Create and manage snapshot/schedule

Snapshot

Create snapshot

Create a new snapshot from the current state of a volume.

caution

If you have expanded the size of a volume, proceed with a full backup even if there were previously created snapshots. If you try to perform an incremental backup, the creation may fail, or the created snapshot may not be usable.

Create snapshot by selecting volume

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.

2. Click the Volume menu, then select Create snapshot from the [More] icon > Create snapshot.

3. Enter the required information and click the [Create] button.

   Category	Description
   Selected volume	The volume to create a snapshot
   Snapshot type	Snapshot creation method - Full: A full backup that creates a copy of all data at the time of backup regardless of data changes   ᄂ e.g. Backing up a 50GB volume creates a 50GB snapshot - Incremental: An incremental backup that only backs up data changed since the last backup point   ᄂ e.g. If only 1GB of data has been added since the last backup, an incremental backup creates an additional 1GB snapshot   ᄂ Possible only after completing the first full backup
   Snapshot name	Automatically set according to the volume name, but can be entered manually - Only English, numbers, and - are allowed
   Description (Optional)	Write a description for the snapshot

Create snapshot from Snapshot tab

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. Click the Snapshot menu, then select Create snapshot located at the top right.
3. Enter the required information and click the [Create] button.

info

Even if you hold a full backup through a snapshot schedule, manual creation of a snapshot requires one initial creation of a full snapshot.

View snapshot list

Check the list of owned snapshots and basic information of each snapshot.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.

2. In the Snapshot menu, find the list of created snapshots.

   Category	Description
   Snapshot Search	Search for volumes using filters or keywords - Filter attributes: Target Volume / Snapshot name / Snapshot Schedule Name / Status / Creator
   Snapshot	Snapshot type - Name and description of the snapshot
   State	State of the snapshot
   Target volume ID	Unique ID of the volume where the snapshot was created.
   Size	The size of the snapshot
   Snapshot schedule	- '-': Snapshots created directly by the user - Snapshot schedule name: Snapshots automatically created by policy
   Creator	User who created the snapshot
   Created	Date and time the snapshot was created - Hovering over the creation date shows elapsed time
   More	Additional functions - Modify snapshot: To modify the name and description of the snapshot - Restore snapshot: To restore a volume from the snapshot - Delete snapshot: Delete a snapshot

Restore volume from snapshot

You can select a snapshot to restore the volume to its state at the time the snapshot was created.

kr-central-2

Restore snapshot from snapshot tab

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Snapshot menu, select Restore Snapshot from the [More] icon > Restore Snapshot
3. Click the [Restore] button.

Restore snapshot from volume tab

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Volume menu, select the volume.
3. Click the Snapshot menu, then select Restore Snapshot from the [More] icon > Restore Snapshot.
4. Click the [Restore] button.

Delete snapshot

You can delete snapshots that are no longer in use.

• If there are multiple incremental backup snapshots for a single volume, only the last created incremental (Incremental) backup snapshot can be deleted.
• If there is more than one incremental backup snapshots, the full backup snapshot of volume cannot be deleted.

caution

Deleted snapshots cannot be recovered.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Snapshot menu, select Delete Snapshot from the [More] icon > Snapshot Delete .
3. Enter the name of the snapshot to delete and click the [Delete] button.

Snapshot schedule

GPU services automatically create and delete snapshots of the current state of a volume according to the set schedule. The method is as follows:

Create snapshot schedule

You can create a new snapshot schedule.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.

2. Select the Snapshot Schedule menu, then click the [Create snapshot schedule] button located at the top right.

3. In the Create snapshot schedule, enter the required information and click the [Create] button.

   

   Item	Description	Description
   Basic information	Snapshot schedule name	Name of the snapshot schedule
   	Description(Optional)	Enter a description for the snapshot schedule
   Backup type	Full	Performs a full backup at every cycle - Creates a copy of all data at the backup point regardless of data changes
   	Incremental	Performs an incremental backup at every cycle - Only backs up data changed since the last backup point - First full backup required
   Snapshot creation cycle		Set the cycle for automatic snapshot creation - Daily, weekly, monthly, yearly
   Deletion schedule		Automatically deletes snapshots created by the schedule after the set elapsed time - Max: Deletes the oldest snapshot first if the number of automatically created snapshots exceeds the set number - Time elapsed : Automatically deletes snapshots when they reach the set elapsed time

View snapshot schedule list

You can view the list of snapshot schedules and their basic information.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.

2. In the Snapshot Schedule menu, find the list of currently operating snapshot schedules.

   Category	Description
   Name	The name of the snapshot schedule specified by the user
   Creation frequency	The snapshot schedule's automatic snapshot creation cycle set by the user
   Deletion schedule	The method and cycle of automatically deleting snapshots created by the snapshot schedule as set by the user
   Attached volume	List of volumes connected to the snapshot schedule
   Created	The date the snapshot schedule was created - Hover the cursor for detailed units
   More	Execute item-specific functions - Modify snapshot schedule: Change the settings of the snapshot schedule - Attach volume: Connect a volume to the snapshot schedule - Detach snapshot schedule: Disconnect a volume connected to the snapshot schedule - Delete snapshot schedule: Delete the created snapshot schedule

Associate snapshot schedule

You can associate a snapshot schedule to a volume.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Snapshot Schedule menu, select Volume from the [More] icon > Create volume for the snapshot schedule.
3. Select the volume to attached to the snapshot schedule and click the [Save] button.

Disassociate snapshot schedule

You can disassociate a snapshot schedule from a volume.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Snapshot Schedule menu, select Detach volume from the [More] icon.
3. Select the volume to detach from the snapshot schedule and click the [Detach] button.

Delete snapshot schedule

You can delete snapshot schedules that are no longer in use.

1. Go to KakaoCloud Console > Beyond Compute Service > GPU.
2. In the Snapshot Schedule menu, select Delete snapshot schedule from the [More] button.
3. Enter the name of the snapshot schedule to delete and click the [Delete] button.

Restore data using snapshot

You can restore an instance's data using a volume created from a snapshot.

Restore data by creating new instance

This section explains how to create a new instance, attach a volume, and restore data.

info

• Data restoration using snapshots is useful when an instance fails to boot or is inaccessible via SSH or RDP due to incorrect configuration.
• This guide explains how to recover only backed-up data, not how to restore a root volume directly containing the OS.

Step 1. Create snapshot of instance requiring recovery (Optional)

💡 If a snapshot already exists, skip this step.

1. Stop the instance requiring recovery and wait until it reaches the Stopped state.
2. Refer to the Create snapshot to create a snapshot of the volume containing the data.

Step 2. Create volume from snapshot containing recovery data (Optional)

💡 If a volume already exists, skip this step. Refer to the Restore volume from snapshot for instructions.

Step 3. Create new instance for recovery

Refer to the Create instance guide to create a new instance with the same configuration, including instance type and volume size.

Step 4. Attach created volume to new instance

Refer to the Attach volume from console guide to attach the volume created in Step 2 to the newly created instance.

Step 5. Access instance and mount attached volume

Linux

1. Refer to the Connect to Linux instance guide to access the new Linux instance.

2. Verify the attached volume.

   Verify device

   lsblk
   Results

   NAME    MAJ:MIN RM  SIZE RO TYPE MOUNTPOINTS
   vda     253:0    0   10G  0 disk 
   ├─vda1  253:1    0    9G  0 part /
   ├─vda14 253:14   0    4M  0 part 
   ├─vda15 253:15   0  106M  0 part /boot/efi
   └─vda16 259:0    0  913M  0 part /boot
   vdb     253:16   0   10G  0 disk 
   ├─vdb1  253:17   0    9G  0 part 
   ├─vdb14 253:30   0    4M  0 part 
   ├─vdb15 253:31   0  106M  0 part 
   └─vdb16 259:1    0  913M  0 part 

3. Check the file system type of the attached volume.

   Use the file -s command to identify the file system type. Below are examples of two common file system types:

   Check file system type

   sudo file -s /dev/vdb
   Results

   (ext4)
   /dev/vdb: Linux rev 1.0 ext4 filesystem data, UUID=a69a34bb-e2fd-4e61-bf93-d462c4438730 (extents) (64bit) (large files) (huge files)
   
   (xfs)
   /dev/vdb: SGI XFS filesystem data (blksz 4096, inosz 512, v2 dirs)

4. Verify the integrity of the attached volume data. Execute the appropriate command based on the file system type and check the results.

   Verify data integrity (ext4)

   Verify data integrity (ext4)

   sudo fsck -n /dev/vdb

   Results

   fsck from util-linux 2.39.3
   e2fsck 1.47.0 (5-Feb-2023)
   /dev/vdb: clean, 11/655360 files, 66753/2621440 blocks

   Verify data integrity (xfs)

   Verify data integrity (xfs)

   sudo xfs_repair -n /dev/vdb

   Results

   Phase 1 - find and verify superblock...
   Phase 2 - using internal log
         - zero log...
         - scan filesystem freespace and inode maps...
         - found root inode chunk
   ...
   Phase 7 - verify link counts...
   No modify flag set, skipping filesystem flush and exiting.

5. Create a directory to mount the attached volume.

   Create directory

   sudo mkdir -p /mnt/restore

6. Mount the volume to the created directory with the appropriate file system type. Since this is for data recovery, mount it as read-only.

   Mount command (connect directory and volume, read-only) (ext4)

   sudo mount -o ro -t ext4 /dev/vdb /mnt/restore
   Mount command (connect directory and volume, read-only) (xfs)

   sudo mount -o ro -t xfs /dev/vdb /mnt/restore

7. Verify that the volume has been mounted correctly.

   Verify mount

   findmnt /mnt/restore
   Results

   TARGET       SOURCE   FSTYPE OPTIONS
   /mnt/restore /dev/vdb ext4   ro,relatime

Step 6. Restore necessary data from the mounted volume

1. Create a path for restoration.

   • In this guide, /path/to/restore is used as an example.
   Create restore path

    sudo mkdir -p /path/to/restore/

2. Copy the data to the restore path (/path/to/restore) using the rsync command.

   • Using the --checksum option may significantly increase the time required if there is a large volume of data.
   Copy data

   sudo rsync -avh --checksum /mnt/restore/ /path/to/restore/
   Results

   sending incremental file list
   backup-data-1
   backup-data-2
   
   sent 2.15G bytes  received 55 bytes  330.46M bytes/sec
   total size is 2.15G  speedup is 1.00

3. Verify that the copied data exists in the restore path.

   Verify data

   ls -alh /path/to/restore/
   Results

   total 2.1G
   drwxr-xr-x 3 root root 4.0K Dec 10 07:24 .
   drwxr-xr-x 3 root root 4.0K Dec 10 07:55 ..
   -rw-r--r-- 1 root root 1.0G Dec 10 07:57 backup-data-1
   -rw-r--r-- 1 root root 1.0G Dec 10 07:57 backup-data-2

Step 7. Detach the connected volume

1. Unmount the volume after data copying is complete.

   Unmount volume

   sudo umount /mnt/restore

2. Follow the steps in the Detach volume guide to detach the recovered volume from the console.

caution

Ensure the recovered volume is properly detached from the console. Leaving it connected may result in data corruption, resource waste, or system crashes.

Windows

1. Refer to the Connect to Windows instance guide to connect to the Windows instance with the attached volume.

2. Verify the volume disk:

   1. Press Start > Run or use the shortcut WIN + R.
   2. Enter diskmgmt.msc and click OK or press Enter.
   3. In Disk Management, verify the attached volume disk.
   4. The recovery volume will have a partition structure similar to the OS volume mounted on the C drive.

   

3. If the attached volume is detected, set its status to Online.

   

4. Verify the mount point of the attached volume.

   • Assume the data to be restored is located in the system's primary partition.
   • Check which path the primary partition is mounted to.
   • In the image below, it is mounted to the (E:) path.

   

5. Open File Explorer to check the mounted drive.
   

Step 6. Restore necessary data from the mounted volume

1. Use File Explorer to navigate between the source and target directories.
2. Restore the data using Copy-Paste or Drag & Drop.

Step 7. Detach the connected volume

1. Change the connected disk status to Offline.
   

2. Follow the steps in the Detach volume guide to detach the recovered volume from the console.

Restore data on an active instance

This section explains how to attach a volume to an existing active instance and restore data.

info

Data restoration using snapshots is useful when an instance fails to boot or is inaccessible via SSH or RDP due to incorrect configuration.

Step 1. Create a volume from a snapshot containing recovery data

Refer to the Restore volume from snapshot guide to create a volume from the snapshot containing the data to be restored.

Step 2. Attach the volume to the instance requiring data recovery

Refer to the Attach volume from console guide to attach the volume created from the snapshot to the instance requiring data recovery.

Linux

Step 3. Access the instance and mount the attached volume

1. Refer to the Connect to Linux instance guide to connect to the Linux instance with the attached volume.

2. Verify the attached volume.

   Verify device attachment

   lsblk
   Results

   NAME    MAJ:MIN RM  SIZE RO TYPE MOUNTPOINTS
   vda     253:0    0   10G  0 disk
   ├─vda1  253:1    0    9G  0 part /
   ├─vda14 253:14   0    4M  0 part
   ├─vda15 253:15   0  106M  0 part /boot/efi
   └─vda16 259:0    0  913M  0 part /boot
   vdb     253:16   0   10G  0 disk

3. Check the file system type of the attached volume.

   Check file system type (ext4)

   Check file system type (ext4)

   sudo file -s /dev/vdb

   Results

   /dev/vdb: Linux rev 1.0 ext4 filesystem data, UUID=a69a34bb-e2fd-4e61-bf93-d462c4438730 (extents) (64bit) (large files) (huge files)

   Check file system type (xfs)

   Check file system type (xfs)

   sudo file -s /dev/vdb

   Results

   /dev/vdb: SGI XFS filesystem data (blksz 4096, inosz 512, v2 dirs)

4. Verify the integrity of the attached volume data. Execute the appropriate command based on the file system type and check the results.

   Verify data integrity (ext4)

   Verify data integrity (ext4)

   sudo fsck -n /dev/vdb

   Results(ext4)

   fsck from util-linux 2.39.3
   e2fsck 1.47.0 (5-Feb-2023)
   /dev/vdb: clean, 11/655360 files, 66753/2621440 blocks

   Verify data integrity(xfs)

   Verify data integrity (xfs)

   sudo xfs_repair -n /dev/vdb

   Results

      Phase 1 - find and verify superblock...
      Phase 2 - using internal log
            - zero log...
            - scan filesystem freespace and inode maps...
            - found root inode chunk
      ...
      Phase 7 - verify link counts...
      No modify flag set, skipping filesystem flush and exiting.

5. Create a directory to mount the attached volume.

   Create directory

   sudo mkdir -p /mnt/restore

6. Mount the volume to the created directory with the appropriate file system type. Since this is for data recovery, mount it as read-only.

   Mount command (connect directory and volume, read-only) (ext4)

   sudo mount -o ro -t ext4 /dev/vdb /mnt/restore
   Mount command (connect directory and volume, read-only) (xfs)

   sudo mount -o ro -t xfs /dev/vdb /mnt/restore

7. Verify that the volume has been mounted correctly.

   Verify mount

   findmnt /mnt/restore
   Results

   TARGET       SOURCE   FSTYPE OPTIONS
   /mnt/restore /dev/vdb ext4   ro,relatime

Step 4. Restore necessary data from the mounted volume

1. Use the rsync command to copy data to the restore path (/path/to/restore).

   • The --checksum option may significantly increase the time required if there is a large volume of data. Use it as needed.
   Copy data

   sudo rsync -avh --checksum /mnt/restore/ /path/to/restore/
   Results

   sending incremental file list
   backup-data-1
   backup-data-2
   
   sent 2.15G bytes  received 55 bytes  330.46M bytes/sec
   total size is 2.15G  speedup is 1.00

2. Verify that the copied data exists in the restore path.

   Verify data

   ls -alh /path/to/restore/
   Results

   total 2.1G
   drwxr-xr-x 3 root root 4.0K Dec 10 07:24 .
   drwxr-xr-x 3 root root 4.0K Dec 10 07:55 ..
   -rw-r--r-- 1 root root 1.0G Dec 10 07:57 backup-data-1
   -rw-r--r-- 1 root root 1.0G Dec 10 07:57 backup-data-2

Step 5. Detach the connected volume

1. Unmount the volume after data copying is complete.

   Unmount volume

   sudo umount /mnt/restore

2. Follow the steps in the Detach volume guide to detach the recovered volume from the console.

Windows

Step 3. Change the attached volume status to online

1. Refer to the Connect to Windows instance guide to connect to the Windows instance with the attached volume.

2. Verify the volume disk:

   1. Press Start > Run or use the shortcut WIN + R.
   2. Enter diskmgmt.msc and click OK or press Enter.
   3. In Disk Management, verify the attached volume disk.

3. If the attached volume is detected, set its status to Online.

4. Verify the mount point of the attached volume. In the image below, the volume is shown as mounted to the (D:) path.

5. Open File Explorer to confirm the mounted drive.

Step 4. Restore necessary data from the mounted volume

1. Use File Explorer to navigate between the source and target directories.

2. Restore the data using Copy-Paste or Drag & Drop.

Step 5. Detach the connected volume

1. Change the connected disk status to Offline.

2. Follow the steps in the Detach volume guide to detach the recovered volume from the console.

caution

Ensure the recovered volume is properly detached from the console. Leaving it connected may result in data corruption, resource waste, or system crashes.