I have an example setup:
overlay /data/all overlay noauto,x-systemd.automount,lowerdir=/data/local:/data/remote 0 0
//myhomeserver/data /data/remote cifs iocharset=utf8,ro,credentials=/data/remote.credentials 0 0
Now, problems arise when myhomeserver can't be reached. Most cases, connection drops, comes back online, but /data/all no longer works (hangs), df hangs a bit and in the end it outputs a classic CIFS-related myhomeserver: Host is down message.
How to handle this in the most graceful way possible with current tech? I'm on latest stable Debian for both servers. Data is always accessed through overlayfs /data/all. I would require something like:
/data/alloverlayfs should never hang or at least do it minimally and recover/data/allshould expose available data from/data/localno matter the state of/data/remove/data/remoteshould be available best-effort, that is, recover automatically after a downtime, and upon recovery,/data/allshould again expose its contents
I think that systemd could somewhat help but don't know how I would use it properly in such a setup, can't find any examples. Maybe autofs could also help? Never used that though..