Issue

When running the jstests/sharding/secondary_reads_waitForShardingInit.js test under the sharding_csrs_continuous_config_stepdown suite, we have sporadically seen a test failure due to a removeShard operation removing a shard, and then getting interrupted due to a CSRS step-down.
The returned error makes the mongos retry the removeShard operation, which in-turn results in a ShardNotFound error back to the test (instead of the usual response document of the removeShard command). This issue is already known and tracked under SERVER-32553.

Fix

To prevent its occurrence for the given case, we should make sure that secondary_reads_waitForShardingInit.js calls removeShard through a helper, which accepts the ShardNotFound error as a correct result, similarly to how it's done in other tests such as jstests/sharding/store_historical_placement_data.js (see SERVER-89414).