Contradictory workloads
Fundamentally, group replication is designed to allow updates from any of the members in the group. Rollback of transactions based on overlap of rows is checked for each of the transactions; the rest are committed and sent to be updated to other members in the group. If several updates on the same row happen frequently, it can result in multiple rollbacks. You might come across cyclic situations where one server updates, requests others to update, and, in parallel, another has already updated for the same row. This would result in rollback.
To prevent such a scenario, you can have the last member of the group apply the update, after which you proceed to another one. You can have similar updates routed only from the ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access