Skip to content

Latest commit

 

History

History
552 lines (445 loc) · 39.8 KB

ChangeLog.md

File metadata and controls

552 lines (445 loc) · 39.8 KB
Raw

v2.4.5

BUG FIXES

  • add aws.LogResponse for S3_DEBUG (affected 2.4.4+ versions)

v2.4.4

BUG FIXES

  • remove aws.LogResponseWithBody for S3_DEBUG to avoid too many logs (affected 2.4.0+ versions)

v2.4.3

IMPROVEMENTS

  • add list command to API /backup/actions, fix 772

BUG FIXES

  • fix behavior for restore_as_attach: true for non-replicated MergeTree, fix 773
  • tables with ENGINE=Dictionary shall create after all dictionaries to avoid retry, fix 771

v2.4.2

IMPROVEMENTS

  • add cpu_nice_priority and io_nice_priority to config, which allow us to throttle CPU and IO usage for the whole clickhouse-backup process, fix 757

BUG FIXES

  • fix restore for object disk frozen_metadata.txt, fix 752
  • fix more corner cases for check_parts_columns: true, fix 747
  • fix applying macros to s3 endpoint in object disk during restore embedded backups, fix 750
  • rewrite GCS clients pool, set default GCS_CLIENT_POOL_SIZE as max(upload_concurrency, download_concurrency) * 3 to avoid stuck, fix 753, thanks @minguyen-jumptrading

v2.4.1

IMPROVEMENTS

  • switch to go-1.21
  • add clickhouse-server:23.8 for integration and testflows tests
  • add FTP_SKIP_TLS_VERIFY config option fix 742

BUG FIXES

  • fix calculation part size for S3 and buffer size for Azure to avoid errors for upload big files, fix 739 thanks @rodrigargar
  • fix GetFileReader for SSE encryption in S3, again fix 709

v2.4.0

IMPROVEMENTS

  • first implementation for properly backup S3/GCS/Azure disks, support server-side copy to back up bucket during clickhouse-backup create and during clickhouse-backup restore, requires add object_disk_path to s3,gcs,azblob section, fix 447
  • Implementation blacklist for table engines during backup / download / upload / restore 537
  • restore RBAC / configs, refactoring restart clickhouse-server via sql:SYSTEM SHUTDOWN or exec:systemctl restart clickhouse-server, add --rbac-only and --configs-only options to create, upload, download, restore command. fix [706]Altinity#706
  • Backup/Restore RBAC related objects from Zookeeper via direct connection to zookeeper/keeper, fix 604
  • Add SHARDED_OPERATION_MODE option, to easy create backup for sharded cluster, available values none (no sharding), table (table granularity), database (database granularity), first-replica (on the lexicographically sorted first active replica), thanks @mskwon, fix 639, fix 648
  • Add support for compression_format: none for upload and download backups created with --rbac / --rbac-only or --configs / --configs-only options, fix 713
  • Add support for s3 GLACIER storage class, when GET return error, then, it requires 5 minutes per key and restore could be slow. Use GLACIER_IR, it looks more robust, fix 614
  • restore functions via CREATE OR REPLACE for more atomic behavior
  • prepare to Make ./tests/integration/ test parallel fix 721
  • touch /var/lib/clickhouse/flags/force_drop_table before every DROP TABLE execution, fix 683
  • add support connection pool for google cloud storage, GCS_CLIENT_POOL_SIZE, fix 724

BUG FIXES

  • fix possible create backup failures during UNFREEZE not exists tables, affected 2.2.7+ version, fix 704
  • fix too strict system.parts_columns check when backup create, exclude Enum and Tuple (JSON) and Nullable(Type) vs Type corner cases, fix 685, fix 699
  • fix --rbac behavior when /var/lib/clickhouse/access not exists
  • fix skip_database_engines behavior
  • fix skip_databases behavior during restore for corner case db.prefix* and corner cases when conflict with --tables="*pattern.*", fix 663
  • fix S3 head object Server Side Encryption parameters, fix 709

v2.3.2

BUG FIXES

  • fix error when backups_to_keep_local: -1, fix 698
  • minimal value for download_concurrency and upload_concurrency 1, fix 688
  • do not create UDF when use --data flag, fix 697

v2.3.1

IMPROVEMENTS

  • add support use_environment_credentials option inside clickhouse-server backup object disk definition, fix 691
  • add but skip tests for azure_blob_storage backup disk for use_embbeded_backup_restore: true, it works, but slow, look ClickHouse/ClickHouse#52088 for details

BUG FIXES

  • fix static build for FIPS compatible mode fix 693
  • complete success/failure server callback notification even when main context canceled, fix 680
  • clean command will not return error when shadow directory not exists, fix 686

v2.3.0

IMPROVEMENTS

  • add FIPS compatible builds and examples, fix 656, fix 674
  • improve support for use_embedded_backup_restore: true, applied ugly workaround in test to avoid ClickHouse/ClickHouse#43971, and applied restore workaround to resolve ClickHouse/ClickHouse#42709
  • migrate to clickhouse-go/v2, fix 540, close 562
  • add documentation for AWS_ARN_ROLE and AWS_WEB_IDENTITY_TOKEN_FILE, fix 563

BUG FIXES

  • hotfix wrong empty files when disk_mapping contains not exists during create, affected 2.2.7 version, look details 676
  • add FTP_ADDRESS and SFTP_PORT in Default config Readme.md section fix 668
  • when use --tables=db.materialized_view pattern, then create/restore backup also for .inner.materialized_view or .inner_id.uuid, fix 613

v2.2.8

BUG FIXES

  • hotfix wrong empty files when disk_mapping contains not exists during create, affected 2.2.7 version, look details 676

v2.2.7

IMPROVEMENTS

  • Auto-tuning concurrency and buffer size related parameters depending on remote storage type, fix 658
  • add CLICKHOUSE_BACKUP_MUTATIONS and CLICKHOUSE_RESTORE_AS_ATTACH config options to allow backup and properly restore table with system.mutations is_done=0 status. fix 529
  • add CLICKHOUSE_CHECK_PARTS_COLUMNS config option and --skip-check-parts-column CLI parameter to watch, create and create_remote commands to disallow backup with inconsistent column data types fix 529
  • add test coverage reports for unit, testflows and integration tests, fix 644
  • use UNFREEZE TABLE in ClickHouse after backup finished to allow s3 and other object storage disks unlock and delete remote keys during merge, fix 423

BUG FIXES

  • apply SETTINGS check_table_dependencies=0 to DROP DATABASE statement, when pass --ignore-dependencies together with --rm in restore command, fix 651
  • add support for masked secrets for ClickHouse 23.3+, fix 640

v2.2.6

BUG FIXES

  • fix panic for resume upload after restart API server for boolean parameters, fix 653
  • apply SETTINGS check_table_dependencies=0 to DROP DATABASE statement, when pass --ignore-dependencies together with --rm in restore command, fix 651

v2.2.5

BUG FIXES

  • fix error after restart API server for boolean parameters, fix 646
  • fix corner cases when restore_schema_on_cluster: cluster, fix 642, error happens on 2.2.0-2.2.4
  • fix Makefile targets build-docker and build-race-docker for old clickhouse-server version
  • fix typo retries_pause config definition in general section

v2.2.4

BUG FIXES

  • fix wrong deletion on S3 for versioned buckets, use s3.HeadObject instead of s3.GetObjectAttributes, fix 643

v2.2.3

BUG FIXES

  • fix wrong parameters parsing from *.state file for resumable upload \ download after restart, fix 641

v2.2.2

IMPROVEMENTS

  • add callback parameter to upload, download, create, restore API endpoints, fix 636

BUG FIXES

  • add system.macros could be applied to path config section to ReadMe.md, fix 638
  • fix connection leaks for S3 versioned buckets during execution upload and delete command, fix 637

v2.2.1

IMPROVEMENTS

  • add additional server-side encryption parameters to s3 config section, fix 619
  • restore_remote will not return error when backup already exists in local storage during download check, fix 625

BUG FIXES

  • fix error after restart API server when .state file present in backup folder, fix 623
  • fix upload / download files from projections multiple times, cause backup create wrong create *.proj as separate data part, fix 622

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

v2.1.3

IMPROVEMENTS

  • during upload sort tables descending by total_bytes if this field present
  • improve ReadMe.md add description for all CLI commands and parameters
  • add use_resumable_state to config to allow default resumable behavior in create_remote, upload, restore_remote and download commands, fix 576

BUG FIXES

  • fix --watch-backup-name-template command line parsing, overridden after config reload, fix 548
  • fix wrong regexp, when restore_schema_on_cluster: cluster_name, fix 552
  • fix wrong clean command and API behavior, fix 533
  • fix getMacro usage in Examples for backup / restore sharded cluster.
  • fix deletion files from S3 versioned bucket, fix 555
  • fix --restore-database-mapping behavior for ReplicatedMergeTree (replace database name in replication path) and Distributed (replace database name in underlying table) tables, fix 547
  • MaterializedPostgreSQL doesn't support FREEZE, fix 550, see also ClickHouse/ClickHouse#32902, ClickHouse/ClickHouse#44252
  • create and restore commands will respect skip_tables config options and --table cli parameter, to avoid create unnecessary empty databases, fix 583
  • fix watch unexpected connection closed behavior, fix 568
  • fix watch validation parameters corner cases, close 569
  • fix --restore-database-mapping behavior for ATTACH MATERIALIZED VIEW, CREATE VIEW and restore --data corner cases, fix 559

v2.1.2

IMPROVEMENTS

  • add watch description to Examples.md

BUG FIXES

  • fix panic when use --restore-database-mapping=db1:db2, fix 545
  • fix panic when use --partitions=XXX, fix 544

v2.1.1

BUG FIXES

  • return bash and clickhouse usergroup to Dockerfile image short, fix 542

v2.1.0

IMPROVEMENTS

  • complex refactoring to use contexts, AWS and SFTP storage not full supported
  • complex refactoring for logging to avoid race condition when change log level during config reload
  • improve kubernetes example for adjust incremental backup, fix 523
  • add storage independent retries policy, fix 397
  • add clickhouse-backup-full docker image with integrated kopia, rsync, restic and clickhouse-local, fix 507
  • implement GET /backup/kill?command=XXX API to allow kill, fix 516
  • implement kill "full command" in POST /backup/actions handler, fix 516
  • implement watch in POST /backup/actions handler API and CLI command, fix 430
  • implement clickhouse-backup server --watch to allow server start watch after start, fix 430
  • update metric last_{create|create_remote|upload}_finish metrics values during API server startup, fix 515
  • implement clean_remote_broken command and POST /backup/clean/remote_broken API request, fix 520
  • add metric number_backups_remote_broken to calculate broken remote backups, fix 530

BUG FIXES

  • fix keep_backups_remote behavior for recursive incremental sequences, fix 525
  • for restore command call DROP DATABASE IF EXISTS db SYNC when pass --schema and --drop together, fix 514
  • close persistent connections for remote backup storage after command execution, fix 535
  • lot of typos fixes
  • fix all commands was always return 200 status (expect errors) and ignore status which passed from application code in API server

v2.0.0

IMPROVEMENTS

  • implements remote_storage: custom, which allow us to adopt any external backup system like restic, kopia, rsync, rclone etc. fix 383
  • add example workflow how to make backup / restore on sharded cluster, fix 469
  • add use_embedded_backup_restore to allow BACKUP and RESTORE SQL commands usage, fix 323, need 22.7+ and resolve ClickHouse/ClickHouse#39416
  • add timeout to azure config AZBLOB_TIMEOUT to allow download with bad network quality, fix 467
  • switch to go 1.19
  • refactoring to remove legacy storage package
  • add table parameter to tables cli command and /backup/tables API handler, fix 367
  • add --resumable parameter to create_remote, upload, restore_remote, donwload commands to allow resume upload or download after break. Ignored for remote_storage: custom, fix 207
  • add --ignore-dependencies parameter to restore and restore_remote, to allow drop object during restore schema on server where schema objects already exists and contains dependencies which not present in backup, fix 455
  • add restore --restore-database-mapping=<originDB>:<targetDB>[,<...>], fix 269, thanks @mojerro

BUG FIXES

  • fix wrong upload / download behavior for compression_format: none and remote_storage: ftp

v1.6.2

IMPROVEMENTS

  • add Azure to every CI/CD run, testing with Azurite

BUG FIXES

  • fix azblob.Walk with recursive=True, for properly delete remote backups

v1.6.1

BUG FIXES

  • fix system.macros detect query

v1.6.0

IMPROVEMENTS

  • add storage_class (GCS_STORAGE_CLASS) support for remote_storage: gcs fix 502
  • upgrade aws golang sdk and gcp golang sdk to latest versions

v1.5.2

IMPROVEMENTS

  • switch to go 1.19
  • refactoring to remove legacy storage package

v1.5.1

BUG FIXES

  • properly execute CREATE DATABASE IF NOT EXISTS ... ON CLUSTER when setup restore_schema_on_cluster, fix 486

v1.5.0

IMPROVEMENTS

  • try to improve implementation check_replicas_before_attach configuration to avoid concurrent ATTACH PART execution during restore command on multi-shard cluster, fix 474
  • add timeout to azure config AZBLOB_TIMEOUT to allow download with bad network quality, fix 467

v1.4.9

BUG FIXES

  • fix download behavior for parts which contains special characters in name, fix 462

v1.4.8

IMPROVEMENTS

  • add check_replicas_before_attach configuration to avoid concurrent ATTACH PART execution during restore command on multi-shard cluster, fix 474
  • allow backup list when clickhouse server offline, fix 476
  • add use_custom_storage_class (S3_USE_CUSTOM_STORAGE_CLASS) option to s3 section, thanks @realwhite

BUG FIXES

  • resolve {uuid} marcos during restore for ReplicatedMergeTree table and ClickHouse server 22.5+, fix 466

v1.4.7

IMPROVEMENTS

  • PROPERLY restore to default disk if disks not found on destination clickhouse server, fix 457

v1.4.6

BUG FIXES

  • fix infinite loop error can't acquire semaphore during Download: context canceled, and error can't acquire semaphore during Upload: context canceled all 1.4.x users recommends upgrade to 1.4.6

v1.4.5

IMPROVEMENTS

  • add CLICKHOUSE_FREEZE_BY_PART_WHERE option which allow freeze by part with WHERE condition, thanks @vahid-sohrabloo

v1.4.4

IMPROVEMENTS

  • download and restore to default disk if disks not found on destination clickhouse server, fix 457

v1.4.3

IMPROVEMENTS

  • add API_INTEGRATION_TABLES_HOST option to allow use DNS name in integration tables system.backup_list, system.backup_actions

BUG FIXES

  • fix upload_by_part: false max file size calculation, fix 454

v1.4.2

BUG FIXES

  • fix --partitions parameter parsing, fix 425

v1.4.1

BUG FIXES

  • fix upload data go routines waiting, expect the same upload speed as 1.3.2

v1.4.0

IMPROVEMENTS

  • add S3_ALLOW_MULTIPART_DOWNLOAD to config, to improve download speed, fix 431
  • add support backup/restore user defined functions, fix 420
  • add clickhouse_backup_number_backups_remote, clickhouse_backup_number_backups_local, clickhouse_backup_number_backups_remote_expected,clickhouse_backup_number_backups_local_expected prometheus metric, fix 437
  • add ability to apply system.macros values to path field in all types of remote_storage, fix 438
  • use all disks for upload and download for multi-disk volumes in parallel when upload_by_part: true fix #400

BUG FIXES

  • fix wrong warning for .gz, .bz2, .br archive extensions during download, fix 441

v1.3.2

IMPROVEMENTS

  • add TLS certificates and TLS CA support for clickhouse connections, fix 410
  • switch to go 1.18
  • add clickhouse version 22.3 to integration tests
  • add S3_MAX_PARTS_COUNT and AZBLOB_MAX_PARTS_COUNT for properly calculate buffer sizes during upload and download for custom S3 implementation like Swift
  • add multithreading GZIP implementation

BUG FIXES

  • fix 406, properly handle path for S3, GCS for case when it begins from "/"
  • fix 409, avoid delete partially uploaded backups via backups_keep_remote option
  • fix 422, avoid cache broken (partially uploaded) remote backup metadata.
  • fix 404, properly calculate S3_PART_SIZE to avoid freeze after 10000 multi parts uploading, properly handle error when upload and download go-routine failed to avoid pipe stuck

v1.3.1

IMPROVEMENTS

  • fix 387, improve documentation related to memory and CPU usage

BUG FIXES

  • fix 392, correct download for recursive sequence of diff backups when DOWNLOAD_BY_PART true
  • fix 390, respect skip_tables patterns during restore and skip all INFORMATION_SCHEMA related tables even skip_tables don't contain INFORMATION_SCHEMA pattern
  • fix 388, improve restore ON CLUSTER for VIEW with TO clause
  • fix 385, properly handle multiple incremental backup sequences + BACKUPS_TO_KEEP_REMOTE

v1.3.0

IMPROVEMENTS

  • Add API_ALLOW_PARALLEL to support multiple parallel execution calls for, WARNING, control command names don't try to execute multiple same commands and be careful, it could allocate much memory during upload / download, fix #332
  • Add support for --partitions on create, upload, download, restore CLI commands and API endpoint fix #378 properly implementation of #356
  • Add implementation --diff-from-remote for upload command and properly handle required on download command, fix #289
  • Add print-config cli command fix #366
  • Add UPLOAD_BY_PART (default: true) option for improve upload/download concurrency fix #324
  • Add support ARM platform for Docker images and pre-compiled binary files, fix #312
  • KeepRemoteBackups should respect differential backups, fix #111
  • Add SFTP_DEBUG option, fix #335
  • Add ability to restore schema ON CLUSTER, fix #145
  • Add support encrypted disk (include s3 encrypted disks), fix #260
  • API Server optimization for speed of last_backup_size_remote metric calculation to make it async during REST API startup and after download/upload, fix #309
  • Improve list remote speed via local metadata cache in $TEMP/.clickhouse-backup.$REMOTE_STORAGE, fix #318
  • Add CLICKHOUSE_IGNORE_NOT_EXISTS_ERROR_DURING_FREEZE option, fix #319
  • Add support for PROJECTION, fix #320
  • Return clean cli command and API POST /backup/clean endpoint, fix #379

BUG FIXES

  • fix #300, allow GCP properly work with empty GCP_PATH value
  • fix #340, properly handle errors on S3 during Walk() and delete old backup
  • fix #331, properly restore tables where have table name with the same name as database name
  • fix #311, properly run clickhouse-backup inside docker container via entrypoint
  • fix #317, properly upload large files to Azure Blob Storage
  • fix #220, properly handle total_bytes for uint64 type
  • fix #304, properly handle archive extension during download instead of use config settings
  • fix #375, properly REMOTE_STORAGE=none error handle
  • fix #379, will try to clean shadow if create fail during moveShadow
  • more precise calculation backup size during upload, for backups created with --partitions, fix bug after #356
  • fix restore --rm behavior for 20.12+ for tables which have dependent objects (like dictionary)
  • fix concurrency by FTP creation directories during upload, reduce connection pool usage
  • properly handle --schema parameter for show local backup size after download
  • fix restore bug for WINDOW VIEW, thanks @zvonand

EXPERIMENTAL

  • Try to add experimental support for backup MaterializedMySQL and MaterializedPostgeSQL tables, restore MySQL tables not impossible now without replace table_name.json to Engine=MergeTree, PostgresSQL not supported now, see ClickHouse/ClickHouse#32902

v1.2.4

HOT FIXES

  • fix 409, avoid delete partially uploaded backups via backups_keep_remote option

v1.2.3

HOT FIXES

  • fix 390, respect skip_tables patterns during restore and skip all INFORMATION_SCHEMA related tables even skip_tables don't contain INFORMATION_SCHEMA pattern

v1.2.2

IMPROVEMENTS

  • Add REST API POST /backup/tables/all, fix POST /backup/tables to respect CLICKHOUSE_SKIP_TABLES

BUG FIXES

  • fix #297, properly restore tables where have fields with the same name as table name
  • fix #298, properly create system.backup_actions and system.backup_list integration tables for ClickHouse before 21.1
  • fix #303, ignore leading and trailing spaces in skip_tables and --tables parameters
  • fix #292, lock clickhouse connection pool to single connection

v1.2.1

IMPROVEMENTS

  • Add REST API integration tests

BUG FIXES

  • fix #290
  • fix #291
  • fix CLICKHOUSE_DEBUG settings behavior (now we can see debug log from clickhouse-go)

v1.2.0

INCOMPATIBLE CHANGES

  • REST API /backup/status now return only latest executed command with status and error message

IMPROVEMENTS

  • Added REST API /backup/list/local and /backup/list/remote to allow list backup types separately
  • Decreased background backup creation time via REST API /backup/create, during avoid list remote backups for update metrics value
  • Decreased backup creation time, during avoid scan whole system.tables when set table query string parameter or --tables cli parameter
  • Added last and filter query string parameters to REST API /backup/actions, to avoid pass to client long JSON documents
  • Improved FTP remote storage parallel upload / download
  • Added FTP_CONCURRENCY to allow, by default MAX_CPU / 2
  • Added FTP_DEBUG setting, to allow debug FTP commands
  • Added FTP to CI/CD on any commit
  • Added race condition check to CI/CD

BUG FIXES

  • environment variable LOG_LEVEL now apply to clickhouse-backup server properly
  • fix #280, incorrect prometheus metrics measurement for /backup/create, /backup/upload, /backup/download
  • fix #273, return S3_PART_SIZE back, but calculates it smartly
  • fix #252, now you can pass last and filter query string parameters
  • fix #246, incorrect error messages when use REMOTE_STORAGE=none
  • fix #283, properly handle error message from FTP server
  • fix #268, properly restore legacy backup for schema without database name

v1.1.1

BUG FIXES

  • fix broken system.backup_list integration table after add required field in Altinity#263
  • fix #274 invalid SFTP_PASSWORD environment usage

v1.1.0

IMPROVEMENTS

  • Added concurrency settings for upload and download, which allow loading table data in parallel for each table and each disk for multi-disk storages
  • Up golang version to 1.17
  • Updated go libraries dependencies to actual version (exclude azure)
  • Add Clickhouse 21.8 to test matrix
  • Now S3_PART_SIZE not restrict upload size, partSize calculate depends on MAX_FILE_SIZE
  • improve logging for delete operation
  • Added S3_DEBUG option to allow debug S3 connection
  • Decrease number of SQL queries to system.* during backup commands
  • Added options for RBAC and CONFIG backup, look to clickhouse-backup help create and clickhouse-backup help restore for details
  • Add S3_CONCURRENCY option to speedup backup upload to S3
  • Add SFTP_CONCURRENCY option to speedup backup upload to SFTP
  • Add AZBLOB_USE_MANAGED_IDENTITY support for ManagedIdentity for azure remote storage, thanks https://github.com/roman-vynar
  • Add clickhouse-operator kubernetes manifest which run clickhouse-backup in server mode on each clickhouse pod in kubernetes cluster
  • Add detailed description and restrictions for incremental backups.
  • Add GCS_DEBUG option
  • Add CLICKHOUSE_DEBUG option to allow low-level debug for clickhouse-go

BUG FIXES

  • fix #266 properly restore legacy backup format
  • fix #244 add read_timeout, write_timeout to client-side timeout for clickhouse-go
  • fix #255 restrict connection pooling to 1 in clickhouse-go
  • fix #256 remote_storage: none, was broke compression
  • fix #266 legacy backups from version prior 1.0 can't restore without allow_empty_backup: true
  • fix #223 backup only database metadata for proxy integrated database engines like MySQL, PostgresSQL
  • fix GCS global buffer wrong usage during UPLOAD_CONCURRENCY > 1
  • Remove unused SKIP_SYNC_REPLICA_TIMEOUTS option

v1.0.0

BUG FIXES

  • Fixed silent cancel uploading when table has more than 4k files (fix #203, #163. Thanks mastertheknife)
  • Fixed download error for zstd and brotli compression formats
  • Fixed bug when old-format backups hadn't cleared

v1.0.0-beta2

IMPROVEMENTS

  • Added diff backups
  • Added retries to restore operation for resolve complex tables dependencies (Thanks @Slach)
  • Added SFTP remote storage (Thanks @combin)
  • Now databases will be restored with the same engines (Thanks @Slach)
  • Added create_remote and restore_remote commands
  • Changed of compression format list. Added zstd, brotli and disabled bzip2, sz, xz

BUG FIXES

  • Fixed empty backup list when S3_PATH and AZBLOB_PATH is root
  • Fixed azblob container issue (Thanks @atykhyy)

v1.0.0-beta1

IMPROVEMENTS

  • Added 'allow_empty_backups' and 'api.create_integration_tables' options
  • Wait for clickhouse in server mode (fix #169)
  • Added Disk Mapping feature (fix #162)

BUG FIXES

  • Fixed 'ftp' remote storage (#164)

v0.6.5

It is the last release of v0.x.x

IMPROVEMENTS

  • Added 'create_remote' and 'restore_remote' commands
  • Changed update config behavior in API mode

BUG FIXES

v1.0.0-alpha1

IMPROVEMENTS

  • Support for new versions of ClickHouse (#155)
  • Support of Atomic Database Engine (#140, #141, #126)
  • Support of multi disk ClickHouse configurations (#51)
  • Ability to upload and download specific tables from backup
  • Added partitions backup on remote storage (#83)
  • Added support for backup/upload/download schema only (#138)
  • Added new backup format select it by compression_format: none option

BROKEN CHANGES

  • Changed backup format
  • Incremental backup on remote storage feature is not supported now, but will support in future versions

v0.6.4

IMPROVEMENTS

  • Added CLICKHOUSE_AUTO_CLEAN_SHADOW option for cleaning shadow folder before backup. Enabled by default.
  • Added CLICKHOUSE_SYNC_REPLICATED_TABLES option for sync replicated tables before backup. Enabled by default.
  • Improved statuses of operations in server mode

BUG FIXES

  • Fixed bug with semaphores in server mode