Rapid Return to Production
(RRP) creates a full production VM from an existing VM created with Instant Virtualization from Block Backups or Instant VM Restore from Agentless Backups. RRP transforms a temporary VM to a working production VM without significant downtime (typically only a few minutes). The new VM is available for use while DPX transfers data behind the scenes, using Storage vMotion, to populate the selected datastore. Secondary storage is automatically released after Storage vMotion transfer completes.
For details, see:
Requirements and Considerations for RRP
Before using RRP the first time, note requirements and considerations. Clients that meet the requirements are eligible for RRP processing and will be listed for selection in the RRP wizard. Requirements and considerations are described in the following topics:
“RRP Requirements”
“RRP General Considerations”
“Power State Considerations for RRP”
“Completing an RRP Job Despite Master Server Interruption”
RRP Requirements
Note: RRP requirements apply to both RRP jobs and Full with RRP jobs. In the latter case, RRP requirements must be met since RRP is used transparently in the background.
|
•
|
A VM created with Instant Virtualization or Agentless Instant VM Restore must already exist. |
|
•
|
For Windows Full Virtualization with RRP (Full with RRP), the radio button Align virtual machine file system on the Virtual Machine Information screen of the virtualization wizard, must be selected. For Linux or Clustered Data ONTAP volumes less than 2 TB in size, the file system will always be aligned and the radio button does not appear. |
Warning: DPX does not support Full Virtualization with RRP from Clustered Data ONTAP if the backup volume was larger than 2 TB. These backups must be restored using , see “Bare Metal Recovery Overview”.
RRP General Considerations
|
•
|
RRP to a vApp is not supported. |
|
•
|
RRP of a VM within a vApp is not supported. |
|
•
|
RDM disks are converted to virtual disks. |
|
•
|
Running an RRP job to an NFS target datastore causes the machine to power off when the job completes. This is the outcome of using a cloning operation, as migration to an NFS target datastore is not supported. This is due to a VMware limitation. |
|
•
|
If the Instant Virtualization VM to be used for RRP (or Full with RRP) was created with a version of DPX that does not support RRP, then RDM compatibility mode will be automatically converted from physical to virtual compatibility mode, which is required by Storage vMotion. |
|
•
|
If a RRP job fails, you cannot define and run another RRP job that uses the same Instant Virtualization VM. If you attempt to do so, a message will indicate that processing has failed because the VM is involved in an active or incomplete RRP job. Re-run the job. |
Power State Considerations for RRP
Note the following considerations related to the power state of the VM used for RRP jobs:
|
•
|
While a RRP job or Full with RRP job is in progress, you cannot change the on/off state of the Instant Virtualization VM in vSphere Client. You must wait for the job to complete. The RRP job may seem to be complete because the new production VM is available almost immediately; however, data transfer is occurring in the background, and the job is not actually complete until this data transfer completes (which can take hours, or even days). |
|
•
|
Once the RRP job starts, don’t attempt to change the power state of the Instant Virtualization VM. If that VM is powered on and you power it off from the VM itself, the RRP job will fail. If the Instant Virtualization VM is powered off, you cannot turn it on until the RRP job completes. The power on state will be grayed out in VMware. |
|
•
|
RRP using Storage vMotion requires the Instant Access VM in the power on state. If the Instant Access VM is off before the RRP job starts, RRP will proceed, but will use the cloning method. In this scenario, the rapid creation of a usable production machine will not be realized. |
|
•
|
If you don’t select the Start VM after it is created check box in the virtualization wizard, that is equivalent to starting a Full with RRP job with the Instant Virtualization VM powered off, and the RRP job will use the cloning method rather than the migration (Storage vMotion) method. |
Completing an RRP Job Despite Master Server Interruption
In the event the master server experiences an interruption due to network or hardware issues, the RRP job will continue. Thus, it is not necessary to restart the RRP job from the beginning (which could lose many hours, or even days, of processing time). However, you do need to take some simple steps to ensure the RRP job completes successfully.
In the event of a master server interruption during a RRP job, the will indicate a status of Abort. However, the master server interruption only affects reporting of job progress; the RRP job continues.
To ensure the RRP job completes successfully, note the following:
|
•
|
To resolve the Abort condition and update job monitoring, re-run the RRP job. Note that this is not actually a job re-run. The re-run simply sync up the master server with the in-progress RRP job, and the job monitor will now indicate actual progress of the job. |
|
•
|
If the RRP job has completed before the job re-run, the re-run will initiate the RRP post-completion phase, LUN cleanup. Although the RRP job continues while Abort is indicated, it will not proceed to LUN cleanup until the job is re-run to resolve the Abort condition. |
|
•
|
If you want to cancel the RRP job for which Abort is indicated, you must re-run the job, then cancel. |
|
•
|
While Abort is indicated, you can not start another RRP job based on the same Instant Virtualization VM. You must resolve the Abort condition first by re-running the job and either allowing it to complete or canceling it. |
Using RRP
Rapid Return to Production (RRP) creates a full production VM from an existing VM created with Instant Virtualization from Block Backups or Instant VM Restore from Agentless Backups. RRP transfers data from a snapshot on secondary storage to primary storage attached to an ESX server. The new production VM is available for use within minutes while data transfer continues transparently in the background. For RRP requirements, see “Requirements and Considerations for RRP”.RRP jobs are defined through the RRP Wizard. RRP jobs can be scheduled, but they are intended for single, not recurring use.
To define an RRP job:
|
1.
|
On the Function Tab Bar, click Restore, then click Block. |
|
2.
|
On the Task Panel, select RRP Wizard; the Welcome dialog appears. |
|
3.
|
Click Next; the Login dialog appears. |
|
4.
|
Log in to the host or vCenter Server for the new VM: |
|
•
|
Enter IP Address / Name. This is a fully qualified network name or IP address of an ESX Server or vCenter Server. |
|
•
|
Enter a valid user name and password. |
|
5.
|
Click Next; the Select Virtual Machine dialog appears. |
|
6.
|
Expand the datacenter to display VMs. |
|
7.
|
Click on a VM for RRP to bring into production. Only machines eligible for RRP will be displayed. The properties of the VM are displayed in the right pane. |
|
8.
|
Click Next; the Select Datastore dialog appears. |
|
9.
|
Select a datastore from which the VM disks will be provisioned. Note the following restrictions: |
|
•
|
Free space must be available on the datastore to accommodate the new VM. |
|
•
|
The datastore type is indicated in the Type column. RRP supports VMFS and NFS/CIFS datastores. |
|
10.
|
Click Next; the Specify Job Information dialog appears. |
|
11.
|
Identify the RRP job: |
|
a.
|
Name the job, specify a folder for the job, and optionally add a comment. |
|
b.
|
Optionally select Delete Job when Done. If checked, your job definition will be deleted once the job completes successfully. It will not be deleted if the job fails. In most cases, you should leave this check box unselected since you will usually want to re-run a job. If you select the check box, the values you entered for it will be lost after the job runs. |
|
c.
|
Optionally schedule the job to run later. If you click Schedule, the job opens. If the job is not scheduled, after you click Finish on the final screen of the wizard, you will be asked if you want to run the job immediately. |
|
12.
|
The Summary screen displays the information you have entered for the RRP job. To change an entry, click Back. To accept all selections, click Finish. |
|
13.
|
After you click Finish, if you have not scheduled the job, you will be asked if you want to run the job immediately. If the job is not scheduled and you don’t choose to run it immediately, you can run the job later from the Monitor Jobs window. |