Issue

Use this article to troubleshoot bulk HA VM migration in a ZSphere/ZStone environment.

Symptoms

• Many alerts report HA migration of VMs.
• Management logs also contain management-node, storage, or physical-host offline errors.

Environment

• Product: ZVF
• Version: 4.10.12
• Component: ZMigrate
• Affected object: HA/storage
• Scenario: High availability issues. Confirm the current version, deployment topology, storage type, and operation timeline.

Possible causes

• Brief management-node, storage, or physical-host disconnection triggers an HA determination.
• ZStone storage-path jitter causes VMs to be determined to require migration.
• Network, storage, and host exceptions occur at the same time and must be correlated by timeline.

Resolution

Perform the following checks in sequence, and record the affected objects, occurrence time, complete error messages, and current state.

1. Organize the HA migration-alert timeline and the lists of involved VMs and hosts.
2. Check management-node, physical-host, and storage-side offline/reconnection logs for the same time window.
3. Confirm whether network jitter, a storage-service exception, or a management-node restart occurred.
4. After bulk migration, check VM state, service impact, and the risk of repeated migration.
5. Before manual intervention, confirm the actual running location of every affected VM. Do not manually start a VM on another host until the risk of split-brain operation has been excluded.

Remediation

1. Based on the alert and log timeline, first restore the offline management node, physical host, or storage service to eliminate the underlying fault that triggered HA migration.
2. After confirming the actual running location of the VM, assess whether manual recovery or migration is required.

Verification

• Management-node, physical-host, and storage-service states return to normal, and HA migration alerts no longer continue.
• Each affected VM runs in only one location, and service access and VM state are normal.