Sharding QA for FB PSMDB v1.19.1

Existing limitations: and .

Tested with FB EVEREST-1884 add PSMDB 1.19.1 · percona/everest@7a0325a

Test 1: Restore to the same and new database after sharding data in mongodb 7.0.15-9

Scenario:
Create a 2 shard, 3 nodes, 3 router, 3 config server database.
Add data 1-2 GB and shard it.
Take 2 scheduled backups
Restore the backup to a new db using PITR, the restore gets stuck in Unknown status.
Result: Restore to a new database gets stuck in Unknown status.

Restore to the same database is successful using PITR.
Result: Restore is successful.

Edit the database, increase the shards from 2 to 3.
Take a scheduled backup
Restore to a new database using PITR.
Result: Restore to a new database gets stuck in Unknown status.

Restore to the same database using PITR.
Result: Restore is successful.

Edit the database, increase the shards from 3 to 4.
Take a scheduled backup.

Restore to a new database using PITR.
Result: Restore to a new database gets stuck in Unknown status.

Restore to the same database using PITR.
Result: Restore is successful.

Tested with FB EVEREST-1884 add PSMDB 1.19.1 · percona/everest@e89e63c

Test 2: Restore to a new database without sharding data

Scenario:
Create a 2 shard, 3 nodes, 3 router, 3 config server database.
Take 1 scheduled backup.
Scale up the eks cluster from 3 to 6 nodes.
Restore the backup to a new db.

Results:
For 8.0.4-1, restore to a new database gets stuck in Unknown status for 2 shards and 3 shards (3 nodes, 3 router, 3 config server database).
For 7.0.15-9, restore to a new database gets stuck in Unknown status for 3 shards.
Seems to be the same issue EVEREST-1644.
For 7.0.15-9, restore to a new database was successful with 2 shards.

FYI