Search Docs...
⌘ K
When a host or SCVM in the cluster experiences an unrecoverable failure and needs to be permanently removed, or when IT administrators need to replace a server or relocate a node, the host can be safely and efficiently removed through the AOC interface.
Note that the operations in this section apply only to the removal of storage nodes in clusters with more than three nodes. If the cluster contains only three nodes, removing one will violate the minimum node requirement of ACOS, resulting in potential data recovery risks. In such cases, replace the node with a new node instead since it cannot be directly removed. For details, refer to the Replacing a node section.
Risk warning
Removing a host from the cluster will result in the permanent loss of all data stored on that host. This operation cannot be undone. Proceed with caution.
Precaution
If you plan to add a new host to replace and rebuild after removing the current host, record the current host's hostname and network configuration in advance, so you can apply the same settings to the new host.
After removing an unresponsive host, do not start it. If this host is intended for other use, make sure to reinstall the operating system before repurposing it.
Procedure
In the host list on AOC, click the ellipsis (...) to the right of the target host, and then select Remove host.
In the opened Remove host dialog, the system will perform a series of prechecks. If any check fails to meet the requirements, follow the recommended actions to manually adjust the configuration, and then click Retry to try removing the host again.
| Precheck item | Suggested operation upon failure |
|---|---|
The current host is not responding or in maintenance mode. |
Place the host in maintenance mode or shut down the host. |
| (AVE only) No virtual machines on the current host except those in the recycle bin. | Manually adjust business virtual machines. If system service virtual machines, such as the file controller, reside on the current host, contact technical support to take the file controller offline and migrate it to another host. |
| (AVE only) No USB devices mounted on virtual machines on the host. | Detach the USB device from the virtual machine. |
| The current host is not a master node. | Convert the host to a storage node. |
| The number of master nodes in this cluster meets the requirement. | Hosts cannot be removed if the requirements are not met. |
| The ZooKeeper and MongoDB services in the cluster are running properly. | Fix the faults. |
The storage services on other hosts in the cluster are not in the removing or idle state. |
Wait for the chunk service on other hosts to recover to a healthy state. |
| After removing the current host, the number of hosts in the cluster that have healthy chunk services and are not in the process of being removed must be no less than the total number of data blocks (K) and parity blocks (M) configured for any virtual volume or NFS export using erasure coding. | Removing the host will cause data loss, and it is not permitted to remove the host. |
| The available storage capacity of other hosts in the cluster is sufficient. | Scaling out the cluster or delete some snapshots or virtual machines to free up cluster capacity. |
| (AVE only) No VM placement group policies for the current host exist. | Remove the current host from the placement group. |
There are no other hosts with the following statuses in the cluster:
|
Wait for the host to complete ongoing tasks and exit the current state. |
| No data recovery in progress in the cluster. | Wait for the data recovery to complete. |
| No dead pextent in the cluster. | Contact technical support to restore cluster data. |
The cluster is not upgrading. |
Wait for the cluster upgrade to complete. |
(Active-active cluster only) The witness node in the active-active cluster is healthy. |
Restore the witness node. |
| (Active-active cluster only) After removing the current host, the number of healthy hosts in the primary availability zone must be no less than 2, the number of healthy hosts in the secondary availability zone must be no less than 1. | Restore unhealthy hosts and proceed with cluster expansion. |
After all checks have passed, enter the name of the host to be removed in the Confirm to delete input box, and click Remove in the lower-right corner of the dialog box.
Once the removal task is submitted, the page will display the task progress and whether the host is successfully removed. If the removal fails, select Retry again to resubmit the task.