Dying Hard Drive

One of the six hard drives in my local file server has acted up one too many times and so I’ve made the decision to migrate all the data off it to a new drive. As the system is set up using Logical Volume Management, this is fairly straight forward: add the new drive to the system and the volume group (in this case, an 8TB Seagate), then pvmove everything from the failing drive to the new one, then remove the old drive from the system.

One small wrinkle though: that acting up. When the drive gets too many read requests too quickly, it starts giving read errors until the system is reset. So I get to transfer 2TB+ of data in small chunks. Manually. I’ve been doing this for a day now, and probably have another two or three to go. At least I don’t have to take the system offline to do the data transfers.

I should be able to continue using the drive in a system that is less disk I/O intensive, like one of the compute nodes I’m in the middle of building. Now to get the next data transfer block going before I head back outside to cut up fallen limbs…