Hi, I just want to make sure I understand import-url
correctly. Say I do:
dvc import-url ssh://<host>/myfile.dat
This creates a myfile.dat.dvc
file in my repo which records in it the remote path ssh://<host>/myfile.dat
, and downloads the myfile.dat
from the remote server. If I later dvc push
, the file myfile.dat
does not get pushed to the remote cache.
If someone else clones my repo and does a dvc pull
, they will download myfile.dat
from ssh://<host>/myfile.dat
as well as recorded in the dvc file, rather than from the remote cache (where it doesn’t exist). I suppose the hash will be used to check the file is still actually the same on the remote.
I can also do dvc update --to-remote myfile.dat.dvc
, which does push the file to the remote cache, if I want to safeguard against it being deleted/changed on ssh://<host>/myfile.dat.
Is this all right?
If I do do dvc update --to-remote
, then will future new dvc pull
s get the file from the remote cache, or from ssh://<host>/myfile.dat
?
Thanks!