5.4.0

Added

• Add post_install_sql topology hook to run SQL after plugins are installed and enabled

Changed

• pre_install_sql now fires unconditionally, even without plugins

Fixed

• Make the HTTP probe check infallible. It now returns false unless the HTTP status code is 200. #365

5.3.2

Changed

• Replace unmaintained serde_yaml with maintained fork serde_norway (47cab00808)

5.3.1

Fixed

• Fix MacOs compatibility issue due to procfs usage 552c00cdb

5.3.0

Added

• Add signal & timeout selection to pike stop command #340

Fixed

• Disable treating "bad" HTTP codes as errors #357
• Kill instance on drop to avoid infinite waiting of alive process #355

5.2.0

Added

• Print Web UI link on cluster start #347

Fixed

• Add aliases for field names in healthcheck response #333

Changed

• Bump picodata-plugin to 26.1.3

5.1.0

Fixed

• Skip symlinked instance directories during cluster shutdown in pike stop #342

Changed

• Enable waiting of vshard initialization by default #338

5.0.0

Breaking Changes

• Switch from liquid to minininja to reduce dependency footprint

Changed

• Switch from reqwest to ureq to reduce dependency footprint
• Bumped picodata-plugin to 26.1.2

4.0.2

Changed

• Switch tls in request to native-tls
• Increase timeout on vshard-readiness to 300 seconds

4.0.1

Fixed

• Check cluster readiness using both /startup and /ready probes
• Improve handling of vshard discovery completion #330

4.0.0

Breaking Changes

• Change API of pike::cluster::run to consume input parameters instead of referencing them and copying inside e99166e81
• Command pike run is checking readiness of cluster instances via HTTP probes introduced in Picodata 26.1 #308

Added

• Revival of terminated instances in sequential pike run commands #319
• Add optional waiting of vshard discovery during cluster startup in pike run #307

Changed

• Bump Picodata version to 26.1.1 in plugin template and CI
• Bump Picotest to 1.9.1 in plugin template