In the concurrency_with_mongot_extension_single_node suite, we get a recurring failure during teardown. This is caused by a known transient race condition which can occur in mongot (see https://jira.mongodb.org/browse/CLOUDP-316333 ).

The root cause here is that the concurrency suite drops the database, which is picked up by the replication on mongot and triggers the index cleanup. The concurrency suite then sends a SIGTERM to mongot, at which time mongot is still in the process of cleaning up the indexes. The sigterm causes mongot to close  the MongotCursorManagerImpl. The index cleanup from the database drop is still going on at this time, and when it tries to clean up another index (which issues a killIndexCursors) after the the SIGTERM was issued, we run into the race condition situation.

The mongot.py test fixture teardown fails because it expects to receive an exit code of 143 from mongot, confirming that it shut down due to the SIGTERM. Instead, we receive an exit code of 1.

As part of this ticket, we'll relax the check. While it's important to know whether mongot shut down gracefully or not, it is not critical for this test. In the teardown block, we already check that mongot is still running immediately before sending the sigterm to mongot - if it's not running, we fail. This catches the case that mongot crashed unexpectedly during the test run itself, which we'd want to flag for sure.