The French host has suffered since Wednesday morning a general failure. It was linked to human error, according to its leader. At around 10:30 am, the sites were accessible again.
According to the Downdetector site, the first outage reports were sent shortly after 9 a.m. on the morning of October 13. Consequence: a large part of the websites hosted by the French giant were no longer responding. The return to normal was observed around 10:30 am.
At 10 o’clock this morning, its leader Octave Klaba published initial explanations, citing “human error” (linked to a copy / paste me) during network reconfiguration, in particular at the “backbone” level.
This term refers to the routers responsible for organizing the transfer of data between servers. It is a failure at the level of such routers which in particular caused the recent massive Facebook outage. According to these explanations, the failure is of software and not hardware origin.
The incident took place in a data center owned by OVH and based in the United States.
“There are two versions of the IP protocol, which is common to the entire Internet and which makes it possible to reach websites using a string of numbers. The IPv4 version, which was deployed in the early days of the Internet, and the more recent IPv6 version. which makes it possible to obtain many more IP addresses, but which is not yet deployed everywhere. It is the IPv4 sites which are affected by the blackout this morning “explains Stéphane Bortzmeyer, specialist in computer networks, to BFMTV.
In a second tweet, Octave Klaba specifies that the update follows a modification of the technical infrastructure of OVH in order to cope with an increase denial of service attacks (DDos). Such attacks use a multitude of machines to attempt to simultaneously connect to a site and saturate its servers.