Hello,
having multiple remotes, which are often unreachable, I get command output like:
git annex drop .
drop 2017 Haus Idstein/26108065_10212412239435248_1115645256_o.jpg (locking horus...) (lockcontent failed) (checking horus...) ssh: Could not resolve hostname horus.local: Name or service not known
(checking S3...) ok
drop 2017 Haus Idstein/26177516_10212412238475224_326557077_o.jpg (locking horus...) (lockcontent failed) (checking horus...) ssh: Could not resolve hostname horus.local: Name or service not known
(checking S3...) ok
drop 2017 Haus Idstein/26178378_10212412238835233_820819105_o.jpg (locking horus...) (lockcontent failed) (checking horus...) ssh: Could not resolve hostname horus.local: Name or service not known
(checking S3...) ok
which takes a long time, because it tries to lock horus first for each, waits for timeout and then tries the next remote. Of course I could use git annex drop --from=S3 but I do not always remember that. Is there a way to auto ignore a remote which is not reachable?
Thanks!
git annex drop --from=S3does not force checking the repo S3, but drops from S3.You can temporarily enable annex-ignore for a remote.
For example:
git -c remote.horus.annex-ignore=true git annex drop ...
Adjusting the annex-cost of the remote may also be useful, so git-annex will avoid using that remote when it can use a lower-cost remote instead. The annex-cost-command could be used to detect the network where the remote access times out and make its cost be very high then.