How can I replace it?
By replacing it and resilvering the pool. This is the first thing one MUST TEST when starting to add any form of complexity to a setup before any productive data is put on such a storage pool. Blindly trusting in technology would work the way it should or the user thinks he understood is a great recipe for future data loss. 
(my rant above is not meant for you but for other users coming accross this thread so at least they might do better)
"The world is a sphere!" This statement is factually correct but not helpful in this case. 
But this should be helpful vor you: Replacing a Failing Hard Drive in a ZFS Pool
You First ADD a drive to a Pool ...
then replace error drive to a new drive .
then detach the error drive ....
I create this type of scenario when i testing my OMV for my Office network .. Hope its Help ..
run this 2 command and post the output ...
ls -lh /dev/disk/by-id/
lsblk
Resilvering 5.99TB now with an ETA of 9h44m
6 * 1000 * 1000 / ( 9.75 * 3600) == 170.x MB/s. A bit high for RAIDZ1 with WD RED. How much capacity is the pool using? Care to provide output from 'zpool list'?
It's doing it now at 151M/s with 8h38m to go, down from the initial estimate of 200M/s.
Well, to me this looks like a lot less real resilver performance since with RAIDZ a resilver takes only care about the used data and not full capacity. So with a 'traditional' RAID5 6TB would need to be resynced (block device layer knowing nothing about filesystems) while with RAIDZ and given I understand the output above correctly (6TB-4.88TB used) it's just a 1.12 TB resilver instead. On fast hosts it's normal that resync/resilver performance slows down a bit over time since HDDs are faster on the outer tracks than on the inner.
649G resilvered, 44.36% done
@sardaukar Successful finished meanwhile?
@sardaukar Successful finished meanwhile?
@cabrio_leo sorry for the late reply - yes, it worked 
Don’t have an account yet? Register yourself now and be a part of our community!
