On Mon, 2019-10-21 at 11:48 -0700, Kevin Fenzi wrote:
Personally, I am willing to try 1 (will need more +1's for
freeze
break), I don't really like 2 and 3 seems a bit scary, but I guess it
could be ok until after freeze.
Thoughts?
I think #1 is fine, and we could even bump it to a very long time. The
goal is to get a consistent *historical* snapshot, and it sounds like
the DB is trying to get us a consistent *current* snapshot. As long as
we can get the consistent snapshot for the backup, it's OK if the
replica lags production by minutes or even hours, as long as it can
also recover. I think tuning it for that use case is sensible.