Diff & Copy to a third location ?

[tiberiusQ]

Dear all,

I have two datasets with slightly different data in it, so I need to diff the difference between them but I only want to copy the diff which is in folderA to a third location for human inspection ;-)

Eg. I have two folders with a lot of subfolders files in it: folderA and FolderB
diff -qr folderA folderB
superimportantfile.pdf. Only in FolderA
etc.

After a while I found and tested a nice way to do this, but this does only works on debian based linux because the cp option -t is not implemented on bsd....
diff -rq folderA folderB | sed -n "s/^Only in \(folderA\): \(.*\)/'\1\/\2'/p" | xargs cp -rtv folderC

Of course I could fire up debian mount folderA folderB and folderC via cifs to run this but due to the amount of data I prefer to run this directly on the truenas box.

Does Anybody has an idea to achieve this on Truenas ?

 

[tiberiusQ]

bump

 

[jgreco]

Right, "-t" isn't POSIX, it's just random crap Linux added. I wish they wouldn't do that, because it significantly muddies the waters.

Is there some reason

# [blabla] | xargs -J % cp -rv % folderC

won't work for you? I concede that use of "-J" is not POSIX either, correct POSIX usage is left as an exercise to the reader. I'm just looking for the quick answer here.

 

[tiberiusQ]

You made my day - Thank you !

 

[tiberiusQ]

Except the posix thing I realized a new behavior which I really do not understand:

user@machine ~ % diff -rq a b
Only in a/folder: important.txt
user@machine ~ % diff -rq a b | sed -n "s/^Only in \(a\): \(.*\)/'\1\/\2'/p" | xargs -J % cp -rv % c
user@machine ~ % ls c
user@machine ~ %

As you can see diff shows the correct missing file in folder a but will not be copied to folder c if the folder exists in folder b.....

It looks like rsync could solve my issue:

rsync -rvcm --compare-dest=b/ a/ c/

 

[tiberiusQ]

Update & Background: About a year ago I copied with rsync -a locally between two smb datasets and got a lot of acl errors indeed and due to the amount of data and time it takes I decided to keep the orig. data and enable the new data set for production. Followed up I just defined acls on the new data-set and no user ever complained about anything, but I am not sure if there is a data diff between those datasets, that's why I try to put the diff to a third location and then back to the production dataset in a dolder called eg. diff. so that the users are able to double check + I am able to delete the orig. data including snapshots.

That's the reason for my odyssey on this thread or task and currently I do get two super different file size outputs between ok, makes sense.
rsync -rvtmc --compare-dest=b/ a/ c/
and
rsync -rvtm --compare-dest=b/ a/ c/

Does rsync -rvtmc --compare-dest=b/ a/ c/ the same what I expected from my diff command above ?

Does anybody has an equal task and can share skills and or can confirm that if you copy with rsync -a between smb datasets no data get lost except acls ? OR is there a zfs command to diff and copy to a third location ?

 

[tiberiusQ]

No commandlinefus onboard ?