Disk Replacement
Your disks will fail, but you’ll usually get some warnings because ZFS proactively checks every occupied bit to guard against silent corruption. This is normally done every month, but you can launch one manually if you’re suspicious. They take a long time, but operate at low priority.
# Start a scrub
zpool scrub pool01
# Check the status
zpool scrub -s pool01
You can check the status of your pool at anytime with the command zpool status. When there’s a problem, you’ll see this:
zpool status
NAME STATE READ WRITE CKSUM
pool01 DEGRADED 0 0 0
raidz1 DEGRADED 0 0 0
/dev/sda ONLINE 0 0 0
/dev/sdb ONLINE 0 0 0
/dev/sdc ONLINE 0 0 0
/dev/sdd FAULTED 53 0 0 too many errors
Time to replace that last drive before it goes all the way bad
# You don't need to manually offline the drive if it's faulted, but it's good practice to as there's other states it can be in
zpool offline pool01 /dev/sdd
# Physically replace that drive. If you're shutting down to do this, the replacement usually has the same device path
zpool replace pool01 /dev/sdd
There’s a lot of strange things that can happen with drives and depending on your version of ZFS it might be using UUIDs or other drive identification strings. Check the link below for some of those conditions.
Feedback
Was this page helpful?
Glad to hear it! Please tell us how we can improve.
Sorry to hear that. Please tell us how we can improve.