Recently I did a policy server upgrade from 12.52 SP1 CR1 to 12.52 SP1 CR9. The same was updates on Policy Store and Admin UI as well.
We have Introscope Agent running on Policy Server and Admin UI using which we monitor the servers performance and services.
We are on Introscope Version: 184.108.40.206 with CA Wily Agent 220.127.116.11 installed on the policy server.
After upgrading the Policy Server is not getting monitored and when checked the logs for EPA Agent it say,
10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] The Agent will attempt to determine its name.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] Unable to determine the Agent name from the System Property "com.wily.introscope.agent.agentName" because that property was either not found or had no value.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] Unable to determine the Agent name from a user specified System Property because the property "introscope.agent.agentNameSystemPropertyKey" in the Introscope Agent Profile was either not found or had no value.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] The Agent will not attempt to automatically determine its name from the Application Server because that feature has been disabled in the Introscope Agent Profile.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] Using the Agent name "SiteMinderReporter" based on the "introscope.agent.agentName" property in the Introscope Agent Profile.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] Started Error Reporting service10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent] Initial agent name set to SiteMinderReporter10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent.Isengard] Initiating connection attempts to the Introscope Enterprise Manager.10/12/18 08:10:57 AM GMT+01:00 [INFO] [IntroscopeAgent.Isengard] The Agent reconnection delay is set to 15 second(s).10/12/18 08:10:58 AM GMT+01:00 [WARN] [IntroscopeAgent.IsengardServerConnectionManager] The Introscope Enterprise Manager rejected the requested agent name ("SiteMinderReporter"). The name has been re-assigned to "SiteMinderReporter%1".10/12/18 08:10:58 AM GMT+01:00 [INFO] [IntroscopeAgent.IsengardServerConnectionManager] Connected Agent to the Introscope Enterprise Manager at 10.251.152.10:5001,com.wily.isengard.postofficehub.link.net.DefaultSocketFactory. Host = "<hostname>", Process = "SiteMinderManager", Agent Name = "SiteMinderReporter%1".
I have checked the EPA Agent property file and the Agent Name set there is SiteminderReporter only. I tried changing it to SiteminderReporter%1 but then it gives error and creates another one by the name SiteminderReporter_1.
I updated the name again to SiteminderReporter_1.
Now on Introscope console I have 3 metrics by the name of SiteminderReporter, SiteminderReporter%1 and SiteminderReporter_1 and only SiteminderReporter has the correct metric which has to be monitored but that is shown greyed out.
Can you please let me know if any change is required on the APM side after a policy server upgrade or how can I resolve the issue.
This sounds like simply an agent name duplication problem.
The Enterprise Manager (EM) rejected the SiteminderReporter name after the EPAgent startup because it was already seeing a connected agent with the same name.
Do you have multiple instances of EPAgent running or should there only be only 1 instance?
You said that agent SiteminderReporter is greyed out. Are both SiteminderReporter%1 and SiteminderReporter_1 active or is only 1 active?
On APM Workstation Investigator use right mouse button on each of the 3 agent nodes
The option "Unmount ..." to manually unmount the agent will only be available if the agent is not connected (greyed out). So you can use that to unmount SiteminderReporter.
Make sure all instances of EPAgent are stopped and then SiteminderReporter%1 and SiteminderReporter_1 should also become greyed out and you can unmount them as well. If EPAgent is running on *nix use "ps-ef|grep java" to check EPAgent is definitely stopped. If EPAgent is running on Windows you can use Task Manager to checkfor java.exe processes.
Then in the IntroscopeEPAgent.properties file revert introscope.agent.agentName to be SiteminderReporter
Start the EPAgent and you should see just SiteminderReporter.
I hope that information helps
It is not the SiteminderReporter which is greyed out when the EPA Agent is running but a sub metric under SiteminderReporter by the name of Siteminder which is greyed out.
When the EPA Agent is running a new metric by the name of SiteminderReporter%1 gets created which does not have the sub metric Siteminder under it and hence the monitoring display nothing.
I tried to follow the steps mentioned above by you. Upon stopping the EPA Agent it is only SiteminderReporter%1 which gets greyed out, SiteminderReporter remains as it is(active) and hence I am not able to unmount it.
I unmounted SiteminderReporter%1 and then restarted the EPA Agent but it re creates the SiteminderReporter%1 and makes it active.
Please note the EPAAgent in the Introscope Properties file is still SiteminderReporter and I have not changed that.
Can you please upload a screenshot(s) of the expanded metric browser showing all relevant nodes to help understanding.
Before upgrade from 12.52 SP1 CR1 to 12.52 SP1 CR9 this was all working correctly?
I am wondering if it could berelated to the compatibility of old Wily SiteMinder Manager 12.04 with SiteMinder 12.52 SP1 CR9. The newer CA Single Sign-On for APM 13.2 does support that version:
PCM for CA Software - Agile Operations Product Compatibility Search
I will check into that.
PFB the snip of the Investigator window.
Yes, it was working as expected before upgrade.
If agent SiteMinderReporter stays active after you shutdown EPAgent I can only think there must be some other "rogue" agent process running.
The SiteMinderReporter|Host:IP Address metric will confirm the IP of the server where that is running so you can check for possible java processes on that server with "ps -ef|grep java" or "ps -ef|grep java|grep -i epa"
if that does not help then because of the impact of losing visibility of the main SiteMinder metric node I would suggest to open a support case for more detailed investigation by us.
Thanks for the suggestions.
There is no Java process running on the server.
I have raised a case with the Support team as well.