Similar to the resharding unit tests that stepdown/stepup during each phase, create unit tests that will induce an unrecoverable error during each phase. After the unrecoverable error, ensure that all tasks are cancelled and appropriately cleaned up.

Additionally, the resharding components use WithAutomaticRetry future chains, however, the "onUnrecoverableError" block remains unused in most places. The error logging is deferred to the "onError" block. Consider adding logging inside the onUnrecoverableError similar to the resharding_batch_oplog_applier.cpp.