Panellicense

Back up OpenVZ 7 containers with R1Soft without double copies

R1Soft on an OpenVZ 7 node copies every container twice by default. Here is the policy setup R1Soft recommends, what it means for new and stopped VPSs, and how restores work.

R1All R1Soft articlesTutorials & how-tos7 min readUpdated 2026-11-15
r1softopenvzvirtuozzoploopdisk-safe
schema: HowToschema: FAQPage

Point R1Soft at an OpenVZ 7 (Virtuozzo 7) node with automatic device detection on and the first full backup comes out roughly twice the size you expected. Every container is stored twice: once as the raw ploop image inside /vz/private, and again as the live ext4 filesystem mounted from /dev/ploopNNNNNp1. Disk Safe storage, replication bandwidth, and backup windows all double for no benefit.

This guide covers the policy layout R1Soft support recommends for OpenVZ 7 nodes, what happens to new and stopped containers, and how file-level and full-container restores work once the duplicate image files are gone. It assumes the agent is already installed on the node and the kernel module loads against the Virtuozzo kernel.

Why OpenVZ 7 gets backed up twice

On OpenVZ 7 each container's filesystem lives in a ploop image under its private area, typically /vz/private/<CTID>/root.hdd/root.hds with a DiskDescriptor.xml beside it. When the container starts, the host mounts that image as a block device and attaches it at /vz/root/<CTID>.

R1Soft's automatic device detection reads /proc/mounts, so it sees both the host volume holding /vz (with the image files inside it) and each mounted /dev/ploop* device. Back up both and you copy the same data twice, in two different formats.

Check what the agent will see on your node:

grep ploop /proc/mounts
du -sh /vz/private/*/root.hdd

The total from the second command is roughly the amount of duplicated data you will avoid.

R1Soft's guidance is to keep backing up both /vz and the ploop devices, and switch on the built-in exclusion for the image files.

Keep automatic device detection on

In Server Backup Manager, leave the Disk Safe on automatic device detection so it includes the host volume (/ or a dedicated /vz) and every mounted /dev/ploop* device. Do not hand-pick devices: ploop device numbers change when containers restart, and a hand-picked list goes stale.

Enable "Exclude Virtuozzo root image files"

  1. Open Policies and edit the policy protecting the node
  2. Open the Advanced Policy Settings tab
  3. Tick Exclude Virtuozzo root image files
  4. Save

The agent now skips the root.hds / root.hdd image files inside /vz/private but still captures everything else on /vz, including each container's ve.conf, DiskDescriptor.xml, and OS templates. Container data comes from the mounted ploop filesystems instead.

Verify the exclusion actually applied

Don't assume it worked. After the next run, open the latest recovery point, browse to /vz/private/<CTID>/root.hdd/ and confirm the image file is missing while DiskDescriptor.xml is present. Then compare Disk Safe usage with the previous full. If the image files are still there, collect the agent report and raise a case with R1Soft support:

/usr/bin/r1soft-report.sh
ls /tmp/r1soft-report-*.tar.gz

As a stopgap, add the per-container root.hdd image files as file excludes on the policy. That works, but you have to maintain the list as containers come and go, which is exactly what the built-in toggle avoids.

What happens to new and stopped containers

New containers are picked up automatically. When you create and start a container, the host mounts its ploop device under /vz/root/<CTID>. The agent rescans /proc/mounts at the start of each run and adds the new device to the backup. Its config files are already covered because /vz is in scope.

Stopped containers are not backed up. OpenVZ unmounts the ploop device when a container stops. With the image exclusion enabled, a container that is powered off during the backup window has neither its mounted filesystem nor its image captured. The previous recovery point remains, but nothing new is taken until the container is running again.

Restores with this layout

File-level restores

Each /dev/ploop* device is backed up as an ordinary ext4 filesystem, so you browse the recovery point and restore files or folders straight into /vz/root/<CTID>/ on a running container. No image mounting, no ploop tooling. This is the restore your support team will run most often, and it is the main reason not to back up /vz alone.

Full container restores

Because the image file is excluded, there is no single root.hds to drop back into place. To bring back a deleted container:

  1. Create an empty container with the same CTID and at least the same disk size as the original. Pull the original values from the backed-up ve.conf in the recovery point.
  2. Mount its empty filesystem without starting it:
prlctl mount 101
  1. In Server Backup Manager, restore the container's ploop device contents to /vz/root/101/ using Restore to alternate location.
  2. Unmount, check the config against the backed-up ve.conf, and start the container:
prlctl umount 101
prlctl start 101

It is more steps than copying an image back, but it works on any node with free space, including one with a different storage layout.

Why not back up /vz only

Disabling the ploop devices and backing up just /vz gives you image-level restores and nothing else. You lose single-file restores, and you snapshot the image while the guest filesystem inside it is live and unfrozen, which risks an inconsistent image. For a hosting node, file-level restores matter far more than image restores.

Databases inside containers are captured crash-consistent either way. If customers run busy MySQL or MariaDB instances in their VPSs, keep an in-container dump job alongside R1Soft. The MySQL restore guide covers recovering from dump files and raw data directories.

Sizing after the change

With images excluded, Disk Safe size tracks the data inside the containers rather than images plus their contents. The du figure from earlier is a good estimate of the saving. Re-check your plan against the SBM sizing guide before buying more backup storage. If the first full run after the change still saturates the node, throttle the policy.

Does R1Soft support OpenVZ 7 containers?+
Yes, with the agent installed on the OpenVZ host node. It cannot run inside a container because containers share the host kernel. One agent on the node protects every container on it.
Why is my R1Soft Disk Safe twice the size of my OpenVZ node data?+
Automatic device detection includes both /vz, which holds each container's ploop image, and the mounted /dev/ploop devices. Enable Exclude Virtuozzo root image files in the policy's Advanced Policy Settings to stop copying the images.
Does R1Soft back up stopped OpenVZ containers?+
Not with the image exclusion enabled. A stopped container's ploop device is unmounted, so neither its filesystem nor its image is captured until it is started again.
Can I restore a single file into an OpenVZ container with R1Soft?+
Yes. Each mounted ploop device is backed up as a normal ext4 filesystem, so you can browse the recovery point and restore files directly into /vz/root/<CTID>/ on the node.
How do I restore a deleted OpenVZ container from R1Soft?+
Create an empty container with the same CTID and disk size, mount it with prlctl mount, restore the backed-up ploop filesystem into /vz/root/<CTID>/, then unmount and start it.

Next steps

Switch in an afternoon

Switch from your current reseller — free.

We migrate active cPanel, Plesk, LiteSpeed and CloudLinux licenses from any reseller. We prorate the first month so you never pay twice, and your customers see zero downtime during the swap.