Data persistence and recovery

Data on root file systems and local storage are not persisted after instance termination and cannot be recovered by Denvr Support. This applies to Virtual Machines as well as Bare Metal.

Virtual Machines

Machine root disks are held on protected storage while an instance is Running or in Stopped state. This is slightly different than local scratch storage which exists on the physical host and is wiped during Stopped state.

Bare Metal

Machine root disks are deployed directly onto the assigned host. The host hardware uses RAID-1 protected boot disks for durability and is retained across host reboot.

Local scratch storage is not affected by node reboot. It is recommended that users configure local block storage using with RAID-1 to protect data, while still gaining the parallel throughput advantage of RAID-0 striping. This can be setup using ZFS, mdraid and similar Linux systems.

Host boot disks and local storage is wiped when reservations are complete.

High-Performance File Storage

Users should use High-Performance File Storage for data that persists beyond the life cycle of the machine instance. This is implemented via NFS and is available to Applications, Virtual Machines, and Bare Metal.

NFS can be mounted by user after a Virtual Machine or Bare Metal host are created.

File storage is not replicated between Regions and is not backed for recovery. The user must assess their own data availability requirements in the unlikely scenario that Denvr Cloud is unavailable.