Snapshot size / report

[Jano]

Hi,


I just upgraded my systems to 11.3 (from 11.2).

I had no any problems like this on 11.2.


I have primary system (PROD-1S) which acts as main NAS and another system (PROD-2S) which is backup of previous.

When I used 11.2 I could have more than 1 month backup (2 snapshots per day per lun [I have 45 luns]).
When I use 11.3 seems I cannot have even half of this due to lack of space.

What is additionally strange is report of partition free space - take a look:

As you can see from 02:00 till about 03:00 I reduce space from 4.55 TiB to zero... in 1h... when connection between both systems is just 1Gbit/s. Impossible specially because real transfer between these systems is up to 60MiB/s.

Looks like sent snapshots "preallocate" space but I do not know for what.

Anyone can explain this please ?


2020-09-02 Updated
Take a look for snapshot size on source (left) and target (right) systems:

Lun size is about 0,5T, so ok, snapshot on target system refers to whole lun size (561G) but size of each snapshot is significantly different... double.

Why ?

 

[Jano]

OK, I solved my problem.

Seems snapshot replicated to another system has pre-reservation set to source lun size (which also include snapshots taken on source side).
Solution for me was to set pre-reservation on target system to 0. Thanks this target system uses real data size.

For more info study command:
zfs get reservation,refreserv <storage>