Right now if we had a large number of DDL changes in a single transaction (e.g., over 1 million), then the approach of RedisDDL of placing all of the changes into a single Redis value would fail.
It's an extreme corner case, and it's unlikely that such a large change set would not end up triggering a re-sync, which would side-step this entire process, but it's still possible.
Right now if we had a large number of DDL changes in a single transaction (e.g., over 1 million), then the approach of RedisDDL of placing all of the changes into a single Redis value would fail.
It's an extreme corner case, and it's unlikely that such a large change set would not end up triggering a re-sync, which would side-step this entire process, but it's still possible.