This section explains how to back up and restore L-Servers.
These functions are only provided for physical L-Servers.
Backup and restore of L-Servers is the function for performing backup and restoration of L-Server system images.
If multiple backups of the same L-Server are collected, the backups are managed by automatically adding version numbers to the backup file name starting from 1. If the number of snapshot versions exceeds the limit (standard limit is 3), the oldest snapshot will be deleted.
The content of the system disk of the L-Server is restored to the state when the system image was collected.
Note
Regardless of the boot environment (local/SAN/iSCSI) and RAID configurations, only the contents of the first disk (boot disk) recognized by the managed server's BIOS can be backed up and restored.
For a server using local boot with a SAN data configuration, it is necessary to configure the target disk of image operations.
For details, refer to "9.1.13 Changing Target Disks of Image Operations" in the "User's Guide VE".
The contents of other disks (data disks) cannot be backed up and restored. To properly backup and restore such data disks, it is recommended to use dedicated backup software, or the copy functions available in storage devices.
When the first disk contains multiple partitions (Windows drive, Linux/VMware partition), all partitions are backed up.
Disk | Windows Disk Name | Windows Drive Name | Target of Backup and Restore |
---|---|---|---|
First | Disk 0 | C: | Yes |
E: | Yes | ||
Second | Disk 1 | D: | No |
F: | No |
As managed servers are restarted during backup and restore operations, their applications should be stopped beforehand.
Restore operations can only be performed for the servers from which a backup has been collected.
When backups have been collected from a managed server with UMC enabled, perform the restore operation on the server with UMC enabled.
The first partition must be a primary partition.
Backups can be collected with the following file systems. Note that LVM (Logical Volume Manager) partitions are not supported.
NTFS
EXT3
EXT4
LinuxSwap
Point
When Red Hat Enterprise Linux 7 is installed with default options, the file system is XFS.
The operations for backup and restore of VM hosts differ depending on the server virtualization software used.
For an explanation of the behavior differences that occur when VM guests are included in the VM host's boot disk, refer to "9.2.2 Functional Differences between Products" in the "Design Guide VE".
If VM guests on the boot disk are not to be backed up (and restored), VM guest files should be moved to another disk.
To preserve the configuration of the server virtualization software used, VM guests should be backed up at the same time as VM hosts. During backup, because the target VM host will be automatically set to VM maintenance mode, the VM host should be in a state that allows VM maintenance mode to be set.
When backing up a VM host in a high-availability configuration, all VM guests stored on shared disks should be migrated to another VM host beforehand.
After backing up the VM host, migrate the VM guests back to their original VM host.
Refer to the server virtualization software manual and "9.2.2 Functional Differences between Products" in the "Design Guide VE" for information on how to back up and migrate VM guests, or about the VM maintenance mode.
To preserve the configuration of the server virtualization software used, VM guests backed up at the time of the VM host's backup should also be restored when restoring a VM host. Note that this is not required if no changes likely to alter the virtualization software configuration were made (e.g. changes such as addition or deletion of a VM guest, or changing the placeholder for VM guest definition files).
During restore, because the target VM host will be automatically set to VM maintenance mode, the VM host should be in a state that allows VM maintenance mode to be set.
If the target VM host is in a high-availability configuration, all VM guests stored on shared disks should be migrated to another VM host beforehand.
After restoring the VM host, migrate the VM guests back to their original VM host.
Refer to the server virtualization software manual and "9.2.2 Functional Differences between Products" in the "Design Guide VE" for information on how to restore and migrate VM guests, or about the VM maintenance mode.
Deleting a managed server will delete its backed up system images at the same time.
It is not possible to backup, restore, or delete a system image from a managed server for which a system image (including different versions) is already being backed up, restored, or deleted.
When restoring a system image on a server whose name was changed after the deployment of a cloning image, check that the "server_name" displayed on the server resource tree and the System Image List match the new server name before restoring the system image.
For managed servers on which the Watchdog function is enabled, backup or restore operations on that server may be aborted by an automatic restart or shutdown. The Watchdog is a function which automatically restarts or shuts down non-responsive servers when their operating system does not respond for a given period.
Therefore, it is highly recommended to disable the Watchdog function before a backup or restore operation.
For details on the Watchdog function, refer to the server manual.
If the disk size of the source (backed up) server differs from that of the destination (restored) server, restore is possible only in cases where the disk size of the destination server is larger than that of the source server.
In that case, unused disk space will remain on the destination server. To use this unused disk space, a partition should first be created from it.
Restoring a system image to a server on which the disk size is smaller than that of the source (backed up) server is not possible. This also applies to server switchover and failback operations that are based on backup and restore, as well as cloning operations.
Therefore, it is also necessary to ensure that spare servers of cloning destination servers have a large enough disk.
When backing up or restoring system images, or collecting and deploying cloning images, up to four processes can be executed simultaneously. If four processes are already being executed, any additional image operations will enter a standby state.
Moreover, server switchover which is executed using the backup and restore method or any restore process performed during failback will also enter a standby state. When Auto-Recovery or manual switchover using the backup and restore method is used, limit the number of image operations (backup or restore of system images, or collection and deployment of cloning images) performed at the same time to 3.
When using backup and restore for PRIMEQUEST servers, check that the boot option settings and BIOS settings of the target servers match. When the settings are different, execute the function after changing the settings so they match. For details on how to change boot options, refer to "7.1.10 Changing Boot Options".
When using PRIMEQUEST 1000 Type2, disable the x2APIC mode of the UEFI.
For the PRIMEQUEST 2000 series, backup and restore is only possible when using Windows managers. PXE boot is only supported by on-board LAN NICs.
When the target managed server is Linux and the disk is recognized using the by-id name, the collected system image cannot be restored after changing disks.
Not supported when using clustering software on physical L-Servers.
To back up and restore L-Server system images, all of the following conditions must be met:
For Backup
The status of the L-Server is that in which resources have already been allocated (allocated)
The L-Server has been started
The L-Server has been registered as an agent
For Restoration
The status of the L-Server is that in which resources have already been allocated (allocated)
Use the following procedure for backup and restoration of system images:
Backup
Right-click the target L-Server in the orchestration tree, and select [Backup/Restore]-[Backup] from the popup menu.
The [Backup System Image] dialog is displayed.
Set the following items as necessary:
Enter a comment to identify the system image.
Up to 128 characters other than percent signs ("%"), back slashes ("\"), double quotes ("), and line feed characters can be specified.
Click the [OK] button.
The system image is backed up.
L-Servers are temporarily powered off while backing up the system image.
The process status can be checked in the Recent Operations area of the ROR console. When the [Cancel] button is clicked in the Recent Operations area, the confirmation dialog to quit the process is displayed.
From the command-line, execute rcxadm image backup.
For details on the rcxadm image backup command, refer to "4.1 rcxadm image" in the "Reference Guide (Command/XML) CE".
Restoration
Right-click the target L-Server in the orchestration tree, and select [Backup/Restore]-[Restore] from the popup menu.
The [Restore System Image] dialog is displayed.
Select the version of the system image to restore.
Confirm the content of the alert message displayed in the [Restore System Image] dialog, and check the [Restore the system image.] check box.
Click the [OK] button.
The system image is restored.
The process status can be checked in the Recent Operations area of the ROR console. When the [Cancel] button is clicked in the Recent Operations area, the confirmation dialog to quit the process is displayed.
From the command-line, execute rcxadm image restore.
For details on the rcxadm image restore command, refer to "4.1 rcxadm image" in the "Reference Guide (Command/XML) CE".
Deletion
This section explains how to delete unnecessary versions of backups of system images.
Right-click the target L-Server in the orchestration tree, and select [Backup/Restore]-[Delete a Backup] from the popup menu.
The [Delete a System Image] dialog is displayed.
Select the version of the system image to delete.
Click the [OK] button.
The backup of the system image is deleted.
From the command-line, execute rcxadm image delete.
For details on the rcxadm image delete command, refer to "4.1 rcxadm image" in the "Reference Guide (Command/XML) CE".