Today [14 May 2007] at approximately 08:15GMT we became aware of an issue affecting the server BART. The server was not responding to web or ping requests.

This server hosts web sites and the main stats server. As a result websites hosted on this server could not be accessed nor could the statistical data for any website hosted on the network.

The server was rebooted. However the server failed to restart. Engineers began checking the server system. It was discovered that the servers RAID array had become corrupt and needed to be reset and rebuilt.

RAID is used to add redundancy to the server. The server runs using two hard disks which mirror the others content. In the event of disk failure the other disk will continue to work and allow the hot swap of the failed disk.

The content mirroring is handled by the RAID array. It is unfortunately this that had become corrupt. As a result it had to be rebuilt. This process took much longer than expected as each piece of data on the two disks is rebuilt and compared via the array.

Service was not restored until 17:00GMT

Following a review there is nothing that we could have done to prevent this happening. However it took longer than anticipated to diagnose the actual problem. This has highlighted a training need within the data centre technical support staff which is now being addressed.

Please note that during the outage no statistical data was lost nor was any website data lost.

We apologise to all customers affected by the outage.