5 1 0:VMware Sandbox Restore
Are you looking for an alternative for your VMware environment?
SEP supports a range of hypervisors, offering snapshot-based online backup of VMs. Click HERE to get an overview of the available SEP modules and explore supported features like CBT, Single File Restore, and more. |
Overview
Note that the VMware sandbox restore procedure in SEP sesam is version specific. For previous documentation versions, see VMware Sandbox Restore in v. Beefalo/Beefalo V2.
SEP sesam supports online VMware sandbox restore via Restore Assistant (if you have the appropriate permissions, see Restore Assistant) to run tests in an isolated virtual environment without network access to production. You can use the SEP sesam sandbox for VMware as a test environment where you can restore and start one or more virtual machines (VMs), troubleshoot VMs, install new software, updates and patches, verify VM backups, etc.
For example, if you want to verify that the restores work, set up one task to verify the restore of Active Directory (AD) and another to verify the restore of Exchange.
Note that in previous SEP sesam versions, you can perform web VMware sandbox restore in basic or advanced restore mode. Now it's only available in advanced restore mode, but you can easily switch between basic and advanced mode by selecting the Advanced View option, see Setting UI mode.
Prerequisites
- Ensure that you have sufficient restore permissions and access rights.
- Make sure that you use a different network for the sandbox restore to avoid IP address conflicts, as the original VM is usually still active.
Accessing the restore assistant
You can access the Restore Assistant in one of the following ways:
- via the GUI: by clicking the Restore Assistant icon in the toolbar or from Activities -> Restore Assistant
- from SEP sesam Web UI: left menu -> Restore Assistant
- or by entering the following address in the browser bar: http://[sesamserver]:11401/sep/ui/restore/.
Note | |
|
Performing VMware sandbox restore
You can perform VMware sandbox restore in the web interface in the advanced UI mode if you have the appropriate permissions.
- Open the Restore Assistant in the browser.
- In the Start window, select VMware vSphere. Then select the Advanced View option in the lower left corner and Sandbox restore.
- In the Virtual Machine window, select the relevant vSphere server and the virtual machine. You can filter VMs by name, location or OS. Click Next.
- In the Task window, under Task selection, select your source task. A backup task defines the source data that was backed up by the client.
Then under Backup selection, select the exact backup version you want to restore. You can use the calendar function in the upper right corner to set a date range for the displayed backups.Note The Generation restore option is enabled by default. When selecting incremental or differential backups, it can be disabled if you only want to restore the selected backup and not the entire backup chain. - Select the restore target: vSphere server, Datacenter, and Virtual machine, and then select the appropriate execution option: Do not overwrite an existing virtual machine, Restore an existing virtual machine with a new name, or Overwrite an existing virtual machine. In the latter case, you can also select to automatically shut down the running VM. Click Next.
- In the Options window, select the following options:
- Under Virtualization restore options, select the target data mover from the drop-down list.
- Under Target options of the virtual machine, select the ESX server, the Data store, and Network interface(s) you want to use for your sandbox restore from the drop-down lists. Optionally, select the target Folder and/or Resource pool/vApp.
- Under Recovery options – Sandbox checks, you can select the checks and actions to perform. Clicking the Edit button opens the list of Activated recovery options.
Use this list to define additional options or to modify and rearrange the existing options:- As you hover over the command, additional icons appear that you can click to customize the list of recovery options.
- Click the New button in the left column to create a new argument. Type the name of your command or click the Actions,Checks or Run buttons to open the drop-down list and select the appropriate option or check. The drop-down lists Actions, Checks and Run contain an extensive list of commands and options that can be used for various sandbox restore scenarios.
For example, after the start action, enter a new action sleep by selecting sleep from the Actions drop-down list: Sleep N seconds and then define the desired VM uptime in seconds. For example, with sleep=300 a VM can be checked for 5 minutes before it is stopped again (actively running on a host). - Use the move up/move down arrows to change the order of the options/checks in the list.
- Use the recycle bin button to remove the currently selected option/check. The option/check is removed from the Activated recovery options list. For the list of available VM sandbox actions for restore refer to Available sandbox actions.
- Under the Optional data source selection, you can select your preferred media pool, drive, used media|barcode, and interface from the drop-down lists.
- An additional set of options is available under the Advanced restore options:
- Use the Log verbosity level drop-down list to change the restore log level for your specific client, see Setting Log Level.
- Click the Edit button under the Options for restore to specify the transport hierarchy by adding transport modes to the Selected transport modes list according to your preferences; for details on transport modes, see Selecting the best VMware transport mode for your environment.
- Use the Restore server options to specify additional options for the SEP sesam Server, for example, sm_restore -v 1 starts a restore with a higher log level.
- You can specify the retention period for the restore (how long (in days) to keep the restore task). If you want to use this sandbox task as a template, set the retention period to 0 (infinity) so that the task is not deleted.
- If you want a pre- or post script to be applied to the restore task, select the check box Execute pre or Execute post; for details, see Pre/Post options.
- In the last step, check the summary of your restore task (restore type (based on task type, selected backup, its date and details, restore options, etc.) and click Start restore.
Available sandbox actions
SEP sesam supports the following actions that can be used mainly for the sandbox restore. Note that some of the actions can be used for VM backup. For additional actions that are available for VMware backup see also Available backup actions.
Check
The check action triggers a test if a virtual machine property matches the specified condition. The action will either succeed or log the cause of the failure. The option onError defines how the failure affects the whole sandbox restore. The check action can appear multiple times within the recovery action sequence.
The general action syntax is:
check=<property><operator><value>[<options>]
where:
- <property> is any named property or tag of the virtual machine (for example, powerState or vmToolsState).
- <operator> is a comparison operator. Available operators are "=", "<>", ">", ">=", "<", "<=" or "%=" (like). The "like" operator should allow the same placeholder as the DB select statement like operator.
- <value> is the value of the named property or tag.
- <options> is a list of options, separated by "#" (hash), customizing the check action behavior. Available options are:
- wait: If the current property value does not match the specified condition, the action will wait for the virtual machine to change the property until it matches the specified condition or a default time out is reached. The default time out is 120 seconds.
- timeout: Sets the timeout to the specified value in seconds.
- onError: Specifies the behavior when the check action fails. Valid values are break, abort, or the absolute path of a script to run on the SEP sesam Server or RDS where the VMWare backup client (sbc_vadp) is executed. break will stop the execution of the sandbox restore and perform a full rollback. abort will stop the execution of the sandbox restore and leave the virtual machine in its current state for further analysis. Note that you should clean up the virtual machine when the analysis is finished.
Note that none of parts are allowed to contain a "." (a dot) or "," (a comma). A dot is the separator for the action in the specified recovery action sequence, and a comma is the separator of the parent parameter -a. Dots have to be replaced with "~" (a tilde).
- Examples
check=powerState=on
- Checks the state of the virtual machine. The following values are available: on, online, poweredOn, running, off, offline, poweredOff, stop, stopped, halted, suspended.
check=powerState=on[wait]
- Checks if the virtual machine is powered on. Wait up to 120 seconds for the virtual machine to get powered on (if offline).
check=powerState=on[wait#timeout=30]
- Checks if the virtual machine is powered on. Wait up to 30 seconds for the virtual machine to get powered on (if offline).
check=vmToolsState=installed
- Checks the state of the VMware guest tools on the virtual machine. The following values are available: installed, running, guestToolsRunning, stopped, notRunning, guestToolsNotRunning, executing, guestToolsExecutingScripts.
check=ip=*
- Checks if any IP address (IPv4 and IPv6) is set for the virtual machine.
check=ip=0~0~0~0/0
- Checks if any IPv4 address is set for the virtual machine.
check=ip=::/0
- Checks if any IPv6 address is set for the virtual machine.
Run
The run action starts the given command on the running virtual machine using the VMWare guest tools. The precondition to use the run action is that the virtual machine boots and has the VMWare guest tools installed and running. The action will either succeed or log the cause of the failure. The option onError defines how the failure affects the whole sandbox restore. The run action can appear multiple times within the recovery action sequence.
The general action syntax is:
run=<command>[<options>]
where:
- <command> is the command to be executed on the virtual machine. It is recommended to use URL encoded form.
- <options> is a list of options, separated by "#" (hash), customizing the run action behavior. Available options are:
- wait: This is the default option for the run action and will force the action to wait untill the command finishes executing on the virtual machine. The output of the executed command is logged in the backup or restore log. The default timeout is set to 0 (wait indefinitely).
- noWait: If this option is specified, the command will start on the virtual machine, but it will not wait for the command to finish nor collect any output.
- timeout: Sets the timeout to the specified value in seconds. If the command does not finish executing within the specified timeout, the run action will no longer wait for the command to finish nor fetch any command output. The command itself will continue to run inside the virtual machine until finished.
- onError: Specifies the behavior when the check action fails. Valid values are break, abort, or the absolute path of a script to run on the SEP sesam Server or RDS where the VMWare backup client (sbc_vadp) is executed. break will stop the execution of the sandbox restore and perform a full rollback. abort will stop the execution of the sandbox restore and leave the virtual machine in its current state for further analysis. Note that you should clean up the virtual machine when the analysis is finished.
- os_username: The name of the user used for executing the command. If omitted, SEP sesam will attempt to get the username from the client or backup/restore task.
- os_password: The password to authenticate the user used for executing the command. If omitted, SEP sesam will attempt to get the password from the client or backup/restore task. The password can be entered in plain text or encoded. If the password is encoded, use curly braces (os_password={<encoded password}).
Recover
The action parameters manage all action steps, done during backup. The actions can be combined with the "." (dot) as the separator between the steps.
The general syntax is:
recover=<action1>.<action2>.<action3>...
For example:
recover=rest.conf.vmdk.data
The table below contains all available Recover actions.
Action | Syntax | Description |
---|---|---|
start | -a recover=start | Start the VM |
stop | -a recover=stop | Stop the VM |
remove | -a recover=remove | Remove the VM |
rmiv | -a recover=rmiv | Remove the VM from inventory of vSphere server, can be reimported later with vCenter client |
rmdi | -a recover=rmdi | Remove all disks from VM |
delsnapall | -a recover=delsnapall | Remove all SEP sesam snapshots of VM |
rest | -a recover=rest | Restore config data, which includes meta data and is required for restore |
conf | -a recover=conf | Create blank VM without VMDKs |
vmdk | -a recover=vmdk | Add VMDK to VM (VM must already exist) |
addvmdk | -a recover=addvmdk | Add further custom VMDK to VM |
data | -a recover=data | Restore VMDK data of VM |
attach | -a recover=attach | Attach VMDK to VM |
attach-ro-nfs | -a recover=attach-ro-nfs | Add NFS datastore in "read-only" mode |
snap | -a recover=snap | Create a VM SEP sesam snapshot |
delsnap | -a recover=delsnap | Remove VM SEP sesam snapshot |
consolidate | -a recover=consolidate | Remove VM snapshot with consolidation |
reloc | -a recover=reloc | Relocate VM |
detach | -a recover=detach | Detach datastore from the host system |
detach_vmdk | -a recover=detach_vmdk | Detach VMDK from VM |
remove_nfs_store | -a recover=remove_nfs_store | Remove NFS datastore |
sleep | -a recover=sleep=10 | Specify wait time in seconds |
exec | -a recover=exec=cleancmd | Run external command, which must be located in sepsesam/bin/sesam
|
run | -a recover=run=clean | Run command directly on VM |
mac | -a recover=mac=[new|orig|<specific MAC Address>] | Will be available in Apollon V2 Handling of the MAC address: generate a new MAC address, set the original MAC address, or apply a specific MAC address provided during the setup. When specifying an explicit MAC address, you can use either no delimiter (e.g., 001B638445E6) or the '-' delimiter (e.g., 00-1B-63-84-45-E6). If the 'mac' option is not specified, the original MAC address is set when the override mode is enabled. |
Monitoring restore
You can monitor the restore progress using the Web UI (Monitoring -> Restores) or view the status in the GUI (Main Selection -> Job State -> Restores). The restore overview provides detailed information on the last run of restore jobs, including task name, status (successful, error, in queue...), start and stop time of the last backup, data size, throughput, etc. For details, see SEP sesam Web UI or Restores by State in the GUI.
See also
Restore Assistant – VMware Restore – VMware Single File Restore