nodes to same other apps from same/other machine and the hosts there are resolving to restart of jvm/ec2. Hence the config varies depending on how agent is associting and how hosts are resolved, Hope that answers your query.

Regarding the error you see, if you see any issue metric reporting of machine agent please send logs in debug mode (change logger level from info to debug in log4j.xml under machien agent directory) to debug further.

Regards,

Arun