When interrupted, the background compaction server generates the following message:

WT_SESSION.compact: [WT_VERB_DEFAULT][ERROR]: __wt_session_compact_check_interrupted, 238: background compact interrupted by application: WT_ERROR: non-specific WiredTiger error

This is consistent with the behaviour of foreground compaction when it is interrupted. This message was not seen before as we don't test the interruption of foreground compaction is test/format or in our Python tests. We have a dedicated C test for it.

However, there are several issues because of this:

• Since test/format is testing this code by turning on/off the background compaction server, we may see that log which could be a red herring if there is another failure.
• If the background compaction is enabled in a Python test, when the test ends and the connection is being closed, if the background compaction was doing some work, the message is generated and the Python suite will see it as a failure as no logs are expected.
• Our parsing rules to triage our failures will pick up on the logs which could be another red herring.

The goal of this ticket is to find a way forward and decide if we should keep the message and update our frameworks to take it into account or stop generating this message. Since we have a stat for interruption, it could be enough.