Troubleshooting the sensor
This topic describes common problems that occur with the Oracle sensor and presents solutions for those problems.
Oracle sensor does not start
- Problem
- The Oracle signature is not matching because either you have renamed the Oracle binaries or you are running a version of the Oracle server that Agile Service Manager does not support (Express Edition, for instance).
- Solution
- Do not change the names of the binaries and ensure you are using a supported version of Oracle. Also make sure that the TNSListener service is started for the Oracle database.
Sensor fails with Unable to find the servers
error
- Problem
- The Oracle database account is not functioning because of one of the following reasons:
- The password is not correct.
- The account is locked.
- The account does not have connect privileges.
- Solution
- Update the access list, unlock the account, or add the connect privilege.
Oracle RAC is not discovered
- Problem
- Oracle RAC is not discovered.
- Solution
- Verify that the discovery user has sufficient privileges to execute RAC commands.
Oracle Sensor executes commands like “srvctl”, “crs_stat”, “cemutlo” and in case any of these command fails, then it may lead to failure in discovery of RAC or Database Instances. This may also result in duplicate Oracle instances.
- Problem
- Oracle sensor finished with a warning message as shown below:
"CTJTD1048W Oracle RAC environment has been detected. The current Oracle SID IP does not represent the SCAN IP. The discovery will be completed using a sensor through one of SCAN IPs".
- Solution
- For Oracle 11.2g onwards, sensor will continue the discovery only in case of SCAN IP. For all other IPs, Oracle sensor will complete with a warning message: CTJTD1048W. This is expected behaviour and can be ignored in case of RAC setup.
- Problem
- Oracle sensor finished with below warning message as shown below:
"CTJTD1049W Oracle RAC environment has been detected. The current Oracle SID IP does not represent the first SCAN IP. The discovery will be completed using a sensor through this SCAN IPs".
- Solution
- For Oracle 11.2g onwards, when a RAC node has more than one SCAN IP configured on it, then the discovery will always be completed using the first SCAN IP (for this node), as determined from the “srvctl config scan” output. For all other SCAN IP, sensor will complete with warning message: CTJTD1049W.
Oracle duplicates occur when instances are discovered by both Veritas cluster sensor and Oracle sensor
- Problem
- When you use both Veritas cluster sensor and Oracle sensor to discover an Oracle instance, duplicates might occur. This happens because Veritas cluster sensor uses the upper case for the instance SID and Oracle sensor uses the lower case for the same SID.
- Solution
- To avoid this problem, modify the
dist/etc/discover-sensors/VeritasClusterSensor.xml file by changing the
following line:
into the following line:<source>Sid</source>
After the change, Veritas cluster sensor creates Oracle instances with the lower case SID.<source>%{Sid}</source>
Note: If you change the line after running discoveries where no duplicates occurred, new duplicates might occur.
During Oracle RAC Discovery, all OracleInstance objects are not linked to the OracleRAC object
- Problem
- During Oracle RAC Discovery, Oracle instances are discovered but all OracleInstances objects are not linked to the OracleRAC object or linked OracleInstance objects gets deleted after an agent OracleDependencyAgent runs.
- Solution
- To avoid this problem, set the following two property to ‘
true
’ incollation.properties
file on the discovery server. By default, it is 'false
'. For more details, refer https://www.ibm.com/support/pages/apar/IJ26967-
com.collation.oracle.sensor.rac.UseHostnameOfCrsStatForMerge=true
-
com.collation.oracle.sensor.rac.AccessOtherScanNodes=true
Setting the above properties will result in improved linking between OracleRAC object and OracleInstance objects.
-