Oshibochka out: storage service GitLab code accidentally deleted almost all of the data
Last night, January 31, a service for storing GitLab code was unavailable. It turns out that the company's system administrator accidentally deleted ... almost everything. Assuming the number of errors at different stages of the copy databases from one server to another, the system administrator from the Netherlands launched the deletion of data from the primary server. When the process has been stopped, out of almost 300 GB of data "survived" only 4, 5 GB.
, a mistake was erased database, which contained requests for change of documentation and user code of the project; user repository (vault) were intact. Representatives GitLab report on each stage to restore the database. In the GitLab noted that the company did not help none of its five systems for storing backups. In one case, the procedure data storage fire with an error, which is why the backup was not created. In this case, the company had no error warning system when creating backups. Now the company is armed with one of the backups that was created manually for six hours before the incident. With it, the company is trying to recover the data.
GitLab was established in 2014 and is trying to compete with platforms such as GitHub, and Atlassian. GitLab advantage is that it is free of charge distributes its code to store tools and allows other organizations to create their own repository based GitLab systems with the addition of its own functions.
It is unknown how many users of the data will be restored intact, but the 6-hour backup givens for this kind of service - this is a very long time.