Comment 5 for bug 1470576

So my [limited] understanding is that, as it stands today, mount_check will cause the REPLICATE request to return a 507. While an in-progress replication cycle may be running, it won't start any NEW replication cycles after that? And therein lies Pete's point that it's as good as it will get.

Perhaps I'm just falling into a cycle where it's trying to replicate over 200GB of data in 1 replication cycle, but that seems unlikely with my data size and partition count.

Furthermore, if the disk fills up entirely, and I start clearing files off of the disk, replication will very quickly start filling it back up again. If the 507 was working, it seems like after a failure during one replication cycle, it would recognize the unmounted device before restarting the replication. (And if it's not, maybe that's the real underlying issue?)