Known Issues in EMQX 5.1

e5.1.1

Inaccurate Error Message Displayed After Successful Rolling Upgrade
Following a rolling upgrade, the emqx.log might present an error message. Nonetheless, this error message is harmless, and the upgrade has indeed succeeded.
When performing a rolling upgrade on a core+replicant cluster, you may observe the error message ** ERROR ** Mnesia post_commit hook failed: error:badarg in the logs. However, it is important to recognize that these errors do not affect the success of the upgrade process, and the cluster can be upgraded without any issues.

e5.1.0

Duplicate "Connection Pool Size" Setting in MongoDB Bridge UI
The MongoDB Bridge UI displays duplicated entries for the "Connection Pool Size" parameter. The second instance is utilized for configuring MongoDB connection options, leading to confusion and potential configuration errors.
Fixed Version: 5.1.1
TimescaleDB Bridge Status Remains Disconnected after Successful Creation
When establishing a password-free login for TimescaleDB, the data bridge is created successfully but remains disconnected. Inputting the username during bridge creation resolves this issue.
Workaround: Configure usernames and passwords for TimescaleDB, then create a data bridge using the correct credentials.
Fixed Version: 5.1.1
Listener Status Not Updated on Dashboard After Command Line Start/Stop
Starting or stopping a listener through the following commands does not update the listener's status on the Dashboard, causing confusion about its actual state.
```
./bin/emqx ctl listeners stop tcp:we
./bin/emqx ctl listeners start tcp:we
```
Plugin Start/Stop Command Only Affects Executing Node, Not Cluster-Wide
Executing the start/stop commands for a plugin as shown below affects only the executing node, not other nodes within the cluster.
```
./bin/emqx ctl plugins stop emqx_plugin_template-5.0.0
./bin/emqx ctl plugins start emqx_plugin_template-5.0.0
```
Workaround: Execute the start/stop command on all nodes in the cluster or use the Dashboard to manage the plugin.
Authentication Required for Coap Gateway Connection Mode
The CoAP gateway only allows connection creation with authentication enabled. Enabling any authentication method resolves the issue.
Workaround: Enable any authentication method and create a connection with the correct authentication information.
Fixed Version: 5.1.1
Dashboard Error When Using "crt" for "Use Peer Certificate field as ClientId" Option
Selecting crt as the option for Use Peer Certificate as Client ID on the General tab of the MQTT Settings page on the Dashboard causes the client ID to display garbled code on the Dashboard and leads to an error when viewing details.
Unable to View and Download Files on Replica Nodes in Cluster with File Transfer Feature
When using the file transfer function to upload files within a cluster, files uploaded to one of the replica nodes can only be accessed with files list API from the node to which they are uploaded.
Workaround: Only view and download files from the node to which the file is uploaded.
Fixed Version: 5.1.1
Dashboard Fails to Reflect Modified Listener Port
Adjusting the listener port in the emqx.conf configuration file as shown below does not update the port displayed on the Dashboard, leading to a mismatch between the displayed and actual port.
```
listeners.tcp.default {
    bind = "0.0.0.1884"
}
```
Fixed Version: 5.1.1
"Max Connections" Parameter Ineffective for "quic" or "ws" Listeners
Configuring "Max Connections" for "quic" or "ws" type listeners through the Dashboard or configuration file does not take effect.
Inaccurate client.disconnected Events Statistics in Rules
Channels with clean_session = false generate two client.disconnected events, leading to inaccurate event statistics.
Parsing Issue with UTF-8 Characters in Rule Action "republish" Payload
When creating a rule with the "republish" action and using ${payload.'msg'} in the payload, clients subscribed to the republish topic receive ${payload.'msg'} instead of the actual ${payload.msg} value.
Fixed Version: 5.2.0
Default Value of "Connection Pool Size" for Oracle Database is Unchangeable
Irrespective of the specified value, the "Connection Pool Size" parameter remains fixed at the default value of 8.
Fixed Version: 5.1.1
Fail to Update for Subscribed Topic's QoS on Dashboard
When a CoAP client subscribes to a topic with QoS 0 and changes the QoS to 1, the Dashboard still displays the original QoS value and the actual QoS remains at 0.
Log Files Deleted on EMQX Restart with Rotation Number > 10
When the log rotation_number is set to a value greater than 10, the log files with numbers higher than 10 (e.g., emqx.log.11, emqx.log.12, etc.) are still getting deleted upon restarting EMQX.
EMQX Cluster Split-Brain Issue when Node Disconnected for >2 Minutes
EMQX 5.x upgrade to OTP25 has encountered a challenge due to the implementation of a more aggressive network fragmentation approach in Erlang OTP 25. Unlike the previous OTP 24 version, even minor network disturbances can now inaccurately trigger split-brain detection in the global component of EMQX.
Fixed Version: 5.2.1