In case the recommendation not filling up your pool over 75 % fails and the pool is filled up to 100% there are some steps which can be done to free up space tough.
As ZFS is a copy-on-write (COW) filesystem it needs free space to delete files!
A remedy after the pool has filled up is to ‚truncate‘ some large files:
root@nas3:˜# cat /dev/null > /path/to/big/file
A more preventive action is to reserve space before shit hits the fan with creation of a dataset:
root@nas3:˜# zfs create vol1-mirror/reserved
root@nas3:˜# zfs set reservation =1G vol1-mirror/reserved
So when the pool fills up to 100% just delete the reserved dataset:
root@nas3:˜# zfs destroy vol1-mirror/reserved
Afterwards it should be able to delete files to free up space.
Don’t forget to recreate the dummy dataset again…