(excepting some replication lag). The slaves can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis tools, or perform administrative tasks (backup, restore, diff partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions up-to-date results. Semi-sync guarantees that at least one replica has the transaction in its relay log, but it has not necessarily been applied yet. The only way Vitess guarantees a fully up-to-date read is

(excepting some replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis tools, or perform administrative tasks (backup, restore, diff partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions up-to-date results. Semi-sync guarantees that at least one replica has the transaction in its relay log, but it has not necessarily been applied yet. The only way Vitess guarantees a fully up-to-date read is

(excepting some replication lag). The slaves can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis tools, or perform administrative tasks (backup, restore, diff partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions least one replica type slave from which queries will always return up-to-date results. Semi-sync guarantees that at least one slave has the transaction in its relay log, but it has not necessarily been applied

(excepting some replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis tools, or perform administrative tasks (backup, restore, diff have been requests to see if we can avoid 2PC and find a solution that could still give the same guarantees. This proposal is for such a solution. ### Existing solution The best existing solution is one partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions

(excepting some replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis tools, or perform administrative tasks (backup, restore, diff partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions up-to-date results. Semi-sync guarantees that at least one replica has the transaction in its relay log, but it has not necessarily been applied yet. The only way Vitess guarantees a fully up-to-date read is

ignore any replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis queries, or perform administrative tasks (backup, restore have been requests to see if we can avoid 2PC and find a solution that could still give the same guarantees. This proposal is for such a solution. Existing solution The best existing solution is one where partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. 95 Single shard

ignore any replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis queries, or perform administrative tasks (backup, restore have been requests to see if we can avoid 2PC and find a solution that could still give the same guarantees. This proposal is for such a solution. ### Existing solution The best existing solution is one partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions

ignore any replication lag). The replicas can serve read-only traffic (with eventual consistency guarantees), execute long-running data analysis queries, or perform administrative tasks (backup, restore have been requests to see if we can avoid 2PC and find a solution that could still give the same guarantees. This proposal is for such a solution. Existing solution The best existing solution is one where partial commits. You could instead use “2PC mode” transactions that give you distributed atomic guarantees. However, choosing this option increases the write cost by approximately 50%. Single shard transactions