zpool man page needs to document implicit checkpoint discard
In short, rewinding to a checkpoint silently discards the checkpoint. A user that wants to take multiple tries at a potentially destructive action could incur severe data loss after the first checkpoint rewind because they expect the checkpoint to stick around after rewinding.
I ran into this, but I was lucky enough that I was working on an unrelated feature so the only thing I've lost is the time to install from scratch and get back to where I was at.
I think this is important enough to fix in gate ahead of whatever OpenZFS may do, and then we can adopt their version whenever it makes its way into gate.
Updated by Electric Monk 7 months ago
- Status changed from New to Closed
- % Done changed from 0 to 100
commit 0f2a8434f3b6c1825df77e8037c61c0f6fb6b300 Author: Brian Bennett <email@example.com> Date: 2021-10-19T21:04:30.000Z 14165 zpool man page needs to document implicit checkpoint discard Reviewed by: Andy Fiddaman <firstname.lastname@example.org> Reviewed by: Dan McDonald <email@example.com> Approved by: Richard Lowe <firstname.lastname@example.org>