OpenThinRDP Manager 1.1.7

OpenThinRDP Manager 1.1.7: anonymous data at three levels, for itself and for the devices of OpenThinRDP 0.3.3.

• Full, Errors only or Off. Settings > Manager > Anonymous data for the project chooses how much the manager sends: Full once a day (its version, Windows, how many devices in ranges, the size of what it reads and the pages and actions used, as counts), Errors only only on a day with errors (how many, never what they say), Off nothing. A manager that sent the anonymous counts before sends Full.
• For the devices. The Anonymous data for the project setting of a site, a group or one device has the same three levels as OpenThinRDP 0.3.3 (ANON_STATS = full, errors, off).
• Where it goes. To PostHog in the EU, with a random id for each send, no person profile and no IP address kept. privacy.md lists every item.

The manager updates itself to this release.

OpenThinRDP Manager 1.1.6

OpenThinRDP Manager 1.1.6: the settings of the site and of a group are back, and the new settings of OpenThinRDP 0.3.2.

• Settings of the site and of a group. On the Devices page, Settings of the site or of a group opens the settings that all its devices get (with Managed here), with Devices to go back. Since the web window it only showed the list of devices again. All the settings of the devices in Settings opens them too.
• A computer on another port. The remote computer or broker takes name:port, the fallback brokers and the other computers a list of them, as devices with 0.3.2 do.
• Screen of the devices. Two new settings for OpenThinRDP 0.3.2: Screen resolution (native or width x height) and Screen scale (auto, 100-200%).

The manager updates itself to this release.

OpenThinRDP Manager 1.1.5

OpenThinRDP Manager 1.1.5: the memory of each VM as the hypervisor really maps it, and several VMs at once.

• Memory of the VMs: large and small pages. The bar of each VM now shows how its memory is mapped by the hypervisor: green the large pages (2 MB, 1 GB), red the small 4 KB pages, which make every memory access of the VM slower. A new column gives the share of small pages (marked from 20%, red from 50%) and the window opens on the worst VM first. Why it matters: on a real site, desktops with dynamic memory had 41-93% of their memory in small pages after a week of uptime, VMs with fixed memory 0-2%. A VM shut down and started again starts clean.
• Actions on VMs side by side. Starting, shutting down or restarting a VM, or signing a user out, now holds only that VM: the other VMs and sessions can be acted on while it runs. Actions on a whole broker, host or node still run alone, and wait until no VM action is running. A VM action that succeeds says so in a short note instead of a window to close.
• Several VMs at once. Desktops > "Shut down VMs..." and "Start VMs...": tick the VMs (each listed with who is signed in), confirm once, and they all run together, each result as it comes.
• JEA: if you use the JEA endpoint of the manager, make it again from this version (the reading of a host changed).

The manager updates itself to this release. The pages of each VM appear with the first reading made by 1.1.5.

OpenThinRDP Manager 1.1.4

OpenThinRDP Manager 1.1.4: the processors of a Hyper-V host as they really are, and the use of each VM one click away.

• Processors of a host: until 1.1.3 the manager showed what Windows of the host itself uses (the root partition), which stays near 0% however busy the VMs are. It now reads the hypervisor: the whole machine, split into its VMs and Windows of the host. Readings below 10% show one decimal (1.5%, not 0 or 2).
• The use of each VM: in Infrastructure, click Memory or Processors of a host (or "Use of the VMs" on the card).
  • Processors: every VM with a bar from 0 to 100% of its own virtual processors, and its share of the whole machine; a VM at 80% or more is marked.
  • Memory: every VM with its assigned memory as a bar: green the fixed part (its startup memory), orange what dynamic memory added, the whole bar its maximum. The memory Windows in the VM asks for is shown beside it; a VM asking for more than it has is marked.
• The busiest VM is named under the processors of the host.
• JEA: if you use the JEA endpoint of the manager, make it again from this version (the reading of a host changed).

The manager updates itself to this release. The figures per VM appear with the first reading made by 1.1.4.

OpenThinRDP Manager 1.1.3

OpenThinRDP Manager 1.1.3: the page of the devices works again.

• Devices: in 1.1.2 the detail of a device was not drawn (the page kept showing the one before). Fixed.
• Behind the scenes, so that it does not happen again: every change of the manager is now checked as code (ESLint on the page, PSScriptAnalyzer on PowerShell) and every page, tab and window is drawn and clicked in the real window, as a user does, before a release. These checks fail on 1.1.2 with exactly this error.

The manager updates itself to this release.

OpenThinRDP Manager 1.1.2

OpenThinRDP Manager 1.1.2: exact numbers for a cluster that loses a node, ignored notices that really step aside, and every page checked on real data.

Room if a node fails

• Counted VM by VM, as the cluster does it: each VM restarts whole on one other node (by its startup memory), with 3 or 4 nodes they spread over the others, 2 GB a node are kept for Hyper-V, a node paused or down takes none. The same count on Home, in the Room for a failure check (now node by node in its facts) and before draining a node. Before, the free memory of all the nodes was added up, the free memory of the node that fails included.
• Home: Spare memory if a node fails, 30 days, below zero when the VMs of a node would not all restart, saved at each reading; the forecast only after the last jump in the data.
• Trends of memory and disks: a jump while the manager was closed (overnight) is no longer read as a trend.

Ignoring a notice

• An ignored notice leaves the counts, the score and the colours, with a state of its own (ignored), not maintenance.
• Until it changes: it comes back when what it says changes, not when a counter goes up; the hardware log comes back with a new event.
• The list of ignored notices says which ones still hide something, since when, and has a Restore button on each row.
• Fixed: with two notices ignored the list was saved wrongly and the next Ignore failed.

Every page and window

• Every action of a right click also behind a ⋯ button on each row.
• The tree of the Health Center shows the item that is open; device names whole and in natural order (2 before 10); versions newest first.
• Experience: the three tables fit their space. Groups: the devices in no group are listed.
• Version of a device: with no saved version the choice says so and how to get one (Apply did nothing).
• Update hours: empty when not set (an example looked like the setting).
• Settings, custom checks and notes with accents read correctly; numbers in the singular; temperatures with °C.

Feedback

• Once a week, until it is sent, the manager asks how it is going: a vote and what it is used for go as anonymous counts; the words open a GitHub discussion already written, which you post or not. Also from Settings > Manager. See privacy.

The manager updates itself to this release.

OpenThinRDP Manager 1.1.1

OpenThinRDP Manager 1.1.1: fixes.

• The window also when the manager starts from a network share. Started from \\server\share, .NET refused to load the WebView2 engine from the network, and the manager opened its classic window without saying so (no Ignore on Home, for one). The engine is now copied once per version to your profile and loaded from there.
• Time servers: a server with several time servers set (for example ntp1 and ntp2) is no longer reported as Changed from what was approved each time Windows moves from one to the other. The Health Center compares the list of the servers set, not the one in use at that moment; an approval made before this release is updated by itself.
• Ignore until it changes: a new choice when ignoring a notice. It comes back as soon as the notice says something new, such as a new critical event in the hardware log, so a fault you have already fixed does not stay on Home for 30 days.
• The help of the Browser setting matches OpenThinRDP 0.3.1 (no browser in the image; only the one chosen, and only with a web page).

The manager updates itself to this release.

OpenThinRDP Manager 1.1.0

OpenThinRDP Manager 1.1.0: the settings of OpenThinRDP 0.3.0, and two tools for a Hyper-V host.

OpenThinRDP 0.3.0

• Environment of a device: Active Directory domain or Workgroup (SETUP). The settings that belong to the other environment are hidden, on the site, on each group and on each device.
• The new settings: other computers a device offers at sign-in, remembered password, Wake-on-LAN (the MAC address of each device's computer, and how long to wait), the live stick that keeps its settings, the browser of your choice (Chromium, Firefox ESR, Brave, Google Chrome, Microsoft Edge), Tailscale as a VPN.

Tools on a Hyper-V host

• Checkpoint of a virtual machine, from the host's row.
• New virtual machine from an ISO: Generation 2, fixed memory, virtual TPM, Secure Boot, first start from the DVD; started at once if you ask. The quick way to try OpenThinRDP, or to make a test machine.

The manager updates itself to this release.

OpenThinRDP Manager 1.0.1

OpenThinRDP Manager 1.0.1: faster, no grey buttons without a reason, and what 1.0.0 got wrong on a real site.

Faster

• The window no longer stops while the servers are read: the model of the Health Center is built in the background (it took up to 2 s of the window after every reading and every part of one).
• During a reading only its progress bar moves; the pages are drawn again only when new data come. Home goes from about a second to under 0.2 s.
• The history of the Health Center is indexed and read once per day: Home no longer gets slower with every week of history.

Buttons and actions

• No grey button with nothing to do: a button that cannot do anything now (resume a node in service, start a VM already on, nothing to release) is not shown.
• Restart / shut down a cluster node in service: it drains the node first (with the checks of a drain), restarts it, waits for it and resumes it with its VMs.
• Restart the SQL primary: a planned failover first, then the restart. Restart the only broker: allowed, with a plain warning.
• Maintenance, VM backup, backup logs, FSMO guide and Approve were greyed by mistake; the button that trusts the new certificate of an iDRAC was missing.
• New Tools on every server, host and VM: PowerShell on it, Remote Desktop, the console of the VM, its C$, its events, Computer Management.

Diagnosis

• The hardware check says what the log of the iDRAC says (for example Comm Error: Riser 1); the hardware log is a tab of its own; the card of the checks can be enlarged.
• A network card a host no longer sees at power on (a riser, a card) is an error, and the link down on the other node is said to be the direct link to it, not a cable.
• Refused SQL sign-ins: who, from which computer, why and when, over 24 hours.
• Experience: Where the time goes, per broker, Hyper-V host and desktop, with the longest phase; the slowest sign-ins show their broker, desktop and host.
• The memory trend starts again after a restart or a drain (no more "full in 28 days" after a reboot).

Fixes

• Updating no longer ends with "the manager closed before opening" when the launcher has a .args with -Folder or -Language.
• VM backup window, group rules, JEA list, the desktop of a user, a scheduled job (fields shown only when they apply), lists of servers on a narrow window, a Refresh button on Home.

Updates from 1.0.0 by itself (Updates > This manager). From 1.0.0 the update shows the old closing error one last time if the launcher has a .args: the new launcher fixes it.

OpenThinRDP Manager 1.0.0

The first release of the manager with releases of its own, numbered apart from the devices. It updates itself from here from now on; managers up to 0.2.27 get here through OpenThinRDP 0.2.28 (two updates in a row, by themselves).

• Faster readings, shown as they come: servers, iDRACs and the rest are read at the same time, the iDRACs with one Redfish session each (a whole site went from 3 minutes to under one), the hosts in bulk. The last reading is shown at once when the manager opens; the new one goes into Home and the Health Center part by part, and its progress is at the top of Home.
• The Health Center no longer stays empty when a number goes down very slowly (the days until it ran out were too many for the reading). A trend says how many days it was measured over, and for the memory of a Hyper-V host it names the usual cause: the ReFS cache of the shared volume.
• A changed iDRAC certificate (a reset iDRAC makes a new one) is said as such, with both thumbprints, and Trust the new certificate takes it after a confirmation.
• No long .NET error when the window closes after Update the manager. An error nobody caught is written to %LOCALAPPDATA%\OpenThinRDP\manager-errors.log and said in a few words.
• Anonymous counts: once a day instead of once a week, one count per send (privacy).

Signed with the project key (LATEST.txt.sig, and MANIFEST.txt.sig inside the zip); the launchers carry the Authenticode signature CN=OpenThinRDP Project.