Instant Restore for ZStack
Instant VM Restore can be used in emergency situations to recover a VM within 1 min, minimizes the downtime of critical businesses. It can be very helpful in emergency situations to directly start a VM from its backup data, without the need to transfer backup data back to production storage then resume the VM.
Create Instant VM Restore Job
To create a ZStack Instant Restore job, please go to VM Backup > Restore > Instant Restore page, select a target VM restore point from ZStack. You can choose from All nodes or specific node to locate the restore points.
Notice
The restore point to be used for instant restore can be a full backup restore point, an incremental restore point or a differential restore point, but you can only select one restore point for each instant restore job.
Select a host from ZStack as the restore destination where you want to run the instant restore VM, and select the backup node IP/domain where the backup storage is mounted.
When a host is selected, in the VM configurations section it will show loading state, it will take a few seconds to request for the information of virtual platform resources. Under General tab, you are able to set VM name, power state, CPU and RAM size for the instant restore VM.
Under Virtual Disk tab, the virtual disk settings are view-only, because instant restore will use VM backup data to run the VM on virtual platform, no data transmission to virtual platform datastore will be involved at this stage, so virtual disk settings are temporarily not needed.
Virtual Network: It allows you to select the virtual network to be connected to and the MAC address assignment of the restored VM.
In the Connect to column, you can select a desired virtual network for specific virtual network interface of the VM, by default it will automatically select one from the available virtual networks. By clicking on Advanced, you can setup the MAC address assignment for the virtual network interface.
By default, the virtual platform will auto generate a new MAC address for the VM, but you can also use the original MAC address or customize the VM MAC address if you prefer.
Other Settings: Currently, Other Settings option will not present to all VM restore job settings. This option will present for certain virtualizations for some additional configurations of the VM restore job.
If the VM backup data is encrypted (Data Encryption enabled in backup job), when restoring the VM you need to provide the data encryption password for verification under Other Settings tab, otherwise without the data encryption password, the VM cannot be restored. Once done, click on OK button to submit the creation of the instant restore job.
Run Instant Restore Job
After creating a new instant restore job, you will be redirected to the Monitor Center > Jobs page, and you’ll be able to see the instant restore job you had created in the job list, it will be in Pending state.
To start the instant restore job, please click on Options, and then select Start Job. The job status will change to Starting, it will take a while (usually a few seconds to a minute depending on the backup storage performance) for the restore process to be completed, when the job status change to Running the VM is ready to use.
If you configured the VM power state preset as On, after the instant restore process is done, the VM will power on automatically and you'll get the services of the VM back online.
But if the power state preset is Off, you need to manually turn on the VM after instant restore process is done. You should not turn on the instant restore VM until the instant restore job status changed into Running state, otherwise the VM might become abnormal and the instant restore will fail.
During the instant restore VM runtime, you'll find that in the ZStack storage list, there's a new NFS datastore, and the storage server should be Vinchin Backup & Recovery, please do not unmount the NFS datastore, otherwise the instant restore VM will become offline. When the instant restore job is terminated, Vinchin Backup & Recovery will automatically unmount the NFS storage.
Notice
All the data of the instant restore VM is stored on Vinchin backup repository which is mounted to the selected ZStack host as NFS datastore, if you terminate the instant restore job, all the data of the instant restore VM will be deleted from the ZStack host (including newly generated data during the runtime of instant restore VM). If you wish to keep the instant restore VM and its newly generated data, do not stop the job until you have migrated all data to the ZStack production datastore.
Do not create snapshot on the instant restore VM, or change any virtual disk settings. Otherwise, the instant restore VM might crash and newly generated data might lose.
Create Live Migration Job
Live migration is a feature based on instant VM restore job, the migration operation is recommended to be done during the non-business hours.
To perform live migration of an instant restore VM, please go to Monitor Center > Jobs page. Click on Options of an instant restore job and then select Migration, on the New Job screen complete the following settings of a live migration job.
Target Host: Please select a ZStack (could be another virtualization) host on which you wish the VM to be migrated to.
Notice
When creating an instant restore job to restore ZStack VMs, the target virtualization can be ZStack or another virtualization (if your Vinchin Backup & Recovery has V2V Migration licensed), no matter which virtualization you had selected as the instant restore destination, the live migration destination can be only the same virtualization as you had selected for instant restore.
VM Configurations: The VM configuration settings for live migration VM are optional, but the modifications you had done for the instant restore VM will be applied to the live migration VM too, so, if you had modified the VM configurations for instant restore VM and you do not wish to apply the same modifications for live migration VM, please change the settings accordingly.
Transfer via: If the migration destination is ZStack, then users could choose to transfer the backup data from backup repository to production datastore via LAN, Encrypted Transfer, SAN (LAN-free), HotAdd and Backup Proxy, if migrating to other virtualizations, the transmission options will be different. For more information of the transmission options, please refer to Data Transmission.
Transfer Threads: By specifying the number of transfer threads, you can enable multithreaded transmission to improve the processing speed of the live migration job. The default value for multithreaded transmission is 3, even if you can set the value from 1 to 8, but usually 3 threads will be enough.
When done the above settings, click on OK to create the live migration job.
Run Live Migration Job
After creating a live migration job, you will be redirected to the Monitor Center > Jobs page. And the instant restore job on which you had created the live migration job, the job type will change from Instant Restore to Migration. By clicking on the job name, you'll be able to view the detailed VM migration process.
The migration process can be divided into 3 stages.
Stage 1:
Migrate the original backup data from backup repository to the production datastore. This stage will take a while depending on the backup data size and the transmission network speed.
Stage 2:
Migrate the new and changed data from backup repository to the production datastore, and when this stage is completed, the instant restore VM will be powered off.
Stage 3:
Migrate the new and changed data during powering off process of the instant restore VM to the production datastore, and when this stage is completed, the live migration VM will be powered on (if the power state preset is On during creating the live migration job).
But if the power state preset is Off, you need to manually turn on the live migration VM after live migration process is fully completed. You should not turn on the live migration VM until the instant restore job status changed from Migration to Running state again, otherwise the VM might become abnormal and the live migration will fail.
The above mentioned processes are automatic, and the whole process can guarantee the data consistency, when the live migration is completed, the live migration VM will take over all services of the instant restore VM with the virtual disks in production datastore.
Stop Instant Restore Job
After live migration completed, the instant restore job is still in a running status. But the VM created by instant restore job had already been powered off and all the data had been migrated to the live migration VM. Now you can stop the instant restore job and delete it.
To delete the instant restore job, please go to Monitor Center > Jobs page. Find the instant restore job which had already been performed live migration, click on Options and then select Stop. A dialog will be shown to warn you that stopping the instant restore job will lose all new and changed data during runtime of the instant restore VM, since you had already performed live migration for the instant restore VM, so please click on OK to confirm.
To stop the instant restore job, it involves the following procedures:
- Delete the instant restore VM.
- Unmount the NFS datastore.
- Clear the new and changed data of the instant restore VM within Vinchin backup repository.
Notice
- Make sure you had performed a live migration for the instant restore VM before stopping the instant restore job, otherwise you'll lose all new and changed data during the runtime of instant restore VM.
- Through out the whole process of instant restore and live migration, the original backup data from which runs the instant restore VM will always maintain unmodified, it will be mounted as an NFS storage in read-only state. The new and changed data of the instant restore VM will be stored in somewhere else within the backup repository.