index_commands_shard_targeting.js runs index related commands with a maxTimeMs of 500ms and expects the following two things to happen:

1. These commands will get stuck waiting for a paused migration to finish and eventually fail with MaxTimeMsExpired error.
2. A log line to appear in shard0 indicating that the collection refresh failed due maxTimeMs expiration.

This works most of the time, but on slow variants it can happen that those requests reach their maxTimeMs at an earlier stage of their execution causing the mongos to cancel the request. In these cases the commands will still fail with MaxTimeMsExpired [1], but they didn't have the chance to reach the shard so [2] won't be fulfilled.