I had to replace a hard drive element of my zfs pool. I've been getting some warnings about errors in writing and reading weeks before. zfs is the last to know about a failing hard disk. If you want to know a disk's health then the SMART logs is your best bet.
This disk was finally made offline a few days before being replaced. The remaining three disks continue to function for the pool. I have a new disk ordered last November online that's waiting for this same situation to happen. I used it to replace the non-functioning disk.
I like that zfs designates the broken disk as /dev/sdb1/old and the new disk as sdb. zfs automatically resilvers the new disk. It took about an hour and everything is functioning as good as new.
Comments