Review the Tivoli Enterprise Portal Servertroubleshooting topics for help with solving problems you encounter with the portal server.
If you do not find the resolution to your problem, review the topics in Chapter 6, “Connectivity troubleshooting,” on page 123.
Performance impacts of the HTTP and HTTPS protocols
Connection protocol options between the Tivoli Enterprise Portal client and the Tivoli Enterprise Portal Server include the default protocol, IIOP, as well as the HTTP and HTTPS protocols. Note, however, that you might encounter a response time impact when you use the HTTP and HTTPS protocols. Affected workspaces include the Linux Process workspace, Linux PAS workspace, and UNIX Process workspace. These workspace have longer response times when you use the HTTP Tivoli Enterprise Portal client instead of the CORBA/IIOP Tivoli Enterprise Portal client. In addition to the response time impact, you might encounter increased CPU consumption for the HTTP and HTTPS protocols.
Users who run the IBM HTTP Server do not have permission to the
content directory
During configuration and startup of the Tivoli Enterprise Portal Server, the system attempts to confirm that the user running the IBM HTTP Server has permission to access the IBM HTTP Server content directory. If not, one of the following
messages are displayed:
KCIIN2723W User who runs IHS (IBM HTTP Server) does not have proper permissions to IHS content directory. Do you want to continue?
or
KCIIN2724W User who runs IHS (IBM HTTP Server) does not have proper permissions to IHS content directory.
This error occurs when the IBM HTTP Server is running under a user that does not match the user specified in theCANDLE_HOME/ARCH/iu/ihs/httpd.conf file. For the portal server to run correctly, the user who is running the IBM HTTP Server must have access to the IBM HTTP Server content directory.
tacmd exportWorkspaces or importWorkspaces receives an out of
memory error
If you get an OutOfMemory Error when running this command, you can increase the maximum Java heap size for the tacmd java JVM by using the
TACMD_JVM_MAX_MEMORY environment variable. This variable specifies the maximum java heap size (in megabytes) for the tacmd Java virtual system. Memory for tacmd is freed when the tacmd invocation finishes. Valid values are 256 through to 2048, inclusive.
For IBM Tivoli Monitoring v6.2.2 Fix Pack 2 or higher, you can set it in the command environment , using SET/export, or you can set it in the environment files (KUIENV on Windows systems, or the $CANDLEHOME/bin/tacmd shell script on UNIX and Linux systems).
The portal server and Warehouse Proxy Agent fail to connect to the
database on a 64-bit Windows system
The ODBC control panel available in the Windows Start menu is for 64-bit ODBC configuration. Since the portal server and the Warehouse Proxy Agent are 32-bit applications the 32-bit ODBC control panel must be used. Launch the 32-bit ODBC window from C:\Windows\SysWOW64\odbccp32.cpl and manually create the "ITM Warehouse"/"TEPS" ODBC source by going to the System DSN tab and clickingAdd.
Failed to log on as sysadmin with portal server LDAP enabled
When LDAP authentication is enabled for the Tivoli Enterprise Monitoring Server, thesysadminID must be defined in the LDAP server. However, when LDAP is enabled for the Tivoli Enterprise Portal Server, the sysadmin ID should exist in the monitoring server's local OS user registry, but should not be defined in the LDAP server, otherwise, the sysadmin ID will not be able to log in.
On AIX systems, newly created users with auto-expire passwords
cause installation failures
When installing on AIX systems, security policies for newly created users
auto-expire the password after the first use and require you to set a new (or same) password as a permanent password. The Tivoli Enterprise Portal Server
configuration interface allows you to create a new user ID for the portal server and warehouse database, but using the interface always fails because the user
password is not set and is expired. You must ssh/telnet into the same server, using the target user ID, and set the password appropriately.
Linux portal server unable to FTP catalog/attribute files
Linux portal server unable to FTP catalog/attribute files with Manage Tivoli Enterprise Monitoring Server. A monitoring server is needed along with the portal server (on the same system) to be able to seed a z/OS monitoring server.
Upgrading the Tivoli Enterprise Portal Server takes a long time
Performing a Tivoli Enterprise Portal Server upgrade, depending on the efficiency of the hardware platform, can take anywhere from 30 minutes to over an hour.
Running the Tivoli Management Services Discovery Library Adapter,
results in a book that does not contain the fully qualified host name
Edit the resulting xml file and change the shortname to the fully qualified host name.
Portal server performance is slow
If you want to increase the performance of your portal server and you are not concerned about security, you can disable Secure Socket Layer data encryption on the portal server.
About this task
If you do not want to use Secure Socket Layer communication between Tivoli Monitoring components and the Tivoli Enterprise Portal Server, use the following steps to disable it:
1. In Manage Tivoli Enterprise Monitoring Services, right-click Tivoli Enterprise Portal Server.
2. ClickAdvanced>Edit ENV file.
3. Find the following line:
kfw_interface_cnps_ssl=Y 4. Change theYtoN.
5. Save the file and exit.
6. ClickYeswhen you are asked if you want to recycle the service.
Cannot create a Tivoli Enterprise Portal Server database
About this task
When using DB2 8.1 or 8.2, you must install the correct fix pack versions of DB2 in order to create a Tivoli Enterprise Portal Server database. These fix pack versions are:
v DB2 V8.1 with Fix Pack 10 or higher fix packs v DB2 V8.2 with Fix Pack 3 or higher fix packs
Also, on AIX systems, a failure occurs when you attempt to install a Tivoli Enterprise Portal Server with a DB2 database. Using the db2 installation user ID (default is db2inst1), do the following:
1. Stop the DB2 server if not already stopped using the following command:
cd /db2inst1/sqllib/adm db2stop
2. Issue the following configuration changes:
export EXTSHM=ON
db2set DB2ENVLIST=EXTSHM db2set -all
3. Using your preferred editor add the following lines to the/db2inst1/sqllib/ db2profilefile:
EXTSHM=ON export EXTSHM
4. Restart the DB2 server using the following command:
cd /db2inst1/sqllib/adm db2start
5. Restart the Tivoli Enterprise Portal Server using the following command:
cd /opt/IBM/ITM/bin ./itmcmd agent start cq
For information on how to modify kernel parameters, see “Modifying kernel parameters (Linux)” (http://publib.boulder.ibm.com/infocenter/db2luw/v9/
topic/com.ibm.db2.udb.uprun.doc/doc/t0008238.htm) in the IBM DB2 Database for Linux, UNIX, and Windows Information Center.
You receive a KFW error when a query is sent to more than 200
managed systems
You receive the following error when a query is sent to more than 200 managed systems:
KFWITM217E Request error: Request to xxx nodes exceeds the limit of 200. Please specify a smaller distribution or increase the maximum.
There is a default limit of 200 nodes for any single query for a workspace view. If the following conditions exist in the query for a workspace view, you must increase the KFW_REPORT_NODE_LIMITenvironment variable for the Tivoli Enterprise Portal server environment variable as described below:
v The query is assigned to a managed system list that contains more than 200
managed systems. OR
v More than 200 managed systems are explicitly assigned to a query in any
workspace view.
Under these conditions, you must increase the following Tivoli Enterprise Portal server environment variable.
KFW_REPORT_NODE_LIMIT=xxx
where xxxis an integer that is equal to or greater than one of the following:
v The number of managed systems defined in a managed system list.
OR
v Explicitly assigned to a query over 200 in a Tivoli Enterprise Portal workspace
view.
You must add theKFW_REPORT_NODE_LIMITenvironment variable or remove the comment marker (#) in the following Tivoli Enterprise Portal Server environment files, and restart the portal server.
v Windows systems:\ibm\itm\cnps\kfwenv v Linux or AIX systems:/opt/IBM/config/cq.ini
After you change theKFW_REPORT_NODE_LIMITvariable, you might receive the following error:
KFWITM217E Request error: SQL1_CreateAccessPlan failed, rc=1.
Typically this problem is caused when too many explicitly defined managed systems are assigned to a query for a workspace view. The best practice for resolving this problem is as follows:
1. Create a managed system list that specifies the explicitly defined managed systems.
2. Remove the explicit assignments from the query.
3. Assign the managed system list to the query.
Alternatively, you can reduce the number of managed systems that you explicitly define in the query.
Non-hub situations are not associated at the Tivoli Enterprise Portal
Server level
Only pure hub situations should be associated to the Tivoli Enterprise Portal Server. However, if you want non-hub situations to be associated at the Tivoli Enterprise Portal Server level, set the Tivoli Enterprise Portal Server environment variable: KFW_CMW_SPECIAL_HUB_ENTERPRISE=N.
When non-hub situations are associated at the Tivoli Enterprise Portal Server, they turn TRUE, meaning they are visible in the situation event console. Through the Situation editor, if you assign all the agents and managed systems lists from a situation, that situation event continues to appear in the situation event console.
Non-root stopping or starting agents causes problems
You might experience issues while starting or stopping agents on servers, when using a non-root user ID. The following message might be received:
KCIIN1191E Cannot execute product_code stop script.
To avoid this situation, use the root account or an account with granted required permissions (itmuser group).
Root password is not accepted during non-root Tivoli Enterprise Portal
Server configuration
While configuring the Tivoli Enterprise Portal Server, when using non-root user, the provided root password is not validated correctly. You should use root account or an account with granted required permissions (itmuser group).
Corba user exception is included in the portal server log when
creating situations
When a user creates a new situation, the situation name must be unique. To verify that the new name is unique, the software attempts to access a situation by the new name. If the situation is found, then the name is already used and the user must select a new name. If the request fails, then the name is not already used. The failure to find the situation name is reflected in the log as the CORBA exception. The CORBA user exception indicates that the name is unique.
Stopping or starting the eWAS subcomponent of the portal server
The eWAS subcomponent of the Tivoli Enterprise Portal Server , named Tivoli Enterprise Portal Server extension server (TEPS/e) is installed automatically with the portal server. If you need to start or stop the application server instance of eWAS on which the portal server is running, you must do it by starting or stopping the portal server. You cannot use the eWAS start and stop commands to control eWAS. Using the eWAS start and stop commands results in an internal error, indicated by KFWITM392E Internal error occurred during login. If you have already used the eWAS commands, see "Starting and stopping eWAS" in the
Setting the TEPS/e maximum heap size
When the dashboard data provider has been enabled on the Tivoli Enterprise Portal Server, you can use monitoring dashboards in the Dashboard Application Services Hub console. In larger environments, it is possible for the Tivoli Enterprise Portal Server extended services (TEPS/e) Java process to run out of memory when large amounts of data are passed to the Dashboard Application Services Hub. You can increase the maximum heap size of the TEPS/e JRE to improve server response time and better accommodate a larger monitored environment.
Before you begin
The memory required by the TEPS/e depends on the size of the monitoring environment and on the Java heap size parameters. The default maximum Java heap size is 1500 MB, which is appropriate for 32-bit systems. On 64-bit systems, if you encounter “out of memory” exceptions, you can increase the maximum heap size of the TEPS/e JRE to values greater than 1500 MB to allow more memory usage and accommodate a larger monitored environment.
About this task
Complete the following steps on the computer where the portal server is installed to adjust the TEPS/e maximum heap size.
Procedure
1. At the command prompt, change to the bin directory of your Tivoli Monitoring installation:
v install_dir\cnpsj\profiles\ITMProfile\bin
v install_dir/platform/iw/profiles/ITMProfile/bin 2. Enter the following command to start the eWAS console:
v wsadmin.bat
v wsadmin.sh
3. At the wsadmin prompt, enter the following command: $AdminConfig modify [$AdminConfig list JavaVirtualMachine] "{maximumHeapSize 2000}"
4. Enter the following command to save your changes:$AdminConfig save
5. Enter the following command to exit:quit
What to do next
Restart the portal server. If you continue to encounter “out of memory” exceptions, try increasing the maximum heap size again.