Docs

Elastic Compute Service

Basic volume operations

Handling failed elastic remote volumes

If your elastic remote volume fails, you can fix it by replacing or deleting the volume. This topic explains how to replace a failed volume using the ECS console.

tip

Elastic remote volumes are currently in closed beta. Contact your account manager to have your account added to the allowlist and start using it.

Introduction

If an elastic remote volume enters an error state due to a software or hardware failure, it is automatically detached from the ECS instance and a failure notification is sent to you. You then need to replace or delete the failed volume, depending on your needs.

Notes

Category	Description
Data cannot be recovered	Replacing or deleting a failed volume will result in the loss of the volume's data. After replacing a volume, the system provides a new, blank one. Before proceeding, confirm that you can discard the data or that you have other copies.
Volume ID remains unchanged	The volume ID remains the same after replacement and does not affect the billing entity. However, the volume UUID changes. If you have set up automatic mounting at startup using the UUID, you must update it when recovering the volume.
Billing continues during failure	An elastic remote volume is still billed even after it has failed.
Only Linux instances require configuration changes	If a volume is detached from an ECS instance after it has failed, the original mount entry in /etc/fstab may prevent the instance from restarting properly. Therefore, for Linux instances, you need to add the nofail parameter to the configuration. You can skip this step for Windows instances and for scenarios where the volume was not attached before the failure.

Prerequisites

Your elastic remote volume is currently in an "Error" status.

Your account has the necessary permissions to perform operations on the volume and its associated ECS instance.

Replacing a failed volume

Follow the steps below to replace the failed volume.

On the "Elastic remote volume" tab, locate the volume and click "... > Replace failed volume" on the right.

In the "Replace failed volume" dialog box, complete the following four steps in sequence: modify configuration, replace failed volume, attach volume, and recover volume.

The volume ID remains the same during the replacement process. Once complete, the volume status becomes "Available".

Modifying configuration

Check the volume status before the failure occurs.

If it meets one of the following conditions, you can skip modifying configuration and proceed to the next one. Otherwise, perform the configuration modification steps below.

The volume was not attached before it failed.

The last ECS instance to which the volume was attached ran the Windows operating system.

In the "Replace failed volume" dialog box, click the ECS instance to which the volume was last attached to access the instance overview page.

Click "Remote Connection" to log in to the ECS instance. For instructions, see Using the console to log in to a Linux instance.

View mount details.

Run the df -h command to view mount details.

Unmount the file system for the volume.

Run the umount command to unmount the file system. For more information, see Unmounting a data volume in Linux.

Add the nofail parameter.

Run the vim /etc/fstab command and add the nofail parameter to the mount options for the volume. This will ensure that the instance can still start if the mount fails.

Example:

UUID=<Original volume UUID> /mnt/data ext4 defaults,nofail 0 2

Proceed to the next step.

Return to the "Replace failed volume" dialog and click "Next".

Replacing a failed volume

Verify the failed volume information.

On the "Replace failed volume" page, confirm the volume ID and specifications.

Replace the failed volume.

tip

The system will create a new, blank volume to replace the failed one. The volume ID will remain the same. This operation will erase all data. Proceed with caution.

Check the box next to "I acknowledge the risk and confirm to replace the failed volume", and then click "Replace failed volume".

Wait until the volume status changes to "Available", and then click "Next".

Attaching the volume

Verify the failed volume information.

On the "Attach Volume" page, select the ECS instance to which you want to attach the volume.

Attach to the original instance (Default): This option attaches the volume back to the instance it was last attached to.

Attach to another instance: Select another ECS instance for attachment.

Attach the volume.

Click "Attach". When "Volume status" changes to "Attached", click "Next".

Recovering the volume

The new volume is blank after replacement. Before using it, you need to create and mount a file system. The following example shows how to create a file system and set up automatic mounting at startup for a Linux instance.

Create a file system.

Run the fdisk -l command to find the name of the new volume.

Run the mkfs -t <file system format> new volume command to create a file system directly on the raw device.

Different file systems support different partition sizes. Select a file system that best meets your business needs.

Run the mkdir <mount point> command to create a mount point.

Run the mount new volume <mount point> command to mount the volume to the newly created mount point.

Run the df -th command to verify the mount result.

Setting up automatic mounting at startup

Run the blkid volume partition command to query the UUID of the partition.

Run the vim /etc/fstab command to open the fstab file.

Press i to enter insert mode. Then, add a new entry similar to the following:

This content is for example purposes only. Use actual values from your environment. The parameters are explained below:

Parameter	Example value	Description
UUID	UUID=0b3040e2-1367-4abb-841********	The UUID of the partition.
Mount directory	/mnt/sdc	The mount directory of the partition. You can find this using the df -th command.
File system format	ext4	The file system format of the partition. You can also find this using the df -th command.
Mount options	defaults,errors=remount-ro	Mount options for the partition. The defaults are usually fine.
Linux dump backup options	0	0 means to not use the Linux dump backup. Since dump backups are generally not used in modern systems, you can set this value to 0. 1 means to use the Linux dump backup.
fsck options	2	Determines whether fsck checks the volume at startup. 0 means no check. Enter 1 if the mount point for the partition is the root directory (/). For other partitions, the value can start at 2 because the system checks partitions in ascending order.

Press "Esc", type :wq, and press "Enter" to save and exit the editor.

Run the mount -a command to reload all entries in the /etc/fstab file.

Run the mount | grep command to view information about the mounted file system.

The following command returns information about /dev/vdc:

mount | grep /mnt/sdc

If the output is similar to the following, automatic mounting is working:

root@ecs-test-0001 ~]# mount | grep /mnt/sdc
/dev/vdc on /mnt/sdc type ext4 (rw,relatime,errors=remount-ro,data=ordered)

Deleting a failed volume

Deleting a failed volume directly releases its resources and renders data recovery impossible. This option is appropriate when the volume is no longer needed.

Modifying configuration

The process is the same as when replacing a failed volume. For Linux instances, clean up the residual mount entries in /etc/fstab to prevent startup failure. For instructions, see the Modifying configuration section above.

Deleting a failed volume

Verify the failed volume information.

On the "Delete failed volume" page, confirm the volume ID and specifications.

Delete the failed volume.

Check the box next to "I am aware of the risk and confirm delete failure", and then click "Confirm delete".

After a volume is successfully deleted, a prompt appears on the "Elastic remote volume" page that reads, "Volume deleted successfully".

Last updated: 2026.08.18 15:57:56