Applicable scenario

• Product: ZCF
• Component: Cloud
• Version: 5.4.x
• Operating system: H84r
• Deployment topology: dual management nodes
• Trigger: The UI stalls when a sub-account deletes a scaling group and requires a UI patch for performance optimization.

Symptoms

• The UI stalls when a sub-account deletes a scaling group.
• The issue is located in the cloud-platform UI layer.
• The UI change is made only on dual management nodes and does not affect other VM service systems.

Determination method

1. Confirm management services are normal before the change.
2. After applying the patch, log in to the UI and verify that page interaction is restored.
3. Confirm resolution by deleting a scaling group.

Cause

The issue is not in underlying hosts or storage. The cloud-platform UI has a performance stall when a sub-account deletes a scaling group, so the UI-side patch content in zstack-ui is replaced.

Resolution

1. Preparation

• Risk level: medium.
• Confirm management services are normal.
• Disable global HA.
• Back up the database and UI directory.

2. Back up the database and UI directory
3. Back up the database on both management nodes:

   zstack-ctl dump_mysql --file-name zstack-cloud-db-master-back
   zstack-ctl dump_mysql --file-name zstack-cloud-db-slave-back

• Expected result: A database rollback baseline is available.

2. Back up the UI service configuration directory on both management nodes:

   tar -czvf zstack-ui-backup.tar.gz ~zstack/zstack-ui/zstack-ui

• Expected result: A rapid UI-directory rollback copy is available.

3. Verify and deploy the patch
4. Upload the patch archive:

   zstack-tic-5286-5.4.6.tar.gz

• Expected result: The patch archive is on both management nodes.

2. Verify the md5:

   md5sum zstack-tic-5286-5.4.6.tar.gz

• Required value:

   65a8afeda4498f09546e544d812cd4a3

• Expected result: The archive is complete and the version is correct.

3. Extract and overlay the UI directory:

   tar -xzvf zstack-tic-5286-5.4.6.tar.gz -C ~zstack/zstack-ui/zstack-ui/

• Expected result: Patch files are overlaid.

4. Verify and complete
5. Log in to ZStack Cloud and check the UI.

• Expected result: The page is accessible normally.

2. Verify that deleting a scaling group no longer stalls the UI.

• Expected result:

   Deleting a scaling group no longer stalls the UI.

The issue is resolved.

3. Enable global HA.

• Outcome: The change process is complete.

Verification

• The patch archive md5 and content match.
• The UI opens normally.
• Deleting a scaling group as a sub-account no longer stalls.
• The dual-management-node environment resumes normal operation.

Risks and rollback

• Risk: A mismatched patch version or md5 can cause incorrect UI file replacement.
• Risk: Without a zstack-ui directory backup, rollback is more difficult.
• Risk level: medium.
• Rollback: Restore the backed-up UI directory and database copy.