Top.Mail.Ru
Журнал изменений | Tdg
 
Журнал изменений
Журнал изменений

Журнал изменений

Журнал изменений

Формат журнала основан на Keep a Changelog, версионирование подчиняется правилам Semantic Versioning.

[1.6.16, 1.7.7] - 2021-05-12

Added

  • Introduce synchronous mode for output processor (tarantool/tdg#688)
  • Allow configure metrics via config (tarantool/tdg2#1001)
  • Handle dependencies on library initialization (however, we still recommend organizing library code to eliminate any dependencies)
  • Update cartridge to d36ebf2
  • SMTP module 0.0.6 (tarantool/tdg2#708)
  • Metrics 0.8.0
  • ODBC module 1.0.1

Fixed

  • “Show more” button on logger page doesn’t return an error (tarantool/tdg#712)
  • Topology change should not affect data access (tarantool/tdg2#934)
  • Don’t show host twice in log message in case of error

[1.6.15, 1.7.6] - 2021-04-26

Added

  • Drop unzip/zip requirement (tarantool/tdg#677)
  • Allow to upload config without HTTP - admin.upload_config_api (tarantool/tdg2#272)
  • Cartridge with fixed stateful failover bug
  • Enterprise bundle 2.7.2-0-g4d8c06890-r399

Fixed

  • Error if uploaded config was invalid zip file (tarantool/tdg2#898)
  • Invalid cursor calculation if trailing nullable fields were part of index (tarantool/tdg#689)
  • Error if null value was passed to some repository.find filters (tarantool/tdg#691)
  • Change :auto_increment to sequence for to_output_processor space to avoid replication conflicts (tarantool/tdg#686)
  • Extend RBAC to support all modern cartridge queries and mutations (e.g. force config reapply) (tarantool/tdg2#919)
  • Revoke access rights to logs from user (affected only new installations)
  • Stop periodical tasks if they are removed from config (tarantool/tdg#699)
  • Rename storage_error to repository_error (tarantool/tdg2#933)
  • Scheduler can’t be run inside multiple replicasets. Add check for singleton role (tarantool/tdg#707)
  • Don’t start to read messages from kafka if token name is invalid (tarantool/tdg#644).
  • Fix disabling guest role on read-only instances

[1.6.14, 1.7.5] - 2021-03-30

Added

  • Cartridge with updated frontend-core (to be able to reload fronted modules)

Fixed

  • OperationError in case options defined for kafka producer
  • Error on user reimport (tarantool/tdg2#419)

[1.6.13, 1.7.4] - 2021-03-23

Added

Fixed

  • Enable coredumps when watchdog aborts Tarantool process
  • Spam into logs on storage startup
  • Incorrect log messages when output processing failed
  • Free resources after kafka/smtp output is removed from config (tarantool/tdg2#816)
  • Unclear vshard errors (tarantool/tdg2#384)
  • «OperationError» if empty config section is uploaded (tarantool/tdg#647)
  • Replace insert with replace in repair storage spaces to avoid replication conflicts (tarantool/tdg2#758)
  • Unique secondary index by time in repair buffer (tarantool/tdg#638)
  • Error if defaults are specified for unions (tarantool/tdg2#118)

[1.6.12, 1.7.3] - 2021-01-26

Added

[1.6.11, 1.7.2] - 2021-01-11

Fixed

Added

  • Cartridge 2.4.0
  • Updates via GraphQL API (tarantool/tdg2#575)
  • Add version compatibility option to config (tarantool/tdg2#431)
  • Deletion of removed indexes from model (tarantool/tdg2#367)
  • Luarapidxml 2.0.1
  • Kafka 1.3.0 (tarantool/tdg2#596)
  • SMTP 0.4.0
  • Metrics 0.6.0
  • Enable CPU usage metrics (tarantool/tdg2#442)
  • Extend datetime sandbox API with custom_datetime_str_to_nsec, millisec_to_formatted_datetime functions
  • Add “soap” module to sandbox
  • Enterprise bundle 2.4.3-0-g5180d98f1-r370.

[1.6.10, 1.7.1] - 2020-12-03

Fixed

Added

[1.7.0] - 2020-11-10

Fixed

Removed:

  • Frontend role
  • Permissions. Use data actions instead

Breaking changes

  • Use ‘autorization’ header (Bearer Authentication) instead of ‘auth-token’

[1.6.9] - 2020-10-20

Fixed

  • Kafka: handle error during consumer closing
  • Fix “field is required by space format is missing” on new roles create

Added

  • Metrics 0.3.0
  • Checks 3.1.0
  • ODBC 0.7.2

[1.6.8] - 2020-10-01

Fixed

Added

  • ODBC 0.7.1

[1.6.7] - 2020-08-20

Fixed

[1.6.6] - 2020-08-05

Fixed

Added

[1.6.5] - 2020-07-15

Fixed

Added:

[1.6.4] - 2020-07-02

Fixed

[1.6.3] - 2020-06-05

Fixed

Added

[1.6.2] - 2020-05-21

Fixed

[1.6.1] - 2020-05-13

Fixed

Added

[1.6.0] - 2020-04-20

Fixed

Added

[1.6.0-alpha] - 2020-04-20

Added

Fixed

  • remove check existence of task_runner when storage init (tarantool/tdg#135)
  • [dev] implement ignore_hard_limits flag for document.tuple_select for some service/development purposes. It calls fiber.yield every force-yield-limit iterations. Can be configured via GraphQL, by default 1000
  • tdgctl: support new option is_master (boolean) for deploy config. The upgrade procedure will only run on such instances
  • decorate an archive name with instance name (tarantool/tdg#183)
  • do not treat responses with code 201, 204 as an error (tarantool/tdg#187)
  • tdgctl: upgrade automatically stops all instances
  • the only request context pass through netbox now (tarantool/tdg#189)
  • repository update for value object case (tarantool/tdg#259)
  • wrong (but without affect) bucket_id calculation in repository.find (tarantool/tdg#266)
  • prohibit the ability to log in if authorization is turned off (tarantool/tdg#277)
  • document: allow to specify multikey index as index check (tarantool/tdg#310, tarantool/tdg#311)
  • uploading configuration with a section removed does remove it clusterwide (tarantool/tdg#314)
  • add cursor field to repair storage get method (tarantool/tdg#344)

Removed:

  • batch_count option from map_reduce

Breaking changes

  • format of the return value in output_processor changed (tarantool/tdg#178)
  • map-reduce interface was updated (tarantool/tdg#192)
  • GraphQL maintenance_get_aggregates function renamed to maintenance.get_aggregates
  • optimistic lock semantic is changed (tarantool/tdg#283)

[1.5.0] - 2019-09-18

Added

  • default value for token_acl list
  • checking permissions by token
  • authorization with token name
  • default user for tasks
  • option to find deleted records in model_accessor
  • system tasks
  • ODBC support in sandbox
  • Kafka consumer in connector
  • request_id prefix for log messages (tarantool/tdg#541)
  • expose tracing into sandbox
  • funcall refactoring (tarantool/tdg#575)
  • Fix: UI: Permission filter fix
  • Fix: rename scheduler api because of clashing with cluster logic
  • Fix: call to utils datetime conversion in model_updater
  • Fix: call task_list API only from master
  • Fix: rename common.tracing role to tracing
  • QA: enable Lua and Python linters
  • UI: show asterisks on password fields (tarantool/tdg#561)
  • UI: logger lamp disable (tarantool/tdg#564)
  • UI: tracing flag for GraphQL and test sections (tarantool/tdg#546)
  • repository.call_on_storage (tarantool/tdg#49)
  • sandbox function this_storage.snapshot
  • connector SMTP output feature (tarantool/tdg#5)
  • delegate arguments parsing to cluster.argparse (tarantool/tdg#93)
  • rename cluster to cartridge (tarantool/tdg#95)
  • Fix: output replication on entity update (tarantool/tdg#61)
  • add permanent_delete option to repository.delete
  • expose if_only_version option to GraphQL mutations
  • GraphQL query time and fail monitoring (tarantool/tdg#101)
  • Add response_body and status_code to failed requests from output_connector (tarantool/tdg#75)
  • Functions refactoring (tarantool/tdg#131). In functions you must get arguments like local x, y = ... instead of local x, y = unpack(...) or using array semantics
  • Tarantool version updated to 2.2

Fixed

[1.4.1] - 2019-09-06

[1.3.0] - 2019-05-31

[1.2.0] - 2018-09-29

Added

  • monitoring a replication lag
  • logger pagination
  • profile the pipelines
  • fixed systemd start service on boot
  • save backup files before deployment
  • ‘try again all’ button in the repair queue
  • singleton mechanism of instance
  • migrated to new logger
  • GraphQL refactoring
  • break the build if any module is missing
  • check deploy permissions
  • download logs with tdgctl
  • add rm data command to tdgctl
  • display cluster memory usage in UI
  • vshard call timeout
  • deploy the project on the bare metal
  • migrate Docker image to static Tarantool and CentOS 7
  • add datacenter name information to every node
  • web console
  • add cluster cookie to deploy script
  • cluster: added GraphQL self query

[1.1.1] - 2018-07-10

Added

  • privileged mode
  • keep DDL in sync across storage nodes
  • simple deployment script based on Docker and Fabric
  • wrap net.box.call with errors

[1.1.0] - 2018-07-01

Added

  • authentication and external app token authentication
  • port repair queue to vshard
  • notifier as separate role
  • collect messages before sending
  • collect metrics from dcos deployment to Prometheus
  • multipart index
  • email notifications about repair queue events

[0.3.0] - 2018-02-28

Added

  • support for sharded storage
  • proof-of-concept of dynamically adding/replacing replicas of shards, using the ‘shard’ module and docker as ‘backend’
  • Avro-based modeling language
  • fixed large GraphQL requests parsing

[0.2.0] - 2018-01-23

Added

  • Prometheus-based monitoring
  • “on insert” subscriptions
  • JSON output adapter

[0.1.0] - 2017-12-20

Added

  • Initial version of connector with SOAP input and Tarantool protocol output
  • Data processing pipelines based on Lua code
  • Schema description language loosely based on graphql-lua
  • Ability to dynamically reload schema
  • Schema verification for incoming objects
  • Basic same-node storage backend
  • GraphQL query API based on the schema
  • Frontend for testing GraphQL queries
  • Frontend for editing schema
  • Frontend for testing SOAP requests