Other installation problems

The following miscellaneous problems might occur:

An installation fails on a UNC mapped drive

You are running an installation with the installation images on a drive mapped using the Universal Naming Convention (UNC). The wizard fails at the first step.

Cause and solution

The Tivoli Workload Scheduler installation wizard methodology does not support UNC mapped drives. Rerun the installation from a drive that is not UNC mapped.

Message "Error writing file = " received

When performing any type of installation on any operating system, you might receive the following error:

Error writing file = There may not be enough temporary disk space. 
Try using -is:tempdir to use a temporary directory on a partition
with more disk space.

Note in particular the absence of an error code, which differentiates this message from a very similar message, with error code 28, that indicates that you are not logged on as root (see Message "Error writing file = 28" received).

Cause and solution

Normally this error means what it says; the solution is as follows.

First, try to redirect the installation to use a different temporary directory, by adding the -is:tempdir .<temp_dir_path> variable to the installation command.

If this oes not work, you must use one of these two methods to give more space to the swdis directory:

Either:
Create a new version in a different file system. The procedure is as follows:
1. Delete or rename both the work and the backup subdirectories and recreate the directories in a file system with more space in it.
2. Link the new directories to the .swdis directory using the ln -s command.
Or:
Create a new backup directory in a file system with more space in it, and modify the /etc/Tivoli/swdis.ini file to point to it.

Ensure to modify the correct section of the swdis.ini file, as follows:
- If you are making a local silent InstallShield wizard installation that uses the disconnected command line (wdinstsp), modify the value of the backup_dir key in the [#MOBILE] section.
- If you are making a remote installation using Tivoli Configuration Manager, you must identify the section relative to the endpoint chosen as the target (for example, [lab133080_aix]), and modify the backup_dir key in that section.

Message "Error writing file = 28" received

When performing any type of installation on any operating system, you might receive the following error:

Error writing file = 28 There may not be enough temporary disk space. 
Try using -is:tempdir to use a temporary directory on a partition
with more disk space.

Note in particular the error code 28, which differentiates this message from a very similar message, without error code 28, that does indicate disk space problems (see Message "Error writing file = " received).

Cause and solution

This error does not mean exactly what it says. When performing a silent installation of a fix pack on UNIX, and possibly in certain other circumstances, this error message might mean that you are not logged on as root.

Make sure that you are logged onto the workstation as root before running the silent installation:

/SETUP.bin -options <path_to_patchInstall.txt> -silent

Message AWSFAB037E is received on UNIX

The twsinst script installation fails on UNIX with the following error message:

AWSFAB037E The twsinst script is being run from the wrong directory. 
AWSFAB038I Mount the TWS installation CD and run the twsinst utility 
placed there.

Cause and solution

These error messages are received when attempting to install Tivoli Workload Scheduler on a UNIX operating system using the twsinst utility copied from the installation DVD to the home directory of the user that is nominated as the TWS_user during the installation. The installation fails and no log files are generated.

You can run twsinst from the following places:

The Tivoli Workload Scheduler DVD
A disk image of the DVD
A copy of the twsinst utility and its associated files placed in any local directory other than the home directory of the user that is going to be nominated as the TWS_user during the installation.

| |

A master domain manager installation fails with message AWSJIS038E

You are installing a domain manger and the installation fails with |the following error message:

AWSJIS038E An internal error has occurred. 
|An unspecified internal error has occurred during the installation process.

moreover, |the output of the installation step list if you are performing an |interactive installation, or the \inst_logfile_dir\ItemConfigure_TWS_DBOut.xml |files if you are performing a silent installation, reports the following |information:

Database DDMSM already present
|Checking if DDMSM is a TWS database
|DDMSM is not a TWS database.

Where inst_logfile_dir is |the directory where all the installation log files are saved.

Cause and solution

You are installing a master domain manager and the installation |fails because you specified in the Database name field |the name of a database already used by a dynamic domain manager. Run |the installation again providing a different name for the database.

| |

A dynamic domain manager installation fails with message AWSJIS038E

You are installing a dynamic domain manager and the installation fails with the |following error message:

AWSJIS038E An internal error has occurred.
|An unspecified internal error has occurred during the installation process.

moreover, |the output of the installation step list if you are performing an |interactive installation, or the \inst_logfile_dir\ItemConfigure_TWS_DBOut.xml file |if you are performing a silent installation, reports the following |information:

Database TWSBM already present
|Checking if TWSBM is not a TWS database
|TWSBM is a TWS database. It must be a TDWB database.
|One or more errors has occurred during the database setup.

Where inst_logfile_dir is the directory where all the |installation log files are saved.

Cause and solution

You are installing a dynamic domain manager, the installation fails because you |specified in the Database name field the name |of a database already used by a master domain manager. Run the installation |again providing a different name for the database.

| |

A dynamic domain manager or backup dynamic domain manager installation fails after the database |configuration

You are installing a dynamic domain manager or a backup dynamic domain manager and the installation |fails after the database configuration step.

Solution

In this case you have to uninstall the dynamic domain manager or backup dynamic domain manager manually |performing the following steps:

From the TWS_home_directory/TDWB/bin directory, |run the following command: |
```
exportserverdata -dbUsr <username> -dbPwd <password> [exportFile <export_file>]
```
where: |
|
username
|
Specify the name of the user required to access the Tivoli Workload Scheduler database |server. |
|
password
|
Specify the password of the user required to access the Tivoli Workload Scheduler database |server |
|
export_file
|
Specify the name of the text file that is created. The default |is server.properties. |
|

Open the file you created and edit it by removing the entry related |to the failed installation. For example: |
```
http://9.168.117.211:31115/JobManagerRESTWeb/JobScheduler
```

Update the database definition by running the following command: |
```
importserverdata -dbUsr <username> -dbPwd <password> [importFile <import_file>]
```
where <import_file> is the name of the file you created |using the exportserverdata command. The |default is server.properties.

An installation fails with a problem with the installation images on an NFS mount

The installation images are on an NFS mount. The installation fails and the log shows messages similar to the following:

cannot start <file_name>
No such file or directory <file_name>"

where <file_name> is a file in the directory structure of the installation images.

Cause and solution

The NFS mount is corrupt. Refresh the NFS mount by issuing unmount and then mount commands. Retry the failed step (or the entire installation, depending on what went wrong and at what point in the installation).

An InstallShield wizard installation fails on AIX or HP-UX with a "run error"

You are trying to install Tivoli Workload Scheduler using the InstallShield wizard on an AIX® or HP-UX operating system. The installation fails giving the following exception in thread "main":

java.lang.NoClassDefFoundError: run error

Cause and solution

This caused by a combination of display, Java, and binary issues.

To resolve this problem, perform the following steps:

Ensure Quality Pack 10 or higher is installed.
Run xhost + and re-export the display.
Retry the installation using the SETUP.bin binary located at the root of the DVD.
This copies the appropriate media to /tmp and runs the installation from there.

Note:

If you have insufficient space in /tmp, run the SETUP.bin command with the -is:tempdir .<temp_dir_path> variable.

| |

An InstallShield wizard installation does not start and displays the |message "A supported JVM is missing"

You are installing on UNIX or Linux and the interactive or silent |installation does not start. The message "A supported Java Virtual Machine is missing" |displays.

Cause and solution

On very slow environments you could experience this installation |problem and error message. However, this InstallShield message is |incorrect. The problem is that the InstallShield Java Virtual Machine verification routine times out |due to the slow environment on which you are running.

InstallShield has the following parameter in the installation launcher:

-is:jvmtime timeout in seconds

where timeout in seconds is the amount of time |InstallShield waits before timing out the installation. If you have |a slow environment, set this parameter to a high value to enable the Java Virtual Machine recognition |routine enough time to obtain the value before timeout. Initially, |set the timeout in seconds parameter |to 60, and increment it by 60 until the installation successfully |launches.

An InstallShield wizard "Add feature" installation fails

You are trying to add a feature to an existing Tivoli Workload Scheduler installation, using the Add feature option in the wizard. You have started the wizard from the Tivoli Workload Scheduler DVD, but have previously copied the DVD image of the feature to your hard disk. When the wizard asks you to supply the path to the feature installation, you eject the product DVD before supplying the hard disk path information. The installation fails.

Cause and solution

This is a known problem with the InstallShield wizard. If an installation starts from a DVD, the InstallShield wizard expects to find a DVD in the DVD drive.

To correct this problem, put the product DVD, or any other DVD, back in the DVD drive (the InstallShield wizard requires a DVD, but it can be any DVD).

A software package block installation fails with the message: DISSE0324E

You launched an installation of either the full product, a component, or a fix-pack, that uses the software package blocks of the software distribution component of IBM® Tivoli Configuration Manager. The installation fails, giving the following messages:

DISSE0282E Error compressing file <file_name> in the software Package block.
DISSE0324E Cannot create backup package 
DISSE0005E Operation unsuccessful

Cause and solution

The installation using a software package block is unable to check that there is sufficient space for the backup it performs. The backup requires at least 80 MB of disk space. The directory used for the backup is determined by the parameter backup_dir in the [#MOBILE] section of the swdis.ini file. The following is an example of this section of the file:

[#MOBILE] 
product_dir=/root/.swdis 
working_dir=/root/.swdis/work 
backup_dir=/root/.swdis/backup 
trace_level=0 
trace_size=1000000 
send_timeout=300 
autopack_dir=/root/.swdis/autopack 
staging_dir=root/.swdis/service 
user_file_variables=/root/.swdis/swdis.var 
import_libraries=spd,libecimp

If there is insufficient space in this directory, the error messages shown above are displayed.

Because the Tivoli Configuration Manager installation was unable to start, the restore script twsRestore cannot be used. The recovery procedure is as follows:

If you are performing a full-product installation, manually delete the following files and directories:
- twsRestore.sh or twsRestore.cmd (as appropriate)
- twsRemove.sh or twsRemove.cmd (as appropriate)
- _uninstall directory
Resolve the file space problem, as follows.
First, try to redirect the installation to use a different temporary directory, by adding the -is:tempdir .<temp_dir_path> variable to the installation command.

If this oes not work, you must use one of these two methods to give more space to the swdis directory:
- Either:
  Create a new version in a different file system. The procedure is as follows:
  1. Delete or rename both the work and the backup subdirectories and recreate the directories in a file system with more space in it.
  2. Link the new directories to the .swdis directory using the ln -s command.
- Or:
  Create a new backup directory in a file system with more space in it, and modify the /etc/Tivoli/swdis.ini file to point to it.
  
  Ensure to modify the correct section of the swdis.ini file, as follows:
  - If you are making a local silent InstallShield wizard installation that uses the disconnected command line (wdinstsp), modify the value of the backup_dir key in the [#MOBILE] section.
  - If you are making a remote installation using Tivoli Configuration Manager, you must identify the section relative to the endpoint chosen as the target (for example, [lab133080_aix]), and modify the backup_dir key in that section.
Run the installation again.

A software package block installation fails to complete successfully

You launched an installation that uses the software package blocks of the software distribution component of IBM Tivoli Configuration Manager. The installation fails.

Cause and solution

Problems when installing remotely with a software package block can often be difficult to solve, because it might be more difficult to set the remote environment correctly so that the installation runs successfully. For this reason, the software package block is supplied with a series of keys that switch on or off its installation activities. These action keys are set by default to true, so that the installation completes normally. If it fails for a reason that you know you can resolve afterwards, you can retry it, setting one or more of these keys to false, so that the installation process does not attempt to perform that or those steps.

For example, if the installation fails when trying to back up the previous installation, and you know that you can proceed without making a backup, you can eliminate this action from the installation by setting its action key to false. You re-launch the installation, which does not perform the backup step, but otherwise completes successfully.

A description of the processing carried out in each step is given in the Tivoli® Workload Scheduler: Planning and Installation Guide.

Note:

The installation steps are always the same, whatever installation method you use.

The details of the action keys are as follows (set any of them to false to not perform that action):

execActionTools = "true": This controls all of the other keys. If you set it to false, none of the other actions take place (their settings are ignored).
execTwsStopAction = "true": This controls whether or not the installation stops existing Tivoli Workload Scheduler processes on the target workstation.
execTwsCleanAction = "true": This controls whether or not the installation cleans up an existing installation before upgrading it.
execTwsUndoAction = "true": This controls whether or not the undoable installation script is run.
execTwsBackupAction = "$(backup)": This controls whether or not the installation takes a backup of an existing installation before commencing the installation. By default this value is set to "false".
execTwsUserAction = "true": This controls whether or not the installation creates or modifies the <TWS_user> details.
execTwsConfigAction = "true": This controls whether or not the installation configures Tivoli Workload Scheduler after the installation.
execTwsStartUpAction = "true": This controls whether or not the installation starts up Tivoli Workload Scheduler after the installation.
execTwsCommitAction = "true": This controls whether or not the installation issues a software distribution commit action to complete the installation (see the Tivoli Configuration Manager publications for more details).

When you have resolved the problem, reset these keys to true for any other installation using the software package block.

See the Tivoli Configuration Manager publications for details of how to change the values of a parameter in a software package block.

The installation fails with the error AWSFAB035E

You are trying to install Tivoli Workload Scheduler or one of its fix packs. The installation fails with the following error:

AWSFAB035E The installation has failed. For more details see 
the log file: /tmp/tws84/summary.log.

The installation log contains the Software Distribution error message DISSE0006E:

DIS:SENG:0006 Operation unsuccessful: Fatal failure.        
Explanation: The operation cannot be completed because of   
an internal error (for example, a memory allocation failure)
System Action: Operation failed.

The log also contains details of the last internal command it ran:

+ wdinstsp -f -D promote=false -D upgrade=false -D fresh_install=true

Cause and solution

This is a problem related to incompatibility between the Tivoli Workload Scheduler installer and the current version of IBM Tivoli Configuration Manager. Follow these steps:

Check the version of Tivoli Configuration Manager that is installed, using the command wlsinst -ah. Check if any patches are installed.
If you are using Tivoli Configuration Manager version 4.2, with fix pack 4.2-SWDGW-F1P1 (or later for the same component), there is a compatibility problem because the installation of Tivoli Workload Scheduler version 8.4 and its fix packs are only compatible with the GA version of the Tivoli Configuration Manager SWDGW component.
In this case you must either choose a different installation method that does not use Tivoli Configuration Manager or uninstall the Tivoli Configuration Manager 4.2-SWDGW fix pack or packs until the installation is complete.
The Tivoli Workload Scheduler GA and fix pack installation uses the Tivoli Configuration Manager version 4.2 disconnected command line. A problem has been discovered with this component which requires you to do the following:
1. Install Tivoli Configuration Manager fix pack 4.2-TCM-FP02 on the workstation.
2. Run the wconvcat command (described in the 4.2-TCM-FP02.README file) to restore the functionality of the disconnected command line.
3. Retry the failed installation.

| |

An agent installation fails if you launch it from a shell where |you set the environment variables using the swd_env script

The following error occurs when you install an agent:

AWSFAB035E The installation has failed.

Cause and solution

You can receive this error if you are running the installation |form a shell script where you previously set up the Software Distribution |environment variables using the swd_env.cmd or |the . . /swd_env.sh.

Resume the installation from a shell script where you did not set |the environment variables and complete the installation.

The installation fails with the error AWSGAB566E

You are trying to install Tivoli Workload Scheduler, and the installation fails with the following error:

AWSGAB566E There is not enough disk space available in the following 
supplied directory to complete the installation: <directory_name>. 
The installation requires <required_space> megabytes, 
but only <available_space> megabytes are available.

Cause and solution

The probable cause of this error is that either the file set where the product is to be installed, or the file set where the temporary installation files are being written, is not large enough. Information about the disk space requirements is given in the Tivoli Workload Scheduler System Requirements Document at http://www.ibm.com/support/docview.wss?rs=672&uid=swg27019747

However, this error is also given if the virtual memory file (sometimes called page file or swap space) on your hard disk is not large enough.

Thus, if there seems to be sufficient disk space, check also the virtual memory that you have allocated to the hard disks. The installation requires at least 256 MB of virtual memory on any operating system.

The commit step fails

You are trying to install Tivoli Workload Scheduler, and the commit step fails

Cause and solution

If the mapped attribute, LDAPUSERIdMap is different from the login attribute, LDAPUserFilter, the commit step fails. You must dump the security file and insert the mapped value (see error message). Then, manually perform the failed composer command (see error message). Finally, set the failed step as successful and resume the installation.

| |

Uninstalling and reinstalling a dynamic agent with the same name |on the same workstation

If you have to uninstall a dynamic agent from a workstation and want |to reinstall it with same name on the same workstation, perform the |following steps to override the existing agent definition:

Obtain the ID of the dynamic agent running the following command: conman |cpuinfo dynamic agent_name on the workstation |in the plan.

Add the ID of the dynamic agent in the jobmanager.ini file using |the AgentID keyword, as shown in the following |example: |

[ResourceAdvisorAgent]
|FullyQualifiedHostname = geneva12.paralab.it.fsx.com
|ResourceAdvisorUrl = https://localhost.localdomain:31116/JobManagerRESTWeb
|/JobScheduler/resourceCPUScannerPeriodSeconds = 10
|ScannerPeriodSeconds = 120
|NotifyToResourceAdvisorPeriodSeconds = 180
|ComputerSystemDisplayName = geneva12_1
|BackupResourceAdvisorUrls = "https://localhost.localdomain:31116/JobManager
|RESTWeb/JobScheduler/resource",
|"https://geneva12.paralab.it.fsx.com:31116/JobManagerRESTWeb
|/JobScheduler/resource"
|AgentID = 4A0B274AC61E11DF995053784667819B

Miscellaneous failures

The installation fails and the cause is not immediately obvious from the log messages.

Cause and solution

The cause of the failure could be any of the following:

The FTP transfer of the files to the node was not done in binary mode

You copied the install directory from the DVD to the local hard disk using FTP, but did not specify the binary option. Make sure the entire directory is transferred by FTP in binary mode.

Note:

The directory on the local hard disk can have any name, but it is important to have a parent directory available for the twsinst installation, because some temporary files need to be located there.

For example:

/temp/HP-UX

/temp/TWS84/HP-UX

There is not enough disk space available for the installation

Check that there is enough disk space for the installation on your chosen fileset.

See the Tivoli Workload Scheduler System Requirements Document at http://www.ibm.com/support/docview.wss?rs=672&uid=swg27019747 for more information about the amount of space necessary for installation.

File names did not retain their original case

On UNIX, check that file names retain their case. For example, the file "TWS_size.txt" cannot be "tws_size.txt".

One or more required files were not copied from the root of the installation DVD

Check that the number of files copied from the DVD is the same as that on the DVD. If not, copy the files again.

You launched a second installation before the first one had successfully finished.

If you launch an installation (of an additional component, for example) while another installation is still running, both installations might fail, or the second installation might try and resume the first, as if you had terminated the first installation part way through and now want to continue.

Depending on the stage that the first installation reached, you might just be able to close the second installation and let the first one finish. However, if one or both have failed, you might need to uninstall and then start the installation again.