Environment

• Product: ZCF
• Component: ZCE-X
• Affected objects: Newly added storage nodes, XDC, Ceph, and the Kylin kernel
• Version: 6.4.200.1
• Trigger condition: New storage nodes may connect to CSI-iSCSI in the future, so XDC kernel mode must be enabled during deployment.

Symptoms

• New storage nodes will connect to CSI-iSCSI in the future.
• XDC kernel mode must be enabled in advance on new primary-storage nodes that are not yet carrying workloads.
• The change applies only to new storage nodes and must not affect compute nodes.

Diagnosis

1. Confirm that the environment is within scope and record versions, system architecture, operating system, deployment mode, and storage type.
2. Confirm that each target is a newly added storage node with no running services.
3. Run ceph -s to check storage health.
4. After the change, run systemctl status xdc to confirm normal XDC service status.

Cause

The new storage nodes may support CSI-iSCSI in the future. Enable XDC kernel mode before they enter service to avoid later kernel-level changes.

Resolution

1. Prerequisites

• Risk level: Low
• Operate only on new storage nodes. Do not operate on compute nodes.
• Process 1 node at a time and confirm normal status before continuing.
• Upload the target kernel package and verify its MD5 checksum.

2. Procedure

1. Check storage status.

   ceph -s

• Expected result: The Ceph cluster is healthy.

2. Upload the kernel package and verify its MD5 checksum.

• File: kernel-4.19.90-89.11.tar.gz
• MD5: e0925e08cadf890e439d3438a6061ffd

   tar xzvf kernel-4.19.90-89.11.tar.gz

• Expected result: The package is extracted.

3. Downgrade the kernel to the specified version.

   cd kernel-4.19.90-89.11/
   yum install --disablerepo=* ./kernel-*

• Expected result: The target kernel is installed.

4. Confirm that the system will boot with kernel 89.11.

   cat /boot/grub2/grubenv

• Expected output: saved_entry=Kylin Linux Advanced Server (4.19.90-89.11.v2401.ky10.x86_64) V10 (Halberd)
• Expected result: The default boot entry uses the target kernel.

5. Enable XDC kernel mode by setting xdc_proxy_feature to false in /etc/xdc/xdc.conf.

• Expected result: XDC is configured for kernel mode.

6. Restart the server.

   reboot

• Expected result: The node starts with the target kernel.

7. Check storage and XDC status before processing the next node.

   ceph -s
   systemctl status xdc

• Expected result: Ceph and XDC are healthy.

Verification

• Run ceph -s after each restart and confirm cluster health.
• Run systemctl status xdc and confirm normal status.
• Confirm the target kernel and xdc_proxy_feature=false on each target node.

Risks and rollback

• Risk: Changing the kernel and restarting the wrong node may affect online services.
• Risk: Changing multiple nodes without individual verification amplifies cluster risk.
• Rollback: Restore the original kernel boot entry and xdc.conf, and confirm Ceph health before deciding whether to retry.