On Oct 4, 2017 10:13 PM, "javocado" <javocado at gmail.com> wrote: Setting vfs.zfs.free_max_blocks to 20k has not helped unfortunately. No, that won't help with this issue as the destroy operation is already in progress and part of a transaction group. But it will mitigate or (hopefully) prevent this issue in the future. Cheers, Freddie