The following describes how to upgrade a High Availability One Identity Safeguard for Privileged Sessions (SPS) cluster to a newer firmware version. One Identity recommends that you always use the latest maintenance release available.
|
Caution:
If you have nodes with the Search Minion role configured, see Upgrading a High Availability One Identity Safeguard for Privileged Sessions (SPS) cluster to avoid a critical error. |
|
Caution:
When upgrading to a new major release (that is, to a new Feature Release or a new Long-Term Supported release), always follow the instructions of the How to upgrade to One Identity Safeguard for Privileged Sessions guide for that release, as it contains more detailed instructions (available at the Safeguard for Privileged Sessions Documentation page). |
To upgrade a High Availability SPS cluster to a newer firmware version
-
Navigate to Basic Settings > System > Firmwares.
-
Upload the new firmware: Browse for the firmware .iso file and then click Upload.
To read the Upgrade Notes of the uploaded firmware, click on the icon. The Upgrade Notes are displayed in a pop-up window.
-
Click Test for the new firmware to check if your configuration can be upgraded to version 7.2.1. If the test returns any errors, correct them before continuing the upgrade process. If you encounter any problems, contact our Support Team.
Caution: Proceed only if the upgrade test is successful.
-
For the new firmware, select After reboot. This will activate the new firmware after reboot.
-
Navigate to Basic Settings > High Availability, and verify that the new firmware is active on the secondary node. This might take a few minutes.
-
In Basic Settings > High Availability > Other node, click Shutdown.
-
Restart the primary node: click This node > Reboot.
SPS attempts to boot with the new firmware. Wait for the process to complete.
-
Login to the SPS web interface to verify that the primary node upgrade was successful.
Navigate to Basic Settings > System > Version details, or check the system log for the version numbers SPS reports on boot. In case you encounter problems, you can find common troubleshooting steps in Troubleshooting.
-
Use the IPMI to power on the secondary node.
The secondary node attempts to boot with the new firmware, and reconnects to the primary node to sync data. During the sync process, certain services (including Heartbeat) are not available. Wait for the process to finish, and the secondary node to boot fully.
-
Navigate to Basic Settings > High Availability and verify that the secondary node is connected, and has the same firmware versions as the primary node.
-
Click Activate Slave to finish the upgrade process and verify HA functionality.
|
Caution:
If you have nodes with the Search Minion role configured, the Search Minion nodes must be upgraded first during High Availability cluster upgrade. If you upgrade the Search Master node first, it is possible that a Search Minion node will create a legacy Elasticsearch index before the init script on the Search Master creates a new one. In this case, the Elasticsearch index will contain invalid schema mapping data, therefore as the High Availability cluster’s schema changes, the Search Minion nodes cannot push their documents into Elasticsearch, resulting in a critical error. |
To avoid the critical error mentioned above, follow the method below to upgrade a High Availability SPS cluster with Search Minion nodes to a newer firmware version
- Set the Search Master node for upgrade so that it uses the newer firmware version for reboot. To do this, complete steps 1-11 of To upgrade a High Availability SPS cluster to a newer firmware version.
- Click Shutdown.
- Upgrade your Search Minion nodes one after the other, using the method in steps 1-2 above.
- Reboot the Search Master node, which will now boot with the newer firmware version.
With this method you detach the Search Minion nodes from the Search Master node and upgrade them separately before any other nodes. As a result, the whole High Availability cluster will use the newer firmware version after reboot.
If you have accidentally upgraded the Search Master node first and encounter this critical error, contact our Support Team.