- Improvements to the internal scheduler that core uses to load balance work between networking and ledger related tasks - this speeds up considerably catchup while reducing latency for processing SCP messages
- Fixed a crash when shutdown is called while catching up
- Reduce chance of triggering catchup when out of sync and out of order ledgers are received - this could happen on nodes that are running on low end hardware
- Avoid replaying many ledgers when rejoining the network - this could cause the node to hang while applying many ledgers
- Improve error messages when running into crypto errors
- Cleanup buckets folder after running a command line catchup - this releases disk space properly.
- Use a more recent version of sqlite