Improvements to data storage resilience

We’re pleased to announce that we’ve recently completed moving all VPS data to a new storage subsystem that mirrors all data across more than one physical server.

Prior to this, VPS data was stored on the RAID card attached to the physical server the VPS was running on. Although this provided good data safety by mirroring across multiple drives, the RAID card itself was a single point of failure.

The new storage subsystem eliminates this single points of failure.  Even if we were to suffer the complete failure of all disks attached to a physical machine and/or the complete failure of a raid card, including the situation where a raid card might inflict data corruption to all disks attached to it, this will not affect the VPS disk data on that machine. Additionally, providing that the physical server OS continues running, all VPS’ would be able to continue running without interruption in the face of such a disaster, due to the surviving network mirror on another physical machine.

Another storage related improvement we are planning to introduce over the coming months is the ability for customers to take snapshot backups of their VPS storage data, which will be stored elsewhere on our network. This would allow customers to restore their VPS disk data to a previous “known good” point in time e.g. if their VPS OS becomes unbootable, or is compromised by malicious activity/trojan/virus. We will announce on this blog as soon as we have rolled out this feature.

Leave a Reply