Cloudera Docs

Data Lake rolling upgrades

The Data Lake rolling upgrade allows you to upgrade the Data Lake Runtime and OS without stopping attached Data Hubs or Data Services. This allows workloads to continue running during the Data Lake upgrade operation.

Similarly to the classic Data Lake upgrade logic, a Data Lake rolling upgrade first upgrades the Runtime version and then the OS.

To run a Data Lake rolling upgrade, the following requirements must be met:

  • The Data Lake must be Runtime version 7.2.17.300+ to perform a rolling upgrade directly to 7.2.18+.
  • The Data Lake OS must be RHEL 8. Rolling upgrades are not generally available from Data Lakes on CentOS.
  • The Data Lake must be an Enterprise Data Lake (EDL). Rolling upgrades to 7.2.18 are not generally available from medium duty or light duty Data Lakes.
If your Data Lake does not meet these requirements, you will first have to do a traditional Runtime upgrade, upgrade your OS from CentOS to RHEL, or resize your Data Lake to EDL before a rolling upgrade can be performed.
Current Runtime version Current OS Current Data Lake Shape Rolling upgrade support?
7.2.17.300+ RHEL 8 EDL Yes, directly to 7.2.18+

In some circumstances, a rolling upgrade may not be supported for a Data Lake cluster, but can be enabled through entitlement. Some cluster services might become unavailable during this type of upgrade, and running workloads could be impacted. The Data Lake upgrade UI displays information about whether a rolling upgrade is available, unavailable, or may be available under entitlement. For instructions on performing a Data Lake upgrade, including rolling upgrades, see Upgrading a Data Lake. For information about obtaining an entitlement for rolling upgrade, contact Cloudera Customer Support.