Hi, Giving this thread a bump, as I still get nervous from the SMART mails and I don't want us to having to act under stress when something goes wrong. Oliver, thanks for the explanation. My suggestion would be to determine whether the disks are ready for replacement. If so, let's just ask Herzner to provide that for us. In the mean time we should verify our backup-policies, especially the off-site backups. Maybe open one or more of those archives to see that everything we expect to be there is in there. I am not convinced it is wise to run buildbots on the same physical device that we run our critical web services. Besides the use of resources which might or might not be not too bad, I am also worried about any physical wear and tear. Anyway, I can be convinced otherwise ;-) Regards, N>