Checklist: Enterprise Manager Cluster Upgrade

This checklist covers the steps that are required to upgrade and deploy an Enterprise Manager cluster. The steps are based on recommended CA practice. See  for configuration options.
apmdevops106
This checklist covers the steps that are required to upgrade and deploy an Enterprise Manager cluster. The steps are based on recommended CA practice. See CA APM Environment Options for configuration options.
You can only use this capability to upgrade from earlier versions of APM to APM 10.7 SP3. You cannot use this capability to upgrade from APM 10.7 to 10.7 SP3.
You can start Cluster Upgrade steps and return to them later. Select
Close
and
Show Wizard
to navigate in and out of the Upgrade Wizard.
Select PDF at the top of the page to download this page in PDF format.
Access Credentials
During the upgrade process, provide access credentials for components. Complete the credentials in this table before you start the cluster upgrade.
  • You must have the same credentials for all cluster components to run Automatic Discovery. For more information, see Run Discovery.
  • PostgreSQL Databases that are not installed by APM Installer are unsupported for Cluster Upgrade.
  • Cluster Upgrade is unsupported for MOMs that run in the high-availability mode. For example, when two MOMs are present for the same cluster.
Component
Hostname/IP Address
Username
Password
WebView Hostnames (all Hostnames if more than one)
Credentials for servers with APM components or SSH Credentials or Private Keys and Passphrase
Enterprise Manager Credentials
PostgreSQL APM Database Superuser Credentials
Download the Command Center Server Upgrade
Download the latest CA APM Command Center to initiate the cluster upgrade.
Item
Action
Notes
Your Notes
Done
1
Go to CA Support and
Download Management
. Download the APM Command Center Cluster Upgrade Bundle.
Connect to the CA Support Download site, download the Command Center Cluster Upgrade Bundle, and transfer it to the target server.
2
Access the Host server with your Administrator credentials.
For more information about installing Command Center, see the following links:
User ID:
Password:
Upgrade the Cluster
Item
Action
Notes
Your Notes
Done
1
Before you begin your upgrade, consider what type of Discovery you must run for your environment.
Carefully review all parts of this section in the documentation: RunDiscovery
.
2
Log in to the Command Center that has the Cluster Upgrade capability.
On the upgraded Command Center Server, point the browser to
http://<hostname>:8088
or
https://<hostname>:8443
and log in as the default user; User ID: , Password:
acc
.
If you log in on HTTP, you are automatically redirected to HTTPS.
3
In the
Servers
tab, discover the cluster.
Auto Discovery runs a check to discover all components in your environment.
If you have Enterprise Team Center (ETC), run Discovery on all clusters before you upgrade the ETC Master cluster.
Run Manual Discovery through Command Center in the following cases:
  • You have Enterprise Managers on both Linux and Windows
  • You have a Windows-only environment
  • You have differing SSH credentials for your components
  • You do not use SSH as the security protocol
  • You use HTTPS and SSL between Command Center Server and Command Center Controller
4
Select space in the Upgrade Wizard.
The Upgrade Wizard becomes available only if Discovery is successful. Select
Upgrade
, and then select
Check Space
.
You can skip these steps (Steps 4 and 5) in the wizard if you already prepared a backup of your environment with your own tools. Skipping these steps in the following procedure disables the Rollback stage in the wizard. You can roll back your environment in Command Center only if you also create the backup through Command Center.
For ETC, start the upgrade with the ETC Master cluster. You must upgrade the ETC Master cluster before you upgrade any Provider clusters.
5
Create a backup.
Use the default values. You must create a backup to be able to roll back your environment.
6
Conduct a pre-upgrade check.
Enter the Admin credentials for the Manager of Mangers (MOM).
The check process collects the current figures for active agents, live metrics, and historical metrics. The process then verifies that sufficient disk space is available for upgrading all components.
7
Start the upgrade.
Verify that Command Center identified the correct cluster and select the box to accept the EULA.
8
Provide the PostgreSQL Database details.
You do not need credentials for the Oracle Database. Retrieve the database credentials from the database connection configuration for PostgreSQL.
Default credentials for the PostgreSQL administrator user:
postgres
9
Test the connection.
This test verifies that the database credentials are correct before the upgrade starts.
10
Wait for the upgrade to complete.
The Upgrade Summary shows the progress of the upgrade of each component.
During this time, Enterprise Manager is down and CA APM does not collect any data for your environment.
11
Verify the upgrade.
Verify that the number of connected agents after the upgrade is the same as before the upgrade. The metrics should be the same or higher than before the upgrade. Verify that components show the new version.
Look at the production environment in Team Center to verify that everything appears as expected.
For ETC, do not upgrade ETC Provider clusters until you are sure that you upgraded the Master cluster successfully.
12
Roll back the upgrade.
(Optional) Roll back the upgrade if your environment changed or you experience problems with the verification check.
If you upgrade an ETC Master cluster, and the Master cluster upgrade fails, roll back the Master cluster and troubleshoot the cause. Do not upgrade ETC Provider Clusters until after you successfully upgraded the Master cluster.
13
Run the cleanup utility to remove temporary upgrade files.
(Optional) When you are satisfied that the upgrade is successful, remove the backup of the upgrade.
  • After you clean up the environment, you cannot use Command Center to roll back your environment.
  • If an ETC Provider cluster upgrade fails, roll back the provider cluster and troubleshoot the cause. Rerun the Provider cluster upgrade from the beginning.
  • If you cannot upgrade an ETC Provider cluster for any reason, rollback all upgraded Master and Provider clusters in reverse order.
    Example:
    Roll back the ETC Provider clusters first beginning with the last Provider cluster that you upgraded and then roll back the Master cluster.