Skip to content

Commit

Permalink
(4.0) tiup: add notice about importing cluster (#6131)
Browse files Browse the repository at this point in the history
  • Loading branch information
Liuxiaozhen12 authored Aug 30, 2021
1 parent 0e11b30 commit c9ce0f7
Showing 1 changed file with 1 addition and 0 deletions.
1 change: 1 addition & 0 deletions upgrade-tidb-using-tiup.md
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ If you have deployed the TiDB cluster using TiDB Ansible, you can use TiUP to im
- `Lightning` / `Importer` is enabled for the cluster.
- You still use the `'push'` method to collect monitoring metrics (since v3.0, `pull` is the default mode, which is supported if you have not modified this mode).
- In the `inventory.ini` configuration file, the `node_exporter` or `blackbox_exporter` item of the machine is set to non-default ports through `node_exporter_port` or `blackbox_exporter_port`, which is compatible if you have unified the configuration in the `group_vars` directory. Or the `node_exporter` or `blackbox_exporter` item of a machine is set to a `deploy_dir` that is different from that of other machines.
- If some nodes in the cluster deployed using TiDB Ansible are deployed without monitoring components, you should first use TiDB Ansible to add the corresponding node information in the `monitored_servers` section of the `inventory.ini` file, and then use the `deploy.yaml` playbook to fully deploy monitoring components. Otherwise, when you perform maintenance operations after the cluster is imported into TiUP, errors might occur due to the lack of monitoring components.
- Support upgrading the versions of TiDB Binlog, TiCDC, TiFlash, and other components.
- Before you upgrade from v2.0.6 or earlier to v4.0.0 or later, you must make sure that no DDL operations are running in the cluster, especially the `Add Index` operation that is time-consuming. Perform the upgrade after all DDL operations are completed.
- Starting from v2.1, TiDB enables parallel DDL. Therefore, clusters **older than v2.0.1** cannot be upgraded to v4.0.0 or later via a direct rolling upgrade. Instead, you can choose one of the following solutions:
Expand Down

0 comments on commit c9ce0f7

Please sign in to comment.