Skip to main content
Version: CES 19 - Squid

Cephadm-Managed Cluster to CES

CES provides container images for cephadm, so an existing cephadm-managed cluster can move to a CES image through the orchestrator upgrade workflow.

Before upgrading​

  1. Read the CES 19.2.6-1 release notes.
  2. Review the Ceph Squid upgrade documentation.

CephX key rotation​

CES 19.2.6-1 contains CVE fixes that require an additional CephX key rotation during the upgrade. cephadm automates the rotation of the keys it manages as part of the orchestrator upgrade process.

Client keys require manual rotation

cephadm does not rotate CephX client keys. This includes every CephX entity whose name starts with client.*. Inventory and rotate these keys separately, then update the keyrings or credentials used by each affected client.

For background and instructions, see the Ceph documentation on upgrading and rotating CephX keys.

The current CES image is:

harbor.clyso.com/ces/ceph/ceph:ces-v19.2.6-1

Start the upgrade​

ceph orch upgrade start \
  --image harbor.clyso.com/ces/ceph/ceph:ces-v19.2.6-1

Monitor progress with:

watch ceph -s
ceph -W cephadm --watch-debug
watch ceph orch upgrade status

To pause, resume, or stop the orchestrator upgrade, use:

ceph orch upgrade pause
ceph orch upgrade resume
ceph orch upgrade stop

Do not stop an upgrade without first understanding the cluster state and contacting support when appropriate.

Verify the migration​

After the orchestrator reports completion, verify the cluster and daemon versions:

ceph -s
ceph versions
ceph orch ps

Confirm that all expected daemons use harbor.clyso.com/ces/ceph/ceph:ces-v19.2.6-1 and that the cluster has returned to a healthy state.

CES 18 to CES 19​

CES 18.2.8-1 is based on Reef 18.2.8, while CES 19.2.6-1 is based on Squid 19.2.6.

There are no known issues to upgrading.

After migration, see the CLYSO Monitoring Module for CES-specific monitoring and diagnostics features.