v2.2.0

IMPROVEMENTS

• switch to go 1.20
• after API server startup, if /var/lib/clickhouse/backup/*/(upload|download).state present, then operation will continue in background, fix 608
• make use_resumable_state: true behavior for upload and download, fix 608
• improved behavior --partitions parameter, for cases when PARTITION BY clause return hashed value instead of numeric prefix for partition_id in system.parts, fix 602
• apply system.macros values when use restore_schema_on_cluster and replace cluster name in engine=Distributed tables, fix 574
• switch S3 storage backend to https://github.com/aws/aws-sdk-go-v2/, fix 534
• added S3_OBJECT_LABLES and GCS_OBJECT_LABELS to allow setup each backup object metadata during upload fix 588
• added clickhouse-keeper as zookeeper replacement for integration test during reproduce 416
• decrease memory buffers for S3 and GCS, change default value for upload_concurrency and download_concurrency to round(sqrt(MAX_CPU / 2)), fix 539
• added ability to set up custom storage class for GCS and S3 depends on backupName pattern, fix 584

BUG FIXES

• fix ssh connection leak for SFTP remote storage, fix 578
• fix wrong Content-Type header, fix 605
• fix wrong behavior for download with --partitions fix 606
• fix wrong size of backup in list command if upload or download was break and resume, fix 526
• fix _successful_ and _failed_ metrics counter issue, happens after 2.1.0, fix 589
• fix wrong calculation date of last remote backup during startup
• fix wrong duration, status for metrics after 2.1.0 refactoring, fix 599
• fix panic on LIVE VIEW tables with option --restore-database-mapping db:db_new enabled), thanks @php53unit