The current wait of 3 seconds is pushing the envelope a bit.
Recovery can miss some updates when they're neither forwarded nor present in replicated index
Replication can still cause index corruption.
Lower solr.cloud.wait-for-updates-with-stale-state-pause back down from 7 seconds.
Use a very low value for solr.cloud.wait-for-updates-with-stale-state-pause in tests.