I should wait patiently until it’s good and ready, but my main zpool is around 94% full. So I’m just trying to decide if I should wait for this feature to be ready, or if should I destroy/rebuild my pool from backup to expand it. I’d rather wait, but for how long?
What do you guys think is the ETA? Like 2-3 months or maybe 6+ months?
I know it’s preferable to start a zpool from scratch rather than expand it, but honestly, all you need to do is rewrite the file, and the new files will be re-striped across all disks (including newly added ones). So that procedure is immensely safer than destroying the whole pool to expand. I use this script to affect zpool changes.
For reference, the PR: raidz expansion feature by don-brady · Pull Request #15022 · openzfs/zfs · GitHub
You have just as much data as we do.
Like any major new feature, given that (according to the PR) extensive testing has not occurred, it is not likely to be ready any time soon. Unfixable flaws may yet be found, however unlikely. Even if no major flaws are found it is unlikely to hit mainstream for a while.
You would be wise to plan for the feature to not be ready in time for your migration. Having spent a number of decades in Enterprise IT, it is never wise to count on an unreleased feature for your short-term planning. If, when you’re ready to pull the trigger on your alternate plan, the feature is rigorously tested and available, let it be a pleasant surprise instead of counting on it and being disappointed that it’s not there when you wanted it.
You could replace the drives with larger drives, one at a time, or add a second vdev since you’re willing to rewrite files.