Applicable scenario

• Product: ZCF
• Component: ZCE-X
• ZCE-X version: 6.4.x
• Storage nodes: multi-node storage cluster
• Trigger: The storage UI does not show the expected branding, requiring a frontend-package update.

Symptoms

• The storage UI does not display the expected branding.
• Updating the frontend RPM package is required to restore the branding.
• The change must not affect the storage service itself.

Determination method

1. Before updating, log in to each node UI and confirm normal accessibility and health.
2. After updating, log in to the UI again and confirm that page branding is restored.
3. Run ceph -s before and after the update and confirm that the cluster remains HEALTH_OK.

Cause

The issue is confined to the frontend UI package. Update the package to restore branding without affecting the storage service.

Resolution

1. Preparation

• Risk level: low.
• Check health of all storage nodes and the UI.
• Update one node at a time, validate the first node, then continue with the remaining nodes.

2. Procedure
3. Log in to each storage-node UI and confirm healthy state:

• <storage node management address>:8056/
• Expected result: The page is accessible and health is normal.

2. Log in to the storage node by SSH and check Ceph:

   ceph -s

• Expected result: HEALTH_OK is displayed.

3. Upload the frontend RPM package to the target storage node:

   ZSTACK-V6-XSCALEROS_6.4.200.1-0.el7.centos.noarch.rpm

• Expected result: The target package is uploaded.

4. Update the package:

   rpm -U --force <rpm package>

• Expected result: The frontend package update completes.

5. Log in to that node's storage UI again and confirm that branding is restored.

• Expected result: UI branding displays normally.

6. Run again:

   ceph -s

• Expected result: The storage cluster remains HEALTH_OK.

7. Update remaining storage nodes one by one using the same procedure.

• Expected result: All storage-node UI frontends are updated.

Verification

• Each storage-node page on 8056 is accessible.
• Page branding is restored.
• ceph -s remains HEALTH_OK after each update.

Risks and rollback

• Risk: If the first node's frontend update fails, stop updating subsequent nodes immediately.
• Risk level: low.
• Rollback: No specific RPM rollback command is available. The frontend update does not affect storage services; if the first node fails, do not continue the rollout.