I was comparing my local BRT use against what fast dedup might get me[^1], and realised I would like a nice breakdown of refcount distribution. So I made one.
Description
Repurposes -TTT to show histograms. The full per-DVA list is now -TTTT.
[ ] I have run the ZFS Test Suite with this change applied.
[x] All commit messages are properly formatted and contain Signed-off-by.
[^1]: Turns out BRT is getting me basically everything I'd get from dedup. dedup would be marginally better at the cost of 11M unique entries. No thanks.
Motivation and Context
I was comparing my local BRT use against what fast dedup might get me[^1], and realised I would like a nice breakdown of refcount distribution. So I made one.
Description
Repurposes
-TTT
to show histograms. The full per-DVA list is now-TTTT
.How Has This Been Tested?
By hand:
Types of changes
Checklist:
Signed-off-by
.[^1]: Turns out BRT is getting me basically everything I'd get from dedup. dedup would be marginally better at the cost of 11M unique entries. No thanks.