Fail fast when nodesync is enabled on a table

[adejanovski]

Project board link

DSE nodesync will prevent running anti-entropy repairs, which makes segments fail repeatedly without failing the repair itself.
We'd need to fail fast if we see that nodesync is indeed enabled on a table that is part of a repair.
We have table specific checks to verify if it is using TWCS and act accordingly. We can probably reuse these code paths to fail repairs if a table in the repair scope has nodesync enabled.

┆Issue is synchronized with this Jira Story by Unito
┆Issue Number: REAP-197