metal-hammer
metal-hammer copied to clipboard
check health of attached disks
machines should only get to state "Waiting" if all attached disks are usable
maybe a small fdisk and newfs or something (when not in "reinstall" mode)?
Should we check all disks or only the disk where we want to install the OS ?
Checking the OS disk makes perfectly sense, because otherwise the later installation will fail and with this we could prevent failures during installation beforehand.
Checking the other disks can be extremely time consuming, imagine our s3 machines with more than 100 TB on dead slow disks, so this would slow down machine de-allocation and the time until this machine is marked as free again by this additional time.
Maybe we should start with the first part and see. WDYT ?