do you see anything in the syslog/journal while that happens?
maybe try to do tail -f /var/log/syslog and journalctl -f on a terminal and keep them running while you do a replication, see what happens in there when you get the timeout
As I understand it, it happens when zfs takes the local snapshot. For some delay, it displays this error. Could it increase the time, so that the timeout does not occur?