Copyright © 2019 MariaDB. All rights reserved.
How crash-safety is guaranteed on slave in case of parallel replication?
For example, if several transactions (A and B) from group are applied on slave in parallel. A is fast, while B is slow.
Slave crashed, when A was applied, while B was still in process.
Will slave continue B after restart? From which position slave will start replication: from next group, from B?
Is there any difference in this for MariaDB 10.0 and 10.1?