IBM Tivoli Composite Application Manager for Application Diagnostics, Version 7.1.0.1

---

Not reporting data problems

The following content provides information about not reporting data problems in Tivoli Enterprise Portal.

Datasources workspace is blank

The question: Why does the Datasources workspace in Tivoli Enterprise Portal for Tivoli Enterprise Monitoring Agent for ITCAM for WebSphere Applications contain no data?

The answer: Check the following points:

1. Tivoli Enterprise Monitoring Agent and data collector code level must be a supported combination, as per fix packs readme files.

2. Datasources must be defined in the WebSphere;

3. Tivoli Enterprise Monitoring Agent Monitoring Level must be set to L2;

4. JDBC collection must be enabled for instrumentation. That is, in DC_home/runtime/svr/custom/toolkit_custom.properties, make sure that com.ibm.tivoli.itcam.toolkit.ai.enablejdbc=true. If this setting does not exist in your toolkit_custom.properties, it is set to true by default;

5. Sampling rate on the Tivoli Enterprise Monitoring Agent side needs to be high enough. For Request Data to be displayed in workspaces, enough user transactions have to be monitored and collected by the Request Aggregator in the Data collector. The Request Collection Sampling Rate (%) setting determines how many requests have to be sampled. If the user load is low and this percentage is set to 2% (the default value), then it is likely that the time duration between transactions getting monitored is large. Hence, these workspaces might not display data for a long time. If you suspect that the Request Collection Sampling Rate is the cause, you can set it to 100% and see if the workspace is populated with data;

6. The application must use Datasources via J2EE Connectors. Consult the application developer to determine if your application uses J2EE connectors to attach to the DB, or its own code. Even if data sources are defined in WAS, the application still might not be using them if the application attaches using its own code. If the application is using WAS data sources and you see SQL activity on the Resource Analysis Tivoli Enterprise Portal workspace, then you can see something on the datasource workspace.

EJB Modules workspace and JCA Connection Pools workspace show no data

The question: Why is there no data on EJB Modules workspace and JCA Connection Pools workspace on Tivoli Enterprise Portal?

The answer: This problem is a limitation of JBoss 5.1. The JCA connection Pools, EJB Modules, and Servlets/JSPs > Web Application tables do not report any data because some MBeans are missing in JBoss 5.1.

Empty workspaces with message KFWITM217E

The question: Some workspaces in Tivoli Enterprise Portal related to ITCAM Agent for WebSphere Applications Monitoring Agent are empty and show error message KFWITM217E.

The answer: KFWITM217E is a generic message and you need to check the return code (rc) to understand the actual root cause of the message. Where do you look up this return code? Usually it is reported as appended text to the error message in the same workspace or on its bottom status area.

For example:

KFWITM217E Request error: SQL1_OpenRequest failed rc=3000

The root cause will likely be an incorrect Historical Data Collection configuration for those historical views showing the problem.

Check if the historical data collection is started on Tivoli Enterprise Portal for the Attribute Groups used in the view that is showing the error.

EJB Containers > Container transactions and EJB Containers > Container objects Pools in Tivoli Enterprise Portal do not enable historical data collection by default.

For rc=3001 it probably means that Historical Data Collection is properly configured and started already but there are no still data available to be shown in the view.

Please make sure there is activity for those data sources and wait for a time at least equal or greater than the Collection Interval specified.

Garbage Collector analysis not reporting data

The problem: The GC analysis workspace depends on the GC log collection feature, which is (optionally) enabled during GC configuration. If GC log collection is not enabled, then only few metrics (PID and Java heap sizes) will be reported from this workspace.

The solution: By default, GC analysis is configured with a 60 second, fixed collection interval; take this into account when navigating into a linked workspace (Allocation Failures, GC Cycles, etc.). Allocation Failures/GC Cycle workspaces are supported for IBM JVM only. For JVM 1.3, ensure that the GC log file location is defined in the <DC_home>/runtime/<platform.node.server>/<platform.node.server>.kwjdc.properties properties file, as follows:

# Set this for JDK 1.3 where GC log file name can't be 
# specified on the command line 
TEMAGCCollector.gclog.path=C:/PROGRA~1/ibm/tivoli/common/CYN/logs/gc.log

Log Analysis table shows no data on IBM i 5.4

The question: Why does the Log Analysis table show no data after installing ITCAM Agent for WebSphere Applications on IBM i 5.4?

The answer: It is a limitation. Log Analysis table is not available because Tivoli Enterprise Monitoring Agent is not supported on IBM i systems.

No data returned for Lock Analysis

The problem: In the Web User Interface of ITCAM for WebSphere Applications, select Problem Determination > Server Activity Display > Lock contentions, and the following error message is received:

CYNVE0850E: There are no classes instrumented for lock analysis.

The solution: Check the following points:

• data collector must be running at L2 or L3 monitoring level;

• Lock analysis must be enabled, specifically the file DC_home/runtime/app_server_version.node_name.server_name/custom/toolkit_custom.properties must contain the following lines:

  am.camtoolkit.gpe.customxml.lock=/opt/IBM/itcam/WebSphere/DC/itcamdc/etc/lock_
  analysis.xml com.ibm.tivoli.itcam.toolkit.ai.enablelockanalysis=true 

  and file DC_home/itcamdc/etc/lock_analysis.xml must exist.

• If the code level is Fix Pack 4 or higher, make sure that the DC_home/runtime/app_server_version.node_name.server_name/custom/*datacollector.properties file contains the following lines, which have been introduced together with Lock Analysis Enhancements:

   The following properties define how much data Lock Analysis will collect at
   various MOD levels.
  
   internal.lockanalysis.collect.LN.lock.events = true | false
   indicates whether or not lock acquisition and release events
   will be collected at MOD level 'N' (1, 2, or 3).
  
   internal.lockanalysis.collect.LN.contend.events = true | false | justone
   indicates whether or not lock contention events
   will be collected at MOD level 'N' (1, 2, or 3). If this  property is set to a value of 'justone', it indicates that only
   one pair of contention events is created for a request that encounters
   contention acquiring a lock. If this property is set to a value of
   of 'true', then multiple pairs of contention records may be created,  if there are multiple threads that acquire the lock prior to the request.
  
   internal.lockanalysis.collect.LN.contention.inflight.reports = true | false
   indicates whether or not the inflight lock contention report is 
   supported from the System Activity Display at MOD level 'N'
   (1, 2, or 3).

  The default settings are as follows:

  internal.lockanalysis.collect.L1.lock.events=false
  internal.lockanalysis.collect.L1.contend.events=false
  internal.lockanalysis.collect.L1.contention.inflight.reports=false
  
  internal.lockanalysis.collect.L2.lock.events=true
  internal.lockanalysis.collect.L2.contend.events=true
  internal.lockanalysis.collect.L2.contention.inflight.reports=true
  
  internal.lockanalysis.collect.L3.lock.events=true
  internal.lockanalysis.collect.L3.contend.events=true
  internal.lockanalysis.collect.L3.contention.inflight.reports=true

Resource metrics not reported

The problem: Resource analysis metrics are not reported

The reason:

• PMI workspaces report data in on-demand collection mode with a 30 second sample age - unlike request and GC workspaces which have fixed intervals.

• For WebSphere, make sure that PMI is enabled in application server configuration.

• Some workspaces are available for specific WASs versions only:

  • Enhanced PMI (Alarm Manager, DCS Stack, Platform Messaging, and so on.) are available for WAS 6.0 or higher.

  • Workload Management workspaces are not available for WAS Base releases.

Request metrics not reported

The problem: Request analysis metrics are not reported.

The solution: Check your Tivoli Enterprise Monitoring Agent configuration for the request monitoring enabled and sampling rate value. By default, the sampling rate is 2, which means that only 2% percent of all requests (randomly selected) are measured.

Selected Request or Datasource/JMS Summary workspaces not reporting data

The problem: Selected Request or Datasource/JMS Summary workspaces do not report data.

The solution: By default, request monitoring is configured for fixed interval collection (60 second expiration interval), take this into account when navigating through workspaces. If you still have this problem, check the request monitoring level in your Tivoli Enterprise Monitoring Agent configuration. By default it is set to Level 1, which means that only edge request times are measured, increase this to Level 2 in order to display data in these workspaces.

Statistics in the Selected Application Summary not always present

The problem: In the "Selected Application Trend at L1" view and "Selected Application Trend at L2/L3" view, the statistics in the Selected Application Summary table are not always present.

The solution: It works as designed. The statistics are only displayed in this table when you are accessing the application.

Parent topic:

Troubleshooting: Tivoli Enterprise Portal

---

+

Search Tips   |   Advanced Search