2 ms·
Yeah I get your feeling. I had it similarly at first, dreading to make changes because I was afraid I'd mess up. And to be fair, I think ZFS could be better in
by magicalhippo 4y ago
Yeah I get your feeling. I had it similarly at first, dreading to make changes because I was afraid I'd mess up.
And to be fair, I think ZFS could be better in this regard. Some commands can put your pool into a very sub-optimal state, and ZFS doesn't warn about this when you enter those commands. Heck even the destroy pool command doesn't flinch if by chance nothing is mounted (which it may well be after recovery on a new system).
I found it helped to watch some of the videos from the OpenZFS conferences that explains the history of ZFS and how the architecture works, like the OpenZFS basics[1] one.
But I agree that the documentation[2] could have a lot more introductory material, to help those who aren't familiar with it.
That said, I echo the suggestion to try it out using file vdev's. For larger changes I do spin up a VM just to make sure. For example, it's possible to mess up replacing a disk by adding new disk as a new single vdev rather than replacing the failing one one, so if I feel unsure about it I take 15 minutes in a VM and write down the steps.
Again, this is something I feel they could improve. Adding a single-disk vdev to a mirrored or raid'ed pool should come with a warning requiring confirmation.
On the bright side, I've been running my pool since 2009, and have never lost data despite a few disk failures and countless unexpected power-outages without PSU. And I just run it on consumer hardware without ECC because that's what I got. Been up to 8 disks, now down to 6 and will soon go down to 4 once the new disks arrive. Send/recive ensures the data is just as it ever was on the new configuration.
[1]: https://www.youtube.com/watch?v=MsY-BafQgj4 https://www.youtube.com/watch?v=MsY-BafQgj4
[2]: https://openzfs.github.io/openzfs-docs/ https://openzfs.github.io/openzfs-docs/