|Oracle® Configuration Manager Installation and Administration Guide
Part Number E37289-03
|PDF · Mobi · ePub|
This chapter includes the following sections:
To enable debug logging by the collector, add the following property to the
$ORACLE_CONFIG_HOME variable was defined for the deployment of the collector, the location of the
collector.properties file is the
Otherwise, add the property to the
Note that debug logging of the collector consumes space. Once you have completed the debugging task, remove the added entry from the
collector.properties file to resume the previous behavior.
To have Perl scripts generate debugging output when the collection is run, set EMAGENT_PERL_TRACE_LEVEL=1 and run emSnapshotEnv.
To see all the steps performed throughout the execution of the setupCCR, configCCR, and deriveCCR commands, set the CCR_DEBUG environment variable to 1. We suggest that the output be redirected to a file. This enables you to send the file to Oracle Support if needed.
This section describes some of the errors you may encounter while you work with Oracle Configuration Manager and provides possible solutions to these errors.
When you run the
installCCRSQL.sh script, it creates the
ORACLE_OCM user and sets up a job to collect database configuration information. The
ORACLE_OCM user requires
EXECUTE privileges on
DBMS_SCHEDULER for database versions 10g or higher, and on the
DBMS_JOB for pre-10g databases. If these privileges are granted to
ORACLE_OCM user inherits these privileges, otherwise these privileges are explicitly granted when the
installCCRSQL.sh script is executed. If the inherited privileges are revoked, the following errors indicating the lack of privileges will be logged in the
ORA-12012: error on auto execute of job 52 ORA-04068: existing state of packages has been discarded ORA-04063: package body "ORACLE_OCM.<package name>" has errors ORA-06508: PL/SQL: could not find program unit being called
To resolve these errors, you must grant the missing
EXECUTE privilege to the
For database versions 10g and higher, grant
EXECUTE privileges on the
DBMS_SCHEDULER packages to the
ORACLE_OCM user by entering the following
SQL> grant execute on UTL_FILE to oracle_ocm; SQL> grant execute on DBMS_SCHEDULER to oracle_ocm; SQL> ALTER PACKAGE oracle_ocm.MGMT_DB_LL_METRICS compile; SQL> ALTER PACKAGE oracle_ocm.mgmt_config compile;
For pre-10g databases, grant
EXECUTE privileges on the
DBMS_JOB package to the
ORACLE_OCM user by entering the following
SQL> grant execute on UTL_FILE to oracle_ocm; SQL> grant execute on DBMS_JOB to oracle_ocm; SQL> ALTER PACKAGE oracle_ocm.MGMT_DB_LL_METRICS compile; SQL> ALTER PACKAGE oracle_ocm.mgmt_config compile;
There may be cases when the
ORACLE_OCM user needs to be granted the required privileges during installation. While granting the privileges, the following error may occur in the
ORA-04021: timeout occurred while waiting to lock object SYS.<package like UTL_FILE
This error may occur if another procedure is using the package for which the privileges are being granted. To resolve this error, retry the install when the package is not being used. This error may occur while granting privileges on
This error may occur if the value of the
MAX_ENABLED_ROLES initialization parameter has been exceeded. To resolve this error, you must increase the value of the
MAX_ENABLED_ROLES parameter and restart the database as follows:
ora file where <
sid> is the database system identifier and increase the value of
MAX_ENABLED_ROLES. If a server parameter (spfile) has been used, alter the
MAX_ENABLED_ROLES parameter by using the following SQL*PLus command:
alter system set MAX_ENABLED_ROLES=<value> scope=spfile
Restart the database.
Once the database has been restarted, re-run the
To ensure that hostnames are displayed with their fully qualified names on My Oracle Support, the
/etc/hosts file must contain an entry that includes both the hostname and the domain in the following format:
<IP-Address> <Full-HostName> <Short-HostName>
10.10.10.10 myhost.mydomain myhost
/etc/hosts file has not been correctly configured, only the short name is displayed on My Oracle Support.
Oracle Configuration Manager does not allow you to run multiple commands simultaneously. If you attempt to do so, the following messages may be displayed:
Message: Another operation is in progress. Please wait...
Description: There are several Oracle Configuration Manager commands that cannot run concurrently. If you try to run one of these commands while another command is in progress, the second command will not be executed until the first command is completed. A message indicating that another command is in progress is displayed. The second command will automatically be run when the first command is completed.
Action: Initially, take no action, the second command will be executed when the first command is completed. But if the command execution takes too long, a timeout will occur. If a timeout occurs, ensure there is no Oracle Configuration Manager activity by executing
stop command. Delete the
lock file and restart the Scheduler by running the
start command. If you are running the command in Disconnected mode, ensure that no collection or update is taking place and then delete the
Message: Operation blocked, waiting...
Description: You cannot run the
emCCR update_components command if any other
emCCR command is running. If you try to run the command, it will be blocked. You also cannot run any
emCCR command while
update_components is running as all other commands will be blocked.
configCCR and most of the
Action: Initially, take no action, the command will get executed when the current command is completed. If a timeout occurs, ensure that there is no Oracle Configuration Manager activity by executing
stop. Delete the
update* files, and restart Oracle Configuration Manager by running
start. If running the command in Disconnected mode, ensure no collection or update is taking place and delete the
op* and the
Message: The Scheduler is down for upgrade.
Description: While upgrading Oracle Configuration Manager, you cannot run any of the
Action: Retry the commands later.
The following errors in the Oracle Database Trace file denote that the ORACLE_OCM user account does not have Execute Privilege on the UTL_FILE located in the SYS.UTL_FILE package.
ORA-12012: error on auto execute of job xxxxx ORA-04063: package body "ORACLE_OCM.MGMT_DB_LL_METRICS" has errors ORA-06508: PL/SQL: could not find program unit being called: "ORACLE_OCM.MGMT_DB_LL_METRICS" ORA-06512: at line 1
To give Execute Privilege to the UTL_FILE, run the following SQL statement:
GRANT EXECUTE ON SYS.UTL_FILE TO ORACLE_OCM;
At the time of instrumentation of the database, Oracle Configuration Manager executes the ccr/admin/execocm.sql script. This script tests to see if PUBLIC access is granted to all the routines Oracle Configuration Manager needs access to. If PUBLIC access is not available, Oracle Configuration Manager grants specific access rights.If you specifically revoke PUBLIC access after installation, you will need to rerun at least the execocm.sql script. Rerunning the collecconfig.sql script accomplishes the same result.
When there is an Oracle Configuration Manager upgrade failure on the Windows environment, the log file is named update2.log.
Customers are seeing messages similar to the following message when a database that's been previously collected but is now dropped.
WARN: Oracle Configuration Manager database objects are not in sync with the installed configuration collection scripts. Refer to the Installation and Configuration documentation on reloading the SQL collection packages.
emCCR disable_target command. The emCCR status does not look at the list of disabled targets to determine whether it conforms to the list of targets for which it should check metadata.
Customers are seeing the host names containing Siebel services twice: once with the long name and the second with the short name.
The Siebel topology is discovered from the host that runs the Siebel Gateway Server. A host definition is created and uploaded to Oracle services with the intent that the host configuration definition will rendezvous at Oracle when an Oracle Collection Manager collector is installed on the physical host where the Siebel service runs. Due to host network configuration, the name of the host is returned as different strings; one from the gateway server and another from the physical host. It is required that all hosts within the Siebel topology be configured to return full canonic names for hosts. This is true for the Siebel Gateway server, as well as from all hosts containing Siebel services.
Reconfigure your host network to return the full canonic name for the local host. On the Gateway Server, the network configuration must return the full canonic name for all hosts that are known to participate in a Siebel system defined by the Gateway Server.
If you are running Enterprise Manager (EM) and the harvester collection fails with the following error in the sysman/log/emoms.log:
[JobWorker 75210:Thread-61] ERROR gcharvester.GcCollectionMgr initOcm.? - GC OCM Harvester: Caught GC Harvester exception from GCInit.init(): The installed version of Oracle Configuration Manager in the ORACLE_HOME (/scratch/aime/work/midlwre8937/oms11g) is prior to 10.3.1. The Grid Control Configuration harvesting requires at a minimum, 10.3.1
the required directory named external is missing in either the $ORACLE_HOME/ccr/hosts/<local_host_name>/state/upload directory (EM release 10.2.0.5) or the ORACLE_INSTANCE_HOME/ccr/state/upload directory (EM releases 11.1 and later).
To resolve this issue, create the directory $ORACLE_HOME/ccr/hosts/<local_host_name>/state/upload/external (EM release 10.2.0.5) or ORACLE_INSTANCE_HOME/ccr/state/upload/external (EM releases 11.1 and later).
If a database instance is not discovered even though the Oracle Configuration Manager instrumentation step reports no errors, ensure that the JOB_QUEUE_PROCESSES parameter is set to a value greater than 0 (zero). This ensures that the DBMS jobs script runs and in turn the Oracle Configuration Manager instrumentation works.
You can set the JOB_QUEUE_PROCESSES parameter using the ALTER command:
alter system set job_queue_processes=<n> scope=both; where <n> represents the number of JOB_QUEUE_PROCESSES
It is suggested that you set the JOB_QUEUE_PROCESSES parameter to 10.
The installation of Oracle Configuration Manager takes a snapshot of the process environment. These key environmental variables are stored such that scheduled collections behave the same even though they are running as detached processes.
If any of the following variables change, update the snapshot file by running the command:
Key environmental variables are: ORACLE_HOME, ORACLE_CONFIG_HOME, IAS_CONFIG_HOME, ORACLE_INSTANCE, JAVA_HOME, TZ, TNS_ADMIN, CRS_HOME, CLUSTER_NAME, LD_PRELOAD, ORAINST_LOC, BEA_HOME, and WL_HOME.
Installation of the Oracle Configuration Manager fails, with the following error in
$ORACLE_HOME/ccr/hosts/<hostname>/log/install-core-*.log, indicating a relink issue. The log file contains records such as:
skipping incompatible /usr/lib64/gcc/x86_64-suse-linux/4.3/libgcc.a when searching for -lgcc /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../x86_64-suse-linux/bin/ld: cannot find -lgcc collect2: ld returned 1 exit status
The problem occurs on a 64-bit Linux system where the 32-bit libraries are not installed. Install the native Linux x86_bit kit located on the My Oracle Support site (