Upgrading a cluster

Upgrade Ceph clusters by using the dashboard.

Before you begin

Verify that your upgrade version path and operating system is supported before starting the upgrade process. For more information, see Compatibility matrix and Supported upgrade and crossgrade paths.

Before starting to upgrade IBM Storage Ceph, see Installation and upgrade requirements.

Before you begin, make sure that you have the following prerequisites in place:

Important: These items cannot be done through the dashboard and must be completed manually, through the command-line interface, before continuing to upgrade the cluster from the dashboard.

Note: For detailed information, see Upgrading cluster in a disconnected environment and complete steps 1 through 9.

The latest cephadm.
```
dnf update cephadm
```
The latest cephadm-ansible.
```
dnf update cephadm-ansible
```

The latest cephadm pre-flight playbook.

ansible-playbook -i INVENTORY_FILE cephadm-preflight.yml --extra-vars "ceph_origin=custom upgrade_ceph_packages=true"

Run the following Ceph commands to avoid alerts and rebalancing of the data during the cluster upgrade:

ceph health mute DAEMON_OLD_VERSION --sticky
ceph osd set noout
ceph osd set noscrub
ceph osd set nodeep-scrub

About this task

Cluster images are pulled automatically from IBM Container Registry on IBM Cloud (cp.icr.io). Optionally, use custom images for upgrade.

Procedure

View if cluster upgrades are available and upgrade as needed from the Administration > Upgrade page on the dashboard.

Note: If the dashboard displays the Not retrieving upgrades message, check if the registries were added to the container configuration files with the appropriate log in credentials to Podman or docker.
Due to known issue BZ #2168121, the updates are not retrieved properly during the ceph orch upgrade ls command that is being used. To upgrade, use the Upgrade using custom image option. Get the latest custom image name from Table 1 in Configuring a private registry for a disconnected installation.

Click Pause or Stop during the upgrade process, if needed. The upgrade progress is shown in the progress bar along with information messages during the upgrade.
Note: When stopping the upgrade, the upgrade is first paused and then prompts you to stop the upgrade.
Optional: View cluster logs during the upgrade process from the Cluster logs section of the Upgrade page.
Verify that the upgrade is completed successfully by confirming that the cluster status displays OK state.

What to do next

After verifying that the upgrade is complete, unset the noout, noscrub, and nodeep-scrub flags.

ceph osd unset noout
ceph osd unset noscrub
ceph osd unset nodeep-scrub

For example,

[ceph: root@host01 /]# ceph osd unset noout
[ceph: root@host01 /]# ceph osd unset noscrub
[ceph: root@host01 /]# ceph osd unset nodeep-scrub