[Tarantool-patches] [PATCH 0/2] Validate result of upserts squash & skip invalid upserts which can't be applied

Aleksandr Lyapunov alyapunov at tarantool.org
Tue Jun 23 15:32:29 MSK 2020


>> I bet we should just remove assert here. If mem_stmt have
>> a valid LSN now - it is a valid statement. The first of them must
>> have n_upserts=0, the second - 1 etc, and n_upsert of following
>> prepared statements must continue the sequence.
> What is the point of keeping counting n_upserts for upserts
> which can't be squashed? For sure we can simply remove assertion,
> but then after accumulating more than 128 upserts which can't
> be applied, vy_squash_process won't be called till dump...
>
Why can't? After accumulating more 128 upserts squash process
must be restarted and squash those upserts that became committed.



More information about the Tarantool-patches mailing list