Data replicated to new pool seems smaller. Problem?

[Z300M]

I used zfs send / receive to copy data from an over-full pool to a new larger one, and the resultant one shows as being a little smaller: 17.46 TB vs. 17.57 TB. Is this something I should worry about? The compression type (lz4) and ratio (1.19) are the same.

df -c

shows somewhat fewer 1K blocks used on the new pool.

Is there a way to at least count the files on each pool, even if I can't (conveniently) do a file-by-file comparison?

 

[dlavigne]

df does not adequately display ZFS usage. Try zfs list instead.

 

[Z300M]

df does not adequately display ZFS usage. Try zfs list instead.

I executed

zfs list -t filesystem > FileList

on each pool

then

diff <Pool1>/FileList <Pool2>/FileList

and no differences were reported, so I assume that everything is OK. Correct?

 

[lopr]

keep in mind, that if you change things like compression on a pool, only new files written are affected. so maybe on the original pool you enabled compression at a later time leaving some files uncompressed that were now compressed when sending/receiving?
I have to admit I have problems wrapping my head around such problems too. But send/recieve a pool when you made some significant changes to your pool is a good way to really apply those changes to pool so I guess your new pool is more consistent than your old. especially if there where no errors when migrating and even the directory listings match I would not worry to much. if you want to be absoluty sure md5 your data and compare, eg. "tar -cf - datadir | md5sum" as suggested here: https://unix.stackexchange.com/ques...e-md5-sum-of-a-directorys-contents-as-one-sum