Your VPS runs on a host node: an underlying physical server that houses individual VPSs. This article explains what happens to your VPS if its host node fails, and how to build redundancy across host nodes yourself.


TABLE OF CONTENTS



If the host node crashes and recovers


If the host node crashes and is recoverable when it is brought back up, the servers on it are evacuated (moved) to the next best available host node.

Important: Scheduled backups are required to restore your VPS after a catastrophic host failure. Changes made after the latest backup are not included. Without a backup, the VPS is reinstalled with its original operating system.


If the host node fails catastrophically


If a host node suffers a catastrophic failure, the latest backup you have for your VPS is restored onto the next best available host node. This depends on the server having scheduled backups; see Automated Backups to enable them. As with any full restore, changes made on the server since that backup are not included.

If the server has no backups, it is reinstalled with the operating system it was provisioned with.


Does BinaryLane keep a live copy of my server?


No. BinaryLane does not keep a live-replicating copy of your VPS on a different host node. If your service must stay available through a host node failure, build the redundancy yourself as described below.


Build redundancy across host nodes


  1. Run two servers that do the same job, for example two web servers with the same content.
  2. Set them as partners with Partner Server: open one server in mPanel, select the settings icon next to its name, choose Set Partner, pick the other server and select Save and Apply. Our systems then do their best to keep the two servers on different host nodes; this is best effort, not a guarantee.
  3. Put an Anycast Load Balancer in front of both servers. It health-checks each backend server every three seconds and passes requests only to healthy servers. The load balancer handles HTTP and HTTPS traffic (ports 80 and 443) only.

Keep scheduled backups on both servers as well, so that a catastrophic host failure restores your latest backup rather than a fresh operating system.


If you require assistance, feel free to submit a support ticket at our helpdesk here: Submit a ticket | BinaryLane