The following release notes describes the issues resolved in this revision of the Leostream Platform.
- The Last Connect Time on the > Resources > Desktops page is now updated only when the Leostream Agent returns a valid username in the connection notification. This change improves the accuracy of the Last Connect Time field, except for the following two cases.
- Leostream Agents installed on Windows 7 machines may not return usernames in the connection notifications, resulting in the Last Connect Time not being updated.
- Leostream Agents installed on VMware templates that were not properly shutdown prior to creating the template return connection notifications when new machines are provisioned from the template, resulting in the Last Connect Time being updated. Ensure that you shut down virtual machines prior to creating templates to use with Leostream.
- Connection Broker 9.1 enables the RESTful API, by default, and therefore requires additional RAM. Ensure that your Connection Broker has at least 8GB prior to upgrading or installing.
- Version 7.4.13 and higher of the Leostream Agent for Windows Operating system and 5.2.10 and higher of the Leostream Agent for macOS and Linux now distinguish between older versions of RGS and newer versions of HP ZCentral Remote Boost. If you define Pools based on the Installed Protocols attribute having a text value of RGS, edit the pool to include machines with Boost installed, as well, to ensure desktops running all versions of the protocol appear in the pool.
- The TGX encryption tool has been updated to version 2022.1.4
- Minor enhancements when defining multiple SAML-based authentication servers using tenants
- Improve handling of center scans that run while there are active provisioning jobs, to avoid removing desktop records associated with virtual machines that are still being created
- Resolve cases where provisioning jobs for virtual machines on Scale Computing HC3 resulted in duplicate desktop records
- Paginate inventory of images from OpenStack centers to handle cases where the OpenStack project has more than 25 images
- Properly log out rogue users that are connected to desktops using HP ZCentral Remote Boost (RGS) based on the Policy option to log out rogue users
- Change desktop power statuses to Stopped for desktops marked as Stopping when a Center scan occurs and confirms that the desktop is powered off
- Improve disconnect logic for PCoIP Remote Workstation Cards to attempt to disconnect the PCoIP session from the Remote Workstation Card if the PCoIP client cannot be disconnected
- Pass the region to the AWS API when checking if an AWS center is online
- Properly fail over to the second value for dynamic tags that use the or syntax and reference hostnames or IP addresses, for example IP_PUBLIC-or-IP
- Improve management of Remote Boost sessions on Rocky Linux 8
- Resolve issue related to the use of NetworkInterface API that could cause the registration with the Connection Broker to fail