This page contains a list of features, changes, known issues and known limitations of the Nerve releases.
This version was released on February 23, 2023. Unless otherwise stated, all known issues and limitations of the previous version are still valid.
|Product features||Introduced Nerve Deterministic Node Automation (DNA), an automated way of deploying workloads to nodes.
Added a new workload type: Docker Compose workloads.
Nerve can now be run as a virtual machine.
|Management System||Added a deploy button for deployment of workloads from the workload version list.
Added a security feature to logging and monitoring options with OpenSearch.
|Local UI||Added the option for setting time zones.|
|Nerve Data Services||Added support for the structure data type in the OPC UA Client input.|
|Documentation||Added the Management System API documentation.|
Refer to What's new in 2.6.0? for more information on new features.
- Improved time synchronization between the host and the RTVM to reduce the time offset, in particular when the host is connected to a faulty NTP server.
- Main open source components have been updated to the latest versions.
- Improved the communication of the system through tooltips, system, error and feedback messages.
- Added a minimize button to the Nerve Connection Manager.
- Multiple UI/UX fixes, improvements and quality of life changes.
- Improved consistency of system behavior between the API and the UI.
- Added more characters that can be used in the label key and value input fields.
- Improved the handling and display of labels.
- Improved how the system handles remote connections when power or network connection to nodes is lost.
- The Nerve Connection Manager is now signed and does not cause any warnings in the installation process on Windows.
- Fixed an issue where an error from an OPC UA Server (error code 0x80000000) would cause the Nerve Data Services Gateway to be stuck in a restart loop.
- Fixed an issue where the Nerve Data Services Gateway was not specified for external ports.
- Fixed an issue where the Save button was not enabled in workload settings after a setting had been edited.
- Fixed an issue where a workload version could be created even though the required fields in the workload settings were not valid.
- Fixed an issue where Nerve Data Services were not listed among node capabilities.
- Fixed an issue where the Nerve Connection Manager required additional packages to work on Debian Bullseye and Ubuntu 22.04.
- Fixed an issue where multiple Virtual Machine workloads could be deployed even though resources were not sufficient.
- Fixed an issue where ports were falsely marked as occupied when trying to establish remote connections.
- Fixed an issue where the Nerve Data Services Gateway would stop sending data through the Azure IoT Hub Device output after a set period of time.
- Fixed an issue where volume paths of other Docker workloads would be suggested when defining volumes for a Docker workload.
- Fixed an issue where a Docker workload could not be deployed if the provisioning was cancelled during the download phase, edited and then properly saved.
- Fixed an issue where a notification in the Management System would cause the UI to freeze.
- Fixed an issue in the Nerve Data Services where the MQTT Subscriber input would not work if the ClientID parameter was not defined.
- Fixed an issue where sorting workloads by CPU or RAM usage in the node details workload list would not sort properly.
Known issues and limitations
- Virtual machine workloads using QCOW2 images cannot be deployed to nodes running on XEN hypervisor.
- Node serial numbers need to be entered with all capital letters when using the API. Using lower case letters causes erroneous behavior of the system.
- The status of a workload deployment is sometimes displayed incorrectly in the deployment log after a failed deployment is restarted and finished successfully.
- The progress bar of node updates only shows progress according to how many nodes have already been updated. It does not take individual update progress into account. So if four nodes are being updated, the progress bar only shows 0%, 25%, 50% and 75%.
- Selecting multiple nodes in the node tree while holding the Shift key causes unwanted duplicates and moving of nodes.
- Trying to upload a file with the wrong file extension when provisioning a workload only gives an error message once.
- Multiple Virtual Machine workloads on one node are started after a node reboot even if their combined resources exceed the limit.
- Remote connections and nodes containing special characters cannot be found in their respective searches.
- Retrying a failed deployment on an offline node results in the deployment log showing -100% progress.
- Adding a Nerve Device through the API with an invalid device model parameter results in the Nerve Device being added as an MFN 100.
- Restarting a failed deployment might lead to conflicting progress status in the deployment log and the deployment details.
- When replacing workload versions, versions containing labels are listed even if deploying them would fail due to label settings.
- When connecting through a remote screen, a message saying Disconnected briefly appears, even if the connections has not been lost.
- Progress percentages shown in dry runs include retries, making percentages seem inaccurate.
- The back button of a browser needs to be used twice on Management System pages containing tables to go back to the previous page.
- The permission to create workloads does not give users the permissions to configure remote connections.
- Some windows and tables might appear with two scrollbars depending on size and resolution.
- The deployment of Docker workloads that were provisioned with uploaded TAR.GZ image files might fail on nodes older than version 2.3.1.
- The deployment of Virtual Machine workloads that have more than three network interfaces might fail.
- In the Management System, the system info lists the current version of the Management System among available updates.
- System notifications might pop-up for one user if another performs changes.
- RDP connections to workloads sometimes fail even though they were correctly configured.
- When self-signed certificates are used, the Cloud URL field check in the node configuration on the Local UI might fail even if the URL is correct. The node can be registered regardless.
- The Nerve Connection Manager is recognized as a threat by some antivirus software. This can be considered a false positive. Take this report into consideration when installing the Nerve Connection Manager.
- Docker Compose workloads that are marked as released can be deleted through the API.
- Removing workloads that were successfully deployed to a node through Nerve DNA while reconfiguration is in progress will result in the node wrongly entering the Applied state instead of the Modified state.
- There is a chance that remote tunnel connections are not established on the first try. If that happens, trying again will establish the connection successfully.
- The node might crash if a workload is deployed from the Management System UI while Nerve DNA reconfiguration is in progress.
- Non-official time zone formats can appear when a time zone is configured for nodes.
- When applying a target configuration through Nerve DNA fails, an error message in the logs refers the user to check the logs for more information. This is false as there is no additional information provided in logs.
- If workload details contain a number in name or version, e.g.
version: 1, the configuration file will not be valid. Escape this by using quotation or annotation marks, e.g
- Downloading Gateway logs fails once the file has grown too large.
- Sorting the list of active remote connections in the Remotes menu of the Management System does not work properly.
- After updating a workload that has an open remote tunnel connection, the status of the remote tunnel is wrongly displayed as green in the Nerve Connection Manager.
- Reducing the window size of a remote VNC screen and increasing the size again leaves the displayed screen at a small size.
- Remote connections can be stuck in connecting phase when network speed is slow.
Scaling and performance limitations
This release has been tested to perform within the following scaling boundaries:
|Maximum number of concurrent devices||200|
|Maximum number of concurrently logged in users||5|
|Maximum workload size||50 GB|
|Maximum number of concurrent remote access sessions||3|
|Maximum number of workloads in workload repository||200|
|Maximum data upload per node||5 datagrams per second with at least 10 sensor values for 200 nodes in parallel.|