Your seedboxes and storage boxes kept running, your data was never touched, and the client area has worked normally since 21:00 UTC on 6 November. On the evening of 6 November a denial-of-service attack flooded our website from about 4,400 addresses in more than 1,400 networks. From about 20:14 to 21:00 UTC the client area (billing, the ... Read More »

Small reads on a shared disk answer much faster on servers with the new read limits: on three busy servers among the first to get them, the wait for a small read fell 93 to 97%. A shared disk serves everyone on the server from one queue. Until now one account reading flat out, a big library scan or a full recheck, could fill that queue, and ... Read More »