@
pettter
Why would there be "hours of downtime" using #
rsync
? They could easily do very quick per-repository downtime when doing the "final sync". No need to do all 9TB at once.