OVH shares overview of data lost in fire

Get a free Techzine subscription!

OVHcloud has shared a list of what data was lost in the fire that raged in the hosting company’s data centre last week. Many of the lost data have been backed up, but some of it is gone forever.

The overview of the possibly affected data is broken down into the different service branches of OVH: Web Cloud Universe, Bare Metal Universe, Hosted Private Cloud Universe and Public Cloud Universe.

Irretrievably lost data

Within Web Cloud Universe, the Plesk service hosted in SGB2 was completely lost. The other services within Web Cloud, such as Sharepoint, Email Pro and Dedicated Exchange, can be recovered.

With the Bare Metal Universe, almost all data is salvageable. Exceptions are:

  • Kimsufi in SBG1/S61B
  • Kimsufi in SBG1/S61C
  • NAS/HA in SBG1/S62B

In the Hosted Private Cloud Universe, a lot of data has been lost. These consist of:

  • Host (no local data) in SBG2/72A
  • Datastore in SBG1/62C
  • Datastore in SBG2/70D
  • Datastore in SBG2/72A
  • vRops in SBG2

A lot has also been lost in the Public Cloud Universe. The status of VPS Automated Backup and VPS Snapshot SSD Range is still being investigated. The information about the PCI Backup Services is also not yet complete. The data that has been definitively lost consists of:

  • VPS OpenStack Zone os-sbg1, os-sbg2 and os-sbg4 in SBG2
  • VPS Additional Storage OpenStack Zone os-sbg1, os-sbg2 and os-sbg4 in SBG2
  • VPS Snapshot Cloud is completely lost, except for VPS Snapshot New Range Ordered since 23 Apr 2021 in OpenStack Zone os-sbg6.
  • All Public Cloud Instances in SBG2 are lost.
  • Snapshot Volume OpenStack Zone os-sbg1 and os-sbg3 are gone.
  • EG-HG-SP Backup Instance
  • All PCS/PCA/CEPH Storage in SBG2 is lost.

Updates from Octave Klaba

Octave Klaba, the founder and CEO of OVHcloud, says that the investigation into the fire is moving faster than expected. He expects to restore power to the servers within a few days. Klaba shares all updates on the fire and its recovery in a Twitter thread.

Fire probably caused by faulty UPS

Last week, one of OVHcloud’s four datacentres in Strasbourg burned to the ground. It concerns datacentre SBG2. The neighbouring datacentre SBG1 also suffered considerable damage. SBG3 and SBG4 are still fully intact but will remain offline until it is certain that power and network connections can be safely restored. This is expected to happen in the course of next week.

The exact cause of the fire is not yet known, but it seems to have originated from two UPS systems. One of these UPSs had undergone extensive maintenance the day before the fire, during which several parts had been replaced.