- The save operation now automatically attempts to recover on failure during termination [see [#166](https://github.com/skytable/skytable/pull/166)]
- More than one process can no longer concurrently use the same data directory, preventing any possible data corruption [see [#169](https://github.com/skytable/skytable/pull/169), [#167](https://github.com/skytable/skytable/issues/167)]
- ⚠ Dropped support for Terrapipe 1.0 (reached EOL)
- ⚠ New disk storage format with much faster reads and/or writes
- Added support for Skyhash 1.0 (see [#147](https://github.com/skytable/skytable/pull/147))
- Fixed persistence bugs (see [#151](https://github.com/skytable/skytable/pull/151))
- Fixed bugs in background services (see [#152](https://github.com/skytable/skytable/pull/152))
- Make BGSAVE recoverable (see [#153](https://github.com/skytable/skytable/pull/153))
- Added `lskeys` action (see [#155](https://github.com/skytable/skytable/pull/155))
- Added `compat` module (see [#158](https://github.com/skytable/skytable/pull/158))
- Added backward compatibility for storage formats all the way upto 0.3.0. See [this wiki article](https://github.com/skytable/skytable/wiki/Disk-storage-formats) for more information
- To enable snapshots, you just have to add the key: there is no need for the `enabled` key. To disable snapshots, you just have to omit the `snapshot` key (block) from your configuration file
The Terrapipe protocol was revised and promoted to 1.0. This will cause all existing client implementations to break, since the protocol has changed fundamentally. The clients have to implement the [latest spec](https://terrabasedb.github.io/docs/Protocols/terrapipe).
The only significant change in the protocol is the new metalayout format: `#a#b#c` instead of the previously proposed `a#b#c#` .
### Disk storage format
The disk storage format was changed rendering existing binary data files incompatible. However, if you have any existing data - which is important, open an issue - because we'll be able to provide a tool that can help you easily migrate your existing datasets - with a one line command - so - no worries!