e5.6.1

emqx/emqx

版本发布时间: 2024-04-18 20:15:43

emqx/emqx最新发布版本:v5.8.2(2024-11-12 22:20:49)

Bug Fixes

#12759 EMQX now automatically removes invalid backup files that fail during upload due to schema validation errors. This fix ensures that only valid configuration files are displayed and stored, enhancing system reliability.
#12766 Renamed message_queue_too_long error reason to mailbox_overflow

mailbox_overflow. The latter is consistent with the corresponding config parameter: force_shutdown.max_mailbox_size.
#12773 Upgraded HTTP client libraries.

The HTTP client library (gun-1.3) incorrectly appended a :portnumber suffix to the Host header for standard ports (http on port 80, https on port 443). This could cause compatibility issues with servers or gateways performing strict Host header checks (e.g., AWS Lambda, Alibaba Cloud HTTP gateways), leading to errors such as InvalidCustomDomain.NotFound or "The specified CustomDomain does not exist."
#12802 Improved how EMQX handles node removal from clusters via the emqx ctl cluster leave command. Previously, nodes could unintentionally rejoin the same cluster (unless it was stopped) if the configured cluster discovery_strategy was not manual. With the latest update, executing the cluster leave command now automatically disables cluster discovery for the node, preventing it from rejoining. To re-enable cluster discovery, use the emqx ctl discovery enable command or simply restart the node.
#12814 Improved error handling for the /clients/{clientid}/mqueue_messages and /clients/{clientid}/inflight_messages APIs in EMQX. These updates address:
- Internal Timeout: If EMQX fails to retrieve the list of Inflight or Mqueue messages within the default 5-second timeout, likely under heavy system load, the API will return 500 error with the response {"code":"INTERNAL_ERROR","message":"timeout"}, and log additional details for troubleshooting.
- Client Shutdown: Should the client connection be terminated during an API call, the API now returns a 404 error, with the response {"code": "CLIENT_SHUTDOWN", "message": "Client connection has been shutdown"}. This ensures clearer feedback when client connections are interrupted.
#12824 Updated the statistics metrics subscribers.count and subscribers.max to include shared subscribers. Previously, these metrics accounted only for non-shared subscribers.
#12826 Fixed issues related to the import functionality of source data integrations and retained messages in EMQX. Before this update:
- The data integration sources specified in backup files were not being imported. This included configurations under the sources.mqtt category with specific connectors and parameters such as QoS and topics.
- Importing the mnesia table for retained messages was not supported.
#12843 Fixed cluster_rpc_commit transaction ID cleanup procedure on replicator nodes after executing the emqx ctl cluster leave command. Previously, failing to properly clear these transaction IDs impeded configuration updates on the core node.
#12882 Fixed an issue with the RocketMQ action in EMQX data integration, ensuring that messages are correctly routed to their configured topics. Previously, when multiple actions shared a single RocketMQ connector, all messages were mistakenly sent to the topic configured for the first action. This fix involves starting a distinct set of RocketMQ workers for each topic, preventing cross-topic message delivery errors.
#12885 Fixed an issue in EMQX where users were unable to view "Retained Messages" under the "Monitoring" menu in the Dashboard.

The "Retained messages" backend API uses the qlc library. This problem was due to a permission issue where the qlc library's file_sorter function tried to use a non-writable directory, /opt/emqx, to store temporary files, resulting from recent changes in directory ownership permissions in Docker deployments.

This update modifies the ownership settings of the /opt/emqx directory to emqx:emqx, ensuring that all necessary operations, including retained messages retrieval, can proceed without access errors.