Unreachable drives and connections

A synchronisation needs both sides: the source it reads from and the target it writes to. When one side is not there – an external disk is not plugged in, a network share is not connected, an SSH server does not answer – the synchronisation is skipped for this pass. Nothing is copied, moved or deleted; the other synchronisations run as usual. As soon as the side is back, the synchronisation carries on by itself.

This is not an error. A backup disk that is plugged in only to back up is absent most of the time – and meant to be. Whether you want to hear about it anyway is your choice per synchronisation.

The option "Warn if the drive or SSH connection is missing"

It sits in the synchronisation's dialog on the Options tab and is switched on for new synchronisations.

Option	What happens when source or target is missing
On (default)	You assume the drive or server has to be there. If it is missing, that should be noticed: After Synchronise once a warning appears that names the synchronisation with source and target – whatever messages you have switched on in the settings. The log holds a line of level `Warning`. A backup proof, if switched on, records that the pass could not take place.
Off	The absence is normal for you. No warning appears, the backup proof stays empty, and the log notes the skip at level `Info` only.

Which setting fits when

Leave it on when the other side should always be there: the second internal disk, the NAS on the home network, the server that holds the backup. If it is missing, something is wrong and you want to know at once.
Switch it off when the other side is there only now and then: the external backup disk, the USB stick, the laptop that sees the home network only at home. The synchronisation catches up as soon as the side is back; a warning on every pass without the disk would be worth nothing.

If several synchronisations go to the same disk, set the option the same way on all of them – otherwise one warns and the other stays silent although both miss the same disk.

Where the absence shows

Place	Behaviour
Synchronise once	A warning only with the option on. If every synchronisation misses a side, the notice that there was nothing to do comes first, then the warning.
Preview	The preview always names skipped synchronisations, because it answers what a pass would do now – and for those synchronisations that is unknown. The message's icon is a warning when one of the named synchronisations has the option on, a notice otherwise.
Start (continuous) and the background service	No message boxes, only the log: warning or info depending on the option. A missing drive is noted once – not on every pass – and again when it goes missing after it has come back.
Assistant	The assistant knows the option of every synchronisation and whether the last pass missed a side. It can therefore explain why a warning came in one case and not in another.

Good to know

What counts as "unreachable": the source has to exist. On Linux the target has to exist as well; on Windows it is enough that the drive is there – a missing target directory is created. Over SSH the server has to answer and the login succeed; the source has to exist there.
Reachability is checked before every pass and the result is cached for a while. After a failure the next check comes after a few seconds, and while the failure lasts at most every five minutes. So it can take up to five minutes until a disk that has come back is noticed.
Because nothing is deleted in the target when the source is missing, an unplugged source disk can never cause the backup to be emptied.