Server migration error: Client.Snapshot.CreateFailed

Symptoms

During a migration in Server Migration Center (SMC), a migration task fails with the error message "Client.Snapshot.CreateFailed".

Causes

During block replication, the SMC client creates a snapshot device for each disk partition on the source system to synchronize data. This error occurs when snapshot creation fails. Possible reasons vary by operating system:

• Windows:

  • Insufficient space on the disk partition.

  • Third-party security or antivirus software is installed.

  • The Volume Shadow Copy storage usage limit for the disk partition is too low.

  • The Volume Shadow Copy service is not working correctly.

• Linux

  • Insufficient space on the disk partition.

  • The available system memory is less than 300 MB.

  • The snapshot device for the disk partition is not working correctly.

Solutions

Choose the solution based on the operating system of the source system.

• Windows

  Open Event Viewer and check for errors related to vss, volsnap, or disk. Then resolve any errors you find.

  • If the error indicates insufficient disk space or that a Volume Shadow Copy snapshot cannot grow:

    • Check the free space on the disk partition. Make sure at least 10% to 15% of the total space is available.

    • Check whether the storage usage limit for the volume is too low. If so, increase the limit or remove it, and then try again. To adjust the limit, open the Local Disk (C:) Properties dialog box, and select the Shadow Copies tab. Select the C:\ volume and click Settings.... In the Settings dialog box, for Maximum size, select Use limit and set the value to 4095 MB. Click OK.

  • If you receive an "access denied" error with error code 0x80070005, third-party security or antivirus software may be blocking the operation. Grant disk backup permissions or disable the software, and then try again.

  • Verify that the Volume Shadow Copy service is running correctly.

• Linux

  1. Check whether the /dev/smcss9 device exists, and run dmesg | grep smcss to check for errors.

     1. Ensure that at least 300 MB of system memory is available.

     2. Check the free space on the disk partition. Make sure at least 10% to 15% of the total space is available.

     3. In the SMC client installation directory, run ./Check/client_check --ssclear smcss to clear residual snapshot devices, and then try again. For details on finding the installation directory, see View the SMC client.

  2. If the error message contains "io scheduler not support", change the I/O scheduler for the affected disk.

     1. Identify the index of the problematic disk.

        Find the msg field in the error message, for example, msg=CreateSnapshot 0-x Failed. In this example, 0-x indicates the first disk, 1-x indicates the second disk, 2-x indicates the third disk, and so on.

     2. Use the disk index from the previous step to find the disk device name.

        Run lsblk to find the corresponding disk device name. The index starts from 0.

        In the command output, index 0 corresponds to the disk device vda (the 40 GB system disk), and index 1 corresponds to the disk device vdb (the 100 GB data disk).

        # lsblk
        NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
        vda    253:0    0    40G  0 disk
        ├─vda1 253:1    0     2M  0 part
        ├─vda2 253:2    0   200M  0 part /boot/efi
        └─vda3 253:3    0  39.8G  0 part /
        vdb    253:16   0   100G  0 disk
        └─vdb1 253:17   0   100G  0 part /data

     3. Use the disk name from the previous step to view the I/O scheduler used by the disk.

        For example, to view the I/O scheduler for vdb, run cat /sys/block/vdb/queue/scheduler.

        The brackets [] indicate the active I/O scheduler. In this example, the active scheduler is bfq.

        cat /sys/block/vdb/queue/scheduler
        mq-deadline kyber [bfq] none

        Note

        The output lists all available I/O schedulers: mq-deadline, kyber, bfq, and none.

     4. Change the I/O scheduler from bfq to another available scheduler.

        For example, to switch to mq-deadline:

        Run sudo echo mq-deadline > /sys/block/vdb/queue/scheduler.

        Then run cat /sys/block/vdb/queue/scheduler to verify the change.

        cat /sys/block/vdb/queue/scheduler
        [mq-deadline] kyber bfq none

        In the output, [mq-deadline] indicates that mq-deadline is the active I/O scheduler.

After you resolve the issue, run the SMC client again to verify the result.