By default, the resharding transaction cloner only writes down its progress every 1000 entries. In stepdown suites, a failover is triggered every 8 seconds. In very slow variants (e.g. tsan debug), the cloner might be unable to process enough records to reach a checkpoint where progress is persisted before the next failover occurs, leaving it unable to make any progress (see BF-32013 for an example of this in practice).

We should reduce the batch size to 1 in stepdown suites to guarantee that the cloner is able to make progress, even if the system is very slow.