Timeline for View ZFS deduplication ratio on a dataset
Current License: CC BY-SA 3.0
8 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Jul 27, 2019 at 7:47 | comment | added | myrdd | fwiw, the above-mentioned ZFS Dedup FAQ is offline, but still available at Wayback Machine | |
Jun 21, 2016 at 13:58 | comment | added | user121391 |
You can also use zdb -S on your pool to get a rough estimate of the deduplication ratio as if it was applied. For more information, see the excellent post about deduplication on the blog of Constantin Gonzalez: constantin.glez.de/blog/2011/07/zfs-dedupe-or-not-dedupe
|
|
Oct 9, 2013 at 13:29 | vote | accept | deed02392 | ||
Dec 20, 2012 at 21:48 | comment | added | Mike Fitzpatrick |
I can't answer that, maybe it was a design decision, I don't know for sure. Deduplication has side effects on reported disk usage and free space and these tend to make more sense when considering the whole pool rather than just a filesystem within it. Refer to the ZFS Dedup FAQ: Deduplicated space accounting is reported at the pool level. You must use the zpool list command rather than the zfs list command to identify disk space consumption when dedup is enabled.
|
|
Dec 20, 2012 at 11:43 | comment | added | deed02392 | But if a duplicate block does exist in the filesystem with it enabled, surely this still yields a greater-than-one deduplication ratio, if only for that filesystem? Why can't I just see the ratio for that filesystem then? | |
Dec 20, 2012 at 1:21 | comment | added | Mike Fitzpatrick | To improve deduplication, ZFS doesn't limit itself to duplicate blocks within just one filesystem. Instead it looks across the whole pool. If dedup is not enabled for a particular filesystem then block writes are performed without passing through the dedup pipeline, even if duplicate blocks exist. | |
Dec 20, 2012 at 0:29 | comment | added | deed02392 | But I only turned it on for a 'filesystem' within the pool, if it has pool wide scope does that mean it's on for the whole pool? I was expecting when I created the dataset only that dataset would be deduped, so shouldn't I be able to see the ratio for only the dataset? | |
Dec 20, 2012 at 0:20 | history | answered | Mike Fitzpatrick | CC BY-SA 3.0 |