Managing patching

The BMC Database Automation (BDA) patch repository provides an easy method for creating a patch package and automatically applying patches to multiple nodes, one at a time or simultaneously. A patch package consists of the patch itself (from either Oracle, Microsoft, or Sybase) and any pre-provisioning or post-provisioning scripts to be run. After a patch package is created, it can be applied to the hosts and databases you specify, and it can be exported to another Manager. In addition to applying interim software patches, you can roll back, where applicable, a previously applied patch, when necessary. Patches can be applied to nodes that have been provisioned outside of BDA in addition to nodes that have been provisioned by BDA.

Starting from version 8.9.02, you can divide the patch application process in the following phases to optimize the patch application process: 

1. Pre-stage the patch: This step copies the patch media to nodes and extracts it.

   Note

   You can pre-stage a patch outside the patch window, which typically has bandwidth and time constraints. After pre-staging, you can apply the patch in the patch window.

2. Apply the patch: This step determines whether the patch media is already present on the node. If the media exists, it is compared with the patch media to be copied using checksum. If the checksum matches, then the patch is applied to the node. If the media is not present or the checksum does not match, the media is copied and then the patch is applied to the node.

BDA 2019 Release 02 adds the following features to the patching process:

• If a patching job fails on a node, the job continues to run on all other nodes until completion. Continuing to run ensures that the first run of the job successfully patches any node that can be patched.
• All services that shut down during the patching job restart when patching is complete, whether or not patching is successful. Thus, the databases return to a functional state without waiting for the entire job to complete.
• When a patching job fails on a node, the job status is set to Failed. This status indicates that the BDA administrator should address the issues that caused the failure and then rerun the job.

• When you rerun a patching job following a failure, the job does not run on the nodes where patching was completed.

  Example

  A patching job runs on nodes 1 through 10. The job successfully patches nodes 1 through 5 but fails on nodes 6 through 10. If you run the job again, it runs only on nodes 6 through 10.

The following topics provide instructions for managing patching by using BDA:

• Managing patch packages for Oracle quarterly packages
• Creating a patch package
• Patching a database
• Rolling back a patch package
• Pre-staging a patch package