Release Notes for mongosync 1.20

New Features

• Adds the INITIALIZING state to the /progress API response. See progress for more details.

• mongosync now requires users to pass the new --enableCappedCollectionHandling flag during start if new capped collections will be created while the migration is ongoing.

Fixed Issues

• Migrations using destination write blocking to destination clusters without users now fail with a helpful message prompting the customer to provision a user on the destination cluster.

• Fixes an issue where mongosync panicked after receiving two SIGINT, SIGPIPE, SIGQUIT, or SIGTERM signals in a row.

Other Changes

• Adds and moves information about the ten largest collections from the entered CEA telemetry event to the entered collection copy event.

• mongosync does not support migrating into destination namespaces with pre-configured shard zone tags. For more information, see Sharded Clusters.

• Renames mongosync internal databases to be consistent with other internal tooling.

  The mongosync_reserved_for_internal_use database is now __mdb_internal_mongosync.

  The embedded verifier's databases mongosync_reserved_for_verification_{src,dst}_metadata are now __mdb_internal_mongosync_verifier_{src,dst}.

• Newly created capped collections during migration can cause mongosync to fail with error message a capped collection was created on the source but capped collection handling is not enabled;.

• Removes the enableUserWriteBlocking parameter from the /start API request. mongosync now manages write blocking on source and destination clusters based on your configuration. For more information, see Write Blocking.

Live Upgrades

Live upgrades to mongosync 1.20 are not supported.