Error "have inconsistent data types for active data part" in v2.6.2

[ziggi]

Just upgraded from 2.4.19 to 2.6.2 and got this error:

2024-10-24 16:21:00.212 ERR pkg/clickhouse/clickhouse.go:1290 > `srv_s3`.`player_joinquit` have incompatible data types []string{"Nullable(Enum8('disconnect' = 1, 'timeout' = 2, 'kicked' = 3, 'fake' = 4))", "Nullable(Enum8('disconnect' = 1, 'timeout' = 2, 'kicked' = 3))"} for "eventType" column
2024-10-24 16:21:00.212 ERR pkg/backup/create.go:278 > b.AddTableToLocalBackup error: `srv_s3`.`player_joinquit` have inconsistent data types for active data part in system.parts_columns table=srv_s3.player_joinquit

ClickHouse server version 23.12.2.59 (official build).

Config

general:
  remote_storage: s3           # REMOTE_STORAGE, choice from: `azblob`,`gcs`,`s3`, etc; if `none` then `upload` and `download` commands will fail.
  max_file_size: 1073741824      # MAX_FILE_SIZE, 1G by default, useless when upload_by_part is true, use to split data parts files by archives
  backups_to_keep_local: 10       # BACKUPS_TO_KEEP_LOCAL, how many latest local backup should be kept, 0 means all created backups will be stored on local disk
                                 # -1 means backup will keep after `create` but will delete after `create_remote` command
                                 # You can run `clickhouse-backup delete local <backup_name>` command to remove temporary backup files from the local disk
  backups_to_keep_remote: 0      # BACKUPS_TO_KEEP_REMOTE, how many latest backup should be kept on remote storage, 0 means all uploaded backups will be stored on remote storage.
                                 # If old backups are required for newer incremental backup then it won't be deleted. Be careful with long incremental backup sequences.
  log_level: info                # LOG_LEVEL, a choice from `debug`, `info`, `warning`, `error`
  allow_empty_backups: false     # ALLOW_EMPTY_BACKUPS
  # Concurrency means parallel tables and parallel parts inside tables
  # For example, 4 means max 4 parallel tables and 4 parallel parts inside one table, so equals 16 concurrent streams
  download_concurrency: 2        # DOWNLOAD_CONCURRENCY, max 255, by default, the value is round(sqrt(AVAILABLE_CPU_CORES / 2))
  upload_concurrency: 2          # UPLOAD_CONCURRENCY, max 255, by default, the value is round(sqrt(AVAILABLE_CPU_CORES / 2))

  # Throttling speed for upload and download, calculates on part level, not the socket level, it means short period for high traffic values and then time to sleep
  download_max_bytes_per_second: 0  # DOWNLOAD_MAX_BYTES_PER_SECOND, 0 means no throttling
  upload_max_bytes_per_second: 0    # UPLOAD_MAX_BYTES_PER_SECOND, 0 means no throttling

  # when table data contains in system.disks with type=ObjectStorage, then we need execute remote copy object in object storage service provider, this parameter can restrict how many files will copied in parallel  for each table
  object_disk_server_side_copy_concurrency: 32
  # when CopyObject failure or object disk storage and backup destination have incompatible, will warning about possible high network traffic
  allow_object_disk_streaming: false

  # RESTORE_SCHEMA_ON_CLUSTER, execute all schema related SQL queries with `ON CLUSTER` clause as Distributed DDL.
  # Check `system.clusters` table for the correct cluster name, also `system.macros` can be used.
  # This isn't applicable when `use_embedded_backup_restore: true`
  restore_schema_on_cluster: ""
  upload_by_part: true           # UPLOAD_BY_PART
  download_by_part: true         # DOWNLOAD_BY_PART
  use_resumable_state: true      # USE_RESUMABLE_STATE, allow resume upload and download according to the <backup_name>.resumable file

  # RESTORE_DATABASE_MAPPING, restore rules from backup databases to target databases, which is useful when changing destination database, all atomic tables will be created with new UUIDs.
  # The format for this env variable is "src_db1:target_db1,src_db2:target_db2". For YAML please continue using map syntax
  restore_database_mapping: {}

  # RESTORE_TABLE_MAPPING, restore rules from backup tables to target tables, which is useful when changing destination tables.
  # The format for this env variable is "src_table1:target_table1,src_table2:target_table2". For YAML please continue using map syntax
  restore_table_mapping: {}

  retries_on_failure: 3          # RETRIES_ON_FAILURE, how many times to retry after a failure during upload or download
  retries_pause: 5s              # RETRIES_PAUSE, duration time to pause after each download or upload failure

  watch_interval: 1h       # WATCH_INTERVAL, use only for `watch` command, backup will create every 1h
  full_interval: 24h       # FULL_INTERVAL, use only for `watch` command, full backup will create every 24h
  watch_backup_name_template: "shard{shard}-{type}-{time:20060102150405}" # WATCH_BACKUP_NAME_TEMPLATE, used only for `watch` command, macros values will apply from `system.macros` for time:XXX, look format in https://go.dev/src/time/format.go

  sharded_operation_mode: none       # SHARDED_OPERATION_MODE, how different replicas will shard backing up data for tables. Options are: none (no sharding), table (table granularity), database (database granularity), first-replica (on the lexicographically sorted first active replica). If left empty, then the "none" option will be set as default.

  cpu_nice_priority: 15    # CPU niceness priority, to allow throttling CPU intensive operation, more details https://manpages.ubuntu.com/manpages/xenial/man1/nice.1.html
  io_nice_priority: "idle" # IO niceness priority, to allow throttling DISK intensive operation, more details https://manpages.ubuntu.com/manpages/xenial/man1/ionice.1.html

  rbac_backup_always: true # always, backup RBAC objects
  rbac_resolve_conflicts: "recreate"  # action, when RBAC object with the same name already exists, allow "recreate", "ignore", "fail" values
clickhouse:
  username: "***"                # CLICKHOUSE_USERNAME
  password: "***"                     # CLICKHOUSE_PASSWORD
  host: localhost                  # CLICKHOUSE_HOST, To make backup data `clickhouse-backup` requires access to the same file system as clickhouse-server, so `host` should localhost or address of another docker container on the same machine, or IP address bound to some network interface on the same host.
  port: 9110                       # CLICKHOUSE_PORT, don't use 8123, clickhouse-backup doesn't support HTTP protocol
  # CLICKHOUSE_DISK_MAPPING, use this mapping when your `system.disks` are different between the source and destination clusters during backup and restore process.
  # The format for this env variable is "disk_name1:disk_path1,disk_name2:disk_path2". For YAML please continue using map syntax.
  # If destination disk is different from source backup disk then you need to specify the destination disk in the config file:

  # disk_mapping:
  #  disk_destination: /var/lib/clickhouse/disks/destination

  # `disk_destination`  needs to be referenced in backup (source config), and all names from this map (`disk:path`) shall exist in `system.disks` on destination server.
  # During download of the backup from remote location (s3), if `name` is not present in `disk_mapping` (on the destination server config too) then `default` disk path will used for download.
  # `disk_mapping` is used to understand during download where downloaded parts shall be unpacked (which disk) on destination server and where to search for data parts directories during restore.
  disk_mapping: {}
  # CLICKHOUSE_SKIP_TABLES, the list of tables (pattern are allowed) which are ignored during backup and restore process
  # The format for this env variable is "pattern1,pattern2,pattern3". For YAML please continue using list syntax
  skip_tables:
    - system.*
    - INFORMATION_SCHEMA.*
    - information_schema.*
  # CLICKHOUSE_SKIP_TABLE_ENGINES, the list of tables engines which are ignored during backup, upload, download, restore process
  # The format for this env variable is "Engine1,Engine2,engine3". For YAML please continue using list syntax
  skip_table_engines: []
  timeout: 5m                  # CLICKHOUSE_TIMEOUT
  freeze_by_part: false        # CLICKHOUSE_FREEZE_BY_PART, allow freezing by part instead of freezing the whole table
  freeze_by_part_where: ""     # CLICKHOUSE_FREEZE_BY_PART_WHERE, allow parts filtering during freezing when freeze_by_part: true
  secure: false                # CLICKHOUSE_SECURE, use TLS encryption for connection
  skip_verify: false           # CLICKHOUSE_SKIP_VERIFY, skip certificate verification and allow potential certificate warnings
  sync_replicated_tables: true # CLICKHOUSE_SYNC_REPLICATED_TABLES
  tls_key: ""                  # CLICKHOUSE_TLS_KEY, filename with TLS key file
  tls_cert: ""                 # CLICKHOUSE_TLS_CERT, filename with TLS certificate file
  tls_ca: ""                   # CLICKHOUSE_TLS_CA, filename with TLS custom authority file
  log_sql_queries: true        # CLICKHOUSE_LOG_SQL_QUERIES, enable logging `clickhouse-backup` SQL queries on `system.query_log` table inside clickhouse-server
  debug: false                 # CLICKHOUSE_DEBUG
  config_dir:      "/etc/clickhouse-server"              # CLICKHOUSE_CONFIG_DIR
  # CLICKHOUSE_RESTART_COMMAND, use this command when restoring with --rbac, --rbac-only or --configs, --configs-only options
  # will split command by ; and execute one by one, all errors will logged and ignore
  # available prefixes
  # - sql: will execute SQL query
  # - exec: will execute command via shell
  restart_command: "exec:systemctl restart clickhouse-server"
  ignore_not_exists_error_during_freeze: true # CLICKHOUSE_IGNORE_NOT_EXISTS_ERROR_DURING_FREEZE, helps to avoid backup failures when running frequent CREATE / DROP tables and databases during backup, `clickhouse-backup` will ignore `code: 60` and `code: 81` errors during execution of `ALTER TABLE ... FREEZE`
  check_replicas_before_attach: true # CLICKHOUSE_CHECK_REPLICAS_BEFORE_ATTACH, helps avoiding concurrent ATTACH PART execution when restoring ReplicatedMergeTree tables
  use_embedded_backup_restore: false # CLICKHOUSE_USE_EMBEDDED_BACKUP_RESTORE, use BACKUP / RESTORE SQL statements instead of regular SQL queries to use features of modern ClickHouse server versions
  embedded_backup_disk: ""  # CLICKHOUSE_EMBEDDED_BACKUP_DISK - disk from system.disks which will use when `use_embedded_backup_restore: true`
  backup_mutations: true # CLICKHOUSE_BACKUP_MUTATIONS, allow backup mutations from system.mutations WHERE is_done=0 and apply it during restore
  restore_as_attach: false # CLICKHOUSE_RESTORE_AS_ATTACH, allow restore tables which have inconsistent data parts structure and mutations in progress
  check_parts_columns: true # CLICKHOUSE_CHECK_PARTS_COLUMNS, check data types from system.parts_columns during create backup to guarantee mutation is complete
  max_connections: 0 # CLICKHOUSE_MAX_CONNECTIONS, how many parallel connections could be opened during operations
azblob:
  endpoint_suffix: "core.windows.net" # AZBLOB_ENDPOINT_SUFFIX
  account_name: ""             # AZBLOB_ACCOUNT_NAME
  account_key: ""              # AZBLOB_ACCOUNT_KEY
  sas: ""                      # AZBLOB_SAS
  use_managed_identity: false  # AZBLOB_USE_MANAGED_IDENTITY
  container: ""                # AZBLOB_CONTAINER
  path: ""                     # AZBLOB_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""         # AZBLOB_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  compression_level: 1         # AZBLOB_COMPRESSION_LEVEL
  compression_format: tar      # AZBLOB_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  sse_key: ""                  # AZBLOB_SSE_KEY
  buffer_size: 0               # AZBLOB_BUFFER_SIZE, if less or eq 0 then it is calculated as max_file_size / max_parts_count, between 2Mb and 4Mb
  max_parts_count: 10000       # AZBLOB_MAX_PARTS_COUNT, number of parts for AZBLOB uploads, for properly calculate buffer size
  max_buffers: 3               # AZBLOB_MAX_BUFFERS
  debug: false                 # AZBLOB_DEBUG
s3:
  access_key: "***"                   # S3_ACCESS_KEY
  secret_key: "***"                   # S3_SECRET_KEY
  bucket: "***"                       # S3_BUCKET
  endpoint: "https://storage.yandexcloud.net"                     # S3_ENDPOINT
  region: ru-central1                # S3_REGION
  # AWS changed S3 defaults in April 2023 so that all new buckets have ACL disabled: https://aws.amazon.com/blogs/aws/heads-up-amazon-s3-security-changes-are-coming-in-april-of-2023/
  # They also recommend that ACLs are disabled: https://docs.aws.amazon.com/AmazonS3/latest/userguide/ensure-object-ownership.html
  # use `acl: ""` if you see "api error AccessControlListNotSupported: The bucket does not allow ACLs"
  acl: private                     # S3_ACL
  assume_role_arn: ""              # S3_ASSUME_ROLE_ARN
  force_path_style: false          # S3_FORCE_PATH_STYLE
  path: ""                         # S3_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""             # S3_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  disable_ssl: false               # S3_DISABLE_SSL
  compression_level: 1             # S3_COMPRESSION_LEVEL
  compression_format: tar          # S3_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  # look at details in https://docs.aws.amazon.com/AmazonS3/latest/userguide/UsingKMSEncryption.html
  sse: ""                          # S3_SSE, empty (default), AES256, or aws:kms
  sse_customer_algorithm: ""       # S3_SSE_CUSTOMER_ALGORITHM, encryption algorithm, for example, AES256
  sse_customer_key: ""             # S3_SSE_CUSTOMER_KEY, customer-provided encryption key use `openssl rand 32 > aws_sse.key` and `cat aws_sse.key | base64`
  sse_customer_key_md5: ""         # S3_SSE_CUSTOMER_KEY_MD5, 128-bit MD5 digest of the encryption key according to RFC 1321 use `cat aws_sse.key |  openssl dgst -md5 -binary | base64`
  sse_kms_key_id: ""               # S3_SSE_KMS_KEY_ID, if S3_SSE is aws:kms then specifies the ID of the Amazon Web Services Key Management Service
  sse_kms_encryption_context: ""   # S3_SSE_KMS_ENCRYPTION_CONTEXT, base64-encoded UTF-8 string holding a JSON with the encryption context
                                   # Specifies the Amazon Web Services KMS Encryption Context to use for object encryption.
                                   # This is a collection of non-secret key-value pairs that represent additional authenticated data.
                                   # When you use an encryption context to encrypt data, you must specify the same (an exact case-sensitive match)
                                   # encryption context to decrypt the data. An encryption context is supported only on operations with symmetric encryption KMS keys
  disable_cert_verification: false # S3_DISABLE_CERT_VERIFICATION
  use_custom_storage_class: false  # S3_USE_CUSTOM_STORAGE_CLASS
  storage_class: STANDARD          # S3_STORAGE_CLASS, by default allow only from list https://github.com/aws/aws-sdk-go-v2/blob/main/service/s3/types/enums.go#L787-L799
  concurrency: 1                   # S3_CONCURRENCY
  part_size: 0                     # S3_PART_SIZE, if less or eq 0 then it is calculated as max_file_size / max_parts_count, between 5MB and 5Gb
  max_parts_count: 10000           # S3_MAX_PARTS_COUNT, number of parts for S3 multipart uploads
  allow_multipart_download: false  # S3_ALLOW_MULTIPART_DOWNLOAD, allow faster download and upload speeds, but will require additional disk space, download_concurrency * part size in worst case
  checksum_algorithm: ""           # S3_CHECKSUM_ALGORITHM, use it when you use object lock which allow to avoid delete keys from bucket until some timeout after creation, use CRC32 as fastest

  # S3_OBJECT_LABELS, allow setup metadata for each object during upload, use {macro_name} from system.macros and {backupName} for current backup name
  # The format for this env variable is "key1:value1,key2:value2". For YAML please continue using map syntax
  object_labels: {}
  # S3_CUSTOM_STORAGE_CLASS_MAP, allow setup storage class depending on the backup name regexp pattern, format nameRegexp > className
  custom_storage_class_map: {}
  # S3_REQUEST_PAYER, define who will pay to request, look https://docs.aws.amazon.com/AmazonS3/latest/userguide/RequesterPaysBuckets.html for details, possible values requester, if empty then bucket owner
  request_payer: ""
  debug: false                     # S3_DEBUG
gcs:
  credentials_file: ""         # GCS_CREDENTIALS_FILE
  credentials_json: ""         # GCS_CREDENTIALS_JSON
  credentials_json_encoded: "" # GCS_CREDENTIALS_JSON_ENCODED
  # look https://cloud.google.com/storage/docs/authentication/managing-hmackeys#create how to get HMAC keys for access to bucket
  embedded_access_key: ""      # GCS_EMBEDDED_ACCESS_KEY, use it when `use_embedded_backup_restore: true`, `embedded_backup_disk: ""`, `remote_storage: gcs`
  embedded_secret_key: ""      # GCS_EMBEDDED_SECRET_KEY, use it when `use_embedded_backup_restore: true`, `embedded_backup_disk: ""`, `remote_storage: gcs`
  skip_credentials: false      # GCS_SKIP_CREDENTIALS, skip add credentials to requests to allow anonymous access to bucket
  endpoint: ""                 # GCS_ENDPOINT, use it for custom GCS endpoint/compatible storage. For example, when using custom endpoint via private service connect
  bucket: ""                   # GCS_BUCKET
  path: ""                     # GCS_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""         # GCS_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  compression_level: 1         # GCS_COMPRESSION_LEVEL
  compression_format: tar      # GCS_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  storage_class: STANDARD      # GCS_STORAGE_CLASS
  chunk_size: 0                # GCS_CHUNK_SIZE, default 16 * 1024 * 1024 (16MB)
  client_pool_size: 500        # GCS_CLIENT_POOL_SIZE, default max(upload_concurrency, download concurrency) * 3, should be at least 3 times bigger than `UPLOAD_CONCURRENCY` or `DOWNLOAD_CONCURRENCY` in each upload and download case to avoid stuck
  # GCS_OBJECT_LABELS, allow setup metadata for each object during upload, use {macro_name} from system.macros and {backupName} for current backup name
  # The format for this env variable is "key1:value1,key2:value2". For YAML please continue using map syntax
  object_labels: {}
  # GCS_CUSTOM_STORAGE_CLASS_MAP, allow setup storage class depends on backup name regexp pattern, format nameRegexp > className
  custom_storage_class_map: {}
  debug: false                 # GCS_DEBUG
  force_http: false            # GCS_FORCE_HTTP
cos:
  url: ""                      # COS_URL
  timeout: 2m                  # COS_TIMEOUT
  secret_id: ""                # COS_SECRET_ID
  secret_key: ""               # COS_SECRET_KEY
  path: ""                     # COS_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""         # GOS_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  compression_format: tar      # COS_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  compression_level: 1         # COS_COMPRESSION_LEVEL
ftp:
  address: ""                  # FTP_ADDRESS in format `host:port`
  timeout: 2m                  # FTP_TIMEOUT
  username: ""                 # FTP_USERNAME
  password: ""                 # FTP_PASSWORD
  tls: false                   # FTP_TLS
  tls_skip_verify: false       # FTP_TLS_SKIP_VERIFY
  path: ""                     # FTP_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""         # FTP_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  compression_format: tar      # FTP_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  compression_level: 1         # FTP_COMPRESSION_LEVEL
  debug: false                 # FTP_DEBUG
sftp:
  address: ""                  # SFTP_ADDRESS
  username: ""                 # SFTP_USERNAME
  password: ""                 # SFTP_PASSWORD
  port: 22                     # SFTP_PORT
  key: ""                      # SFTP_KEY
  path: ""                     # SFTP_PATH, `system.macros` values can be applied as {macro_name}
  object_disk_path: ""         # SFTP_OBJECT_DISK_PATH, path for backup of part from clickhouse object disks, if object disks present in clickhouse, then shall not be zero and shall not be prefixed by `path`
  concurrency: 1               # SFTP_CONCURRENCY
  compression_format: tar      # SFTP_COMPRESSION_FORMAT, allowed values tar, lz4, bzip2, gzip, sz, xz, brortli, zstd, `none` for upload data part folders as is
  compression_level: 1         # SFTP_COMPRESSION_LEVEL
  debug: false                 # SFTP_DEBUG
custom:
  upload_command: ""           # CUSTOM_UPLOAD_COMMAND
  download_command: ""         # CUSTOM_DOWNLOAD_COMMAND
  delete_command: ""           # CUSTOM_DELETE_COMMAND
  list_command: ""             # CUSTOM_LIST_COMMAND
  command_timeout: "4h"          # CUSTOM_COMMAND_TIMEOUT
api:
  listen: "localhost:7171"     # API_LISTEN
  enable_metrics: true         # API_ENABLE_METRICS
  enable_pprof: false          # API_ENABLE_PPROF
  username: ""                 # API_USERNAME, basic authorization for API endpoint
  password: ""                 # API_PASSWORD
  secure: false                # API_SECURE, use TLS for listen API socket
  ca_cert_file: ""             # API_CA_CERT_FILE
                               # openssl genrsa -out /etc/clickhouse-backup/ca-key.pem 4096
                               # openssl req -subj "/O=altinity" -x509 -new -nodes -key /etc/clickhouse-backup/ca-key.pem -sha256 -days 365 -out /etc/clickhouse-backup/ca-cert.pem
  private_key_file: ""         # API_PRIVATE_KEY_FILE, openssl genrsa -out /etc/clickhouse-backup/server-key.pem 4096
  certificate_file: ""         # API_CERTIFICATE_FILE,
                               # openssl req -subj "/CN=localhost" -addext "subjectAltName = DNS:localhost,DNS:*.cluster.local" -new -key /etc/clickhouse-backup/server-key.pem -out /etc/clickhouse-backup/server-req.csr
                               # openssl x509 -req -days 365000 -extensions SAN -extfile <(printf "\n[SAN]\nsubjectAltName=DNS:localhost,DNS:*.cluster.local") -in /etc/clickhouse-backup/server-req.csr -out /etc/clickhouse-backup/server-cert.pem -CA /etc/clickhouse-backup/ca-cert.pem -CAkey /etc/clickhouse-backup/ca-key.pem -CAcreateserial
  integration_tables_host: ""  # API_INTEGRATION_TABLES_HOST, allow using DNS name to connect in `system.backup_list` and `system.backup_actions`
  allow_parallel: false        # API_ALLOW_PARALLEL, enable parallel operations, this allows for significant memory allocation and spawns go-routines, don't enable it if you are not sure
  create_integration_tables: false # API_CREATE_INTEGRATION_TABLES, create `system.backup_list` and `system.backup_actions`
  complete_resumable_after_restart: true # API_COMPLETE_RESUMABLE_AFTER_RESTART, after API server startup, if `/var/lib/clickhouse/backup/*/(upload|download).state2` present, then operation will continue in the background
  watch_is_main_process: false # WATCH_IS_MAIN_PROCESS, treats 'watch' command as a main api process, if it is stopped unexpectedly, api server is also stopped. Does not stop api server if 'watch' command canceled by the user.

[Slach]

thanks for reporting as workaround, you can change configuration

clickhouse:
  check_parts_columns: false

or use CLICKHOUSE_CHECK_PARTS_COLUMNS=false environment variable