Problem Symptom
During the independent upgrade of DPC nodes or the upgrade of DeployManager, the available space and memory of the nodes change. As a result, the CloudA component cannot pass the pre-installation check after being uninstalled. As a result, the CloudA upgrade fails.
Problem Diagnosis
1. Check CloudA deployment log file /var/log/deploy/clouda/script/clouda.deploy.script.log. The log file shows that CloudA fails to be upgraded due to insufficient memory and disk space.
Causes
The process to upgrade CloudA is to uninstall CloudA of the earlier version and then install that of the latest version. If the node environment changes after CloudA is uninstalled, such as the usable space or memory decreases, CloudA will fail the pre-installation check. As a result, CloudA of the latest version fails to be installed.
Solution
1. Run the free command to check the node environment information. Ensure that the usable memory of the node is larger than 4.5 GB and the usable spaces of the /opt, /tmp, and /var/log directories are larger than 4.5 GB, 5 GB, and 50 MB, respectively. After ensuring that the node environment meets the CloudA installation requirements, perform steps 2 and 3 to continue the upgrade. Rollback is not supported.
2. In DeployManager upgrade scenarios, manually upgrade CloudA on each node by referring to 3.6.14-CloudA Fails to Be Automatically Upgraded Due to a Node Fault.
3. If independent DPC node upgraded scenarios, upgrade the DPC nodes again by referring to the OceanStor 100D 8.0.3 Compute Node Installation, Upgrade, Rollback, and Uninstallation Guide
