We ran into a recent issue where if we did a backup using the latest 4.2.7 version of MongoDB and mongodump, when we try to restore that backup on a 4.2.6 version of mongorestore and database, it hangs.  Seems to hang on an index restore near the end of the restore process., with no error message.  Restore process has to be killed manually.  This was using the community version.

Not sure if it's related to this: https://jira.mongodb.org/browse/TOOLS-2580

A point release shouldn't introduce incompatibilities in the dump/restore process, right?

After we restored the 4.2.7 to an Atlas instance running 4.2.6, the restore didn't hang but then we started getting errors like the following on Atlas during writes to the db with our app.  Not sure if this is a related issue or not:

2020-06-15T15:42:15.321226866Z F, 2020-06-15T15:42:15.321133 #15 FATAL – : [5a85fbc07361ceaf1cace2a927fde307]
2020-06-15T15:42:15.321236266Z [5a85fbc07361ceaf1cace2a927fde307] Mongo::Error::OperationFailure (Retryable write with txnNumber 7 is prohibited on session 68ecef31-38c1-43a9-aad4-2bcd9274fd6f - ddqot68vLpBUPnOuqFj2cdClv/dv/vF9ZVSwknYHEsE= because a newer retryable write with txnNumber 26 has already started on this session. (225) (on shore-tgln-uat-mongodb-shard-00-02-sw9w6.azure.mongodb.net:27017, modern retry, attempt 1)):
2020-06-15T15:42:15.321240266Z [5a85fbc07361ceaf1cace2a927fde307]
2020-06-15T15:42:15.321243066Z [5a85fbc07361ceaf1cace2a927fde307] app/concepts/mixin/contract/save.rb:21:in `save_contract'
2020-06-15T15:42:15.321245666Z [5a85fbc07361ceaf1cace2a927fde307] app/controllers/application_controller.rb:30:in `conduct'
2020-06-15T15:42:15.321248366Z [5a85fbc07361ceaf1cace2a927fde307] app/controllers/api/v1/recipients_controller.rb:23:in `update'