10-01-2018 07:01 AM - last edited on 10-05-2018 03:07 AM by UBNT-Radek
UNMS version 0.13.0 has been released, see the changelog below. For detailed installation and update, instructions view our guide: Installation & Update. Please, continue to share your thoughts and requests.
- It is recommended to create a backup of your UNMS data. Here is a guide on how to do it.
- Unfortunately, you will not be able to keep your old device log records after upgrading to 0.13.0. The data layer had been completely changed and moved to PostgreSQL. There was no possibility to make such a huge change backward compatible.
- If you are using the mobile version of UNMS, you need at least version 2.7.1 for Android and 2.2.3 for iOS.
- Topology map: The most important feature of the 0.13.0 release by far. The core of this feature is to map all data links in your network thus creating a network topology map. In this release, data links are created automatically between wireless and fiber devices, but ethernet connections need to be added and edited manually.
- Ability to recognize 3rd-party devices and include them to the topology map: This is a critical part of creating the complete topology map. The info about 3rd-party devices is very limited by design.
- New device discovery tool: With this release, device discovery can search devices in public ranges, thanks to support for TCP protocol. The tool will show all devices it could find (it will show all devices that replied to the discovery ping) and it will download basic info about each device through UBNT discovery protocol and SNMP v1.
- Performance upgrade: This is a big performance leap as we’re introducing support for multicore processing and multiple Node.js workers with load balancing capability. UNMS will automatically set the number of workers according to the server performance. At least two workers are used in all cases since one is needed for API and the other for devices connectors. It is possible to override the number of workers in the installation command by changing the value of --workers attribute, it is also possible to change it through CLI. We urge you to never set more workers then there are cores in your server and ideally there should be at least 0.5GB of memory for each worker.
- Device maintenance mode: With this feature, you can mark any device to stop contributing to the overall network health and from generating outage alerts.
- NetFlow support: NetFlow will not only measure data throughput to all devices but also recognize unknown IP addresses in a network and use that information to create a Third-party device. There is a new help article with more detailed info about NetFlow.
- Wireless config of airMAX devices: please be very cautious when using this feature to update device configuration.
- Smart backups: If you create a backup of a device it now always contains the correct AES part of the UNMS key, it is also possible to download a backup with generic UNMS key, which is important if you want to easily switch faulty HW for the same model.
- Full compatibility with the latest OLT FW 3.0.0.
- Full compatibility with the latest airCube FW 2.4.0.
Known issues and missing features:
- In .csv imports clients sometimes don't pair correctly!
- The support for airFiber devices has to move in 0.13.1 unfortunately. We found some bugs on the UNMS side and we need more time to fix them.
- The Custom IP feature will be available in 0.13.1 as well.
- There is an issue when OLT with FW 3.1 is to be connected. We will fix this in 0.13.1.
- Outage email alert sometimes doesn't reflect custom settings. Emails are being sent earlier and more frequently even for shorter outages. It will be fixed in 0.13.1.
- We had to pause the development of the Windows installer for the time being. Even though we invested a lot of time and resources into it, we were not satisfied with the reliability, ease of use and the performance of the solution. We felt too many compromises had to be made in order to install UNMS on Windows. For those reasons, we decided to focus the capacity of our DevOps developers into the building of our cloud-based UNMS, which will significantly reduce the time we need to finish it.