Rollback

This section describes how to roll back after upgrading an on-premises ALM Octane server. This may be necessary if for some reason the upgrade fails or performance is slow.

Depending on when you want to roll back, there are different steps to perform.

Note: To roll back you need the pre-upgrade backups of all configuration files including octane.conf from each node.

After the upgrade's setup validation phase

You can roll back after the upgrade's setup validation phase, whether it passed or failed.

If the upgrade reached setup validation, the following have been modified: 

  • Previously-deployed files

  • ALM Octane configuration files, including octane.conf

To roll back the deployed files, including octane.conf

  1. Revert to the previous rpm file: rpm -Uvh --oldpackage <filename>

  2. Revert to the backup of octane.conf, and any other configuration files you previously used.

  3. Re-initialize the ALM Octane server (the octane service).

Back to top

After a site schema has been upgraded

You can roll back after the site schema has been upgraded.

If the upgrade upgraded the site schema, the following has been modified:

  • The site schema (database)

  • Elasticsearch indexes

  • ALM Octane configuration files, including octane.conf

To roll back the site schema

  1. Stop the ALM Octane server (the octane service).

  2. Revert to a backup of the site schema.

  3. Revert to a backup of Elasticsearch indexes.

  4. Revert to the previous rpm file: rpm -Uvh --oldpackage <filename>

  5. Revert to backups of ALM Octane configuration files, including octane.conf.

    Note: In previous versions, these configuration files were located in the opt/octane/conf directory.

  6. Re-initialize the ALM Octane server (the octane service).

Back to top

After space schema has been upgraded

If the upgrade upgraded the space schema, the following have been modified:

  • Previously-deployed files

  • Elasticsearch indexes

  • ALM Octane configuration files, including octane.conf

  • The site schema

  • The space schema

Rolling back a single space is relevant after upgrade of a space failed. In this case, fixes are required depending on the cause of the failure, as seen in the logs and in the UI.

Follow the steps for one of the following options.

Rollback option Steps
To roll back changes to the space schema
  1. Revert to the backup of the space schema.

  2. Revert to the backups of Elasticsearch indexes related to the specific space.

    Space-specific indexes can be identified by the space logical name embedded in their name, using the pattern mqm_{space logical name}_*.

    Note: There are multiple Elasticsearch indexes for each space. Make sure to roll back all of them.

  3. Revert to the repository backup of this specific space.

  4. Fix what caused the upgrade to fail.

  5. Run the following API to repair the space:

    POST {octane server}/admin/shared_spaces/repair?ids={space_id}

    Tip: To repair multiple spaces, provide the space_ids separated by commas.

  6. Upgrade again.

To roll back the entire upgrade

Follow the steps for To roll back the site schema.

Back to top

After upgrade completed

If the upgrade completed successfully, the following have been modified:

  • ALM Octane configuration files, including octane.conf

  • The site schema

  • The space schema

  • Elasticsearch indexes

  • ALM Octane repository files

To roll back the entire upgrade

  1. Follow the procedure To roll back the site schema.

  2. Revert to backups of all space schemas.

  3. Revert to backups of all Elasticsearch indexes.

  4. Revert to backup of the previous repository.

Back to top

After upgrading cluster nodes

If you upgraded additional cluster nodes, the following has been modified on the cluster nodes: 

  • Previously-deployed files

  • ALM Octane configuration files, including octane.conf

To roll back to the rpm package

  1. Revert to the previous rpm file on each cluster node: rpm -Uvh --oldpackage <filename>

  2. Revert to backups of ALM Octane configuration files, including octane.conf.

    Note: In previous versions, these configuration files were located in the opt/octane/conf directory.

  3. Re-initialize the ALM Octane server (the octane service) on each cluster node.

Back to top