IBM Tivoli Monitoring > Version 6.3 Fix Pack 2 > Installation Guides > Installation Guide > Deploy monitoring agents across your environment > Bulk agent deployment > Best-practice deployment procedures

IBM Tivoli Monitoring, Version 6.3 Fix Pack 2


Deployment

The following best practices are employed to initiate and perform actual deployments. These best practices apply regardless of whether the deployment is an initial agent installation, an agent upgrade, or an agent configuration update.


It is important to recognize that the groups created for deployment are applicable to all agent maintenance operations throughout the agent lifecycle.

  1. Submit deployment.

    Submitting the deployment differs slightly depending on the type of deployment activity you desire, but they are all very similar.

    • createNode

      Use the tacmd createNode command when performing initial deployments of OS and SSM agents. This includes deploying bundle groups that contain OS agents grouped together with application agents.

        # tacmd createnode -g DBServers -b OS_DBAgentBundles

    • addSystem

      Use the tacmd addSystem command when initially deploying application agents.

        # tacmd addsystem -g AppServers -b AppAgentBundles

    • updateAgent

      Use the tacmd updateAgent command when performing agent or SSM upgrades.

        # tacmd updateagent -g DBServers -b LatestAgentfix packs
        # tacmd updateagent -g AppServers -b LatestAppAgentfix packs

      The KUIWINNT.dsc file on Windows systems and the uiplatform.dsc files (where platform is the platform name) on Linux and UNIX systems have been added or updated so that the KUI package (the tacmd commands) can be remotely deployed. Use the following command:

      where:

      node

      is the IBM Tivoli Monitoring node on which the KUI package is to be remotely deployed.

    • configureSystem

      Use the tacmd configureSystem command when performing configuration updates to agents or SSMs.

        # tacmd configuresystem -g DBServers -b DBServerAgentConfig

    • removeSystem

      Use the tacmd removeSystem command when removing application agents or SSM patches from a system.

        # tacmd removesystem -g AppServers -b AppAgentBundles

  2. Monitor deployment.

    There are many mechanisms that you can use for monitoring deployment status. These include:

    • Use the CLI.

      The tacmd getDeployStatus command returns the complete status of requested deployment transactions.

        # tacmd getdeploystatus -g group_transactionID

    • Use portal workspaces.

      The Tivoli Enterprise Portal provides a number of workspaces that display the status of deployments. These provide an easy mechanism to rapidly visualize deployment status summaries as well as detailed status. See the or the portal online help for more information.

    • Use automation.

      Automation can be driven in two ways. Situations can be created to indicate deployment failures to initiate corrective activities. Also, automation scripts can use the tacmd getDeployStatus command to retrieve and parse status information and initiate corrective activities. Situations and automation scripts can be used in tandem to monitor deployment status.

    • Regardless of the specific mechanism for monitoring deployment status, the best practice involves the following steps.

      1. Check the transaction status to ensure it is queued properly.

      2. Wait some period of time for the deployments to proceed.

      3. Check the transaction status for successful completions and failures. This involves either using the command interface or the portal interface or checking for situation events that indicate deployment failures.

      4. Perform any necessary corrective actions.

      5. Repeat starting with step 2b until all transactions have completed.

  3. Correct problems.

    When a deployment transaction fails, some corrective action is usually required before reinitiating the deployment. Consult the return message portion of the transaction status to identify the problem and possible resolutions. Rerun the deployment once errors are corrected. For example, when using a deploy group with the addSystem or updateAgent commands, remote deployment may fail to locate the existing managed-system name for some hosts. The message received is KDY0012E The target target_host_name is incorrect or is offline. The command did not complete because the value for the target is incorrect or the target is offline. This message normally indicates that the OS agent is not online. If the agent is, in fact, online, cancel current operations to this node:

      # tacmd cleardeploystatus  -h hostname

    Then issue the operation directly using the managed system name parameter instead of the deploy group:

  4. Purge status.

    When all deployment transactions have completed or the deployment status is no longer required, clear the transaction's deployment status from the systems.


Results

While the steps above describe the complete process of deployment, it is common to perform deployments in small sets of consumable transactions. That is, perform an initial deployment on a smaller group of systems to ensure that there are no issues that would inhibit the successful completion of deployment. Once that deployment is complete, continue with one or more larger deployments using target deploy groups with more members. Work through each deploy group representing some portion of the enterprise until you have completed all deployments for the enterprise.


Parent topic:

Best-practice deployment procedures

+

Search Tips   |   Advanced Search