Summary

We should add a spec test that ensures that large bulk_write operations inside an explicit transaction work properly.

When performing large bulk_writes, the write may internally be split into multiple smaller batches. When the bulk writes are executed inside an explicit transaction, there is a room for errors. For example, there was a user reported bug in Ruby driver, and as a result of this bug such bulk writes always failed - https://jira.mongodb.org/browse/RUBY-3044

Motivation

Who is the affected end user?

Users who uses large bulk write operations in transactions

How does this affect the end user?

There is a workaround - splitting the batch into smaller batches on the app side. But it must be very annoying, the driver should take care of this.

How likely is it that this problem or use case will occur?

We received only one bug report for the Ruby driver regarding this problem, even though this bug existed for quite a long time. So, it seems to be rather a rare case.

Is this issue urgent?

No.

Is this ticket only for tests?

This ticket is to create a spec tests.