Hi Ylian, I just installed MeshCentral (AWS instance) and it's looking quite amazing so far :-) I followed your instructions from your web page to the point, came upon a few issues with the doc (for example, the pkill command as noted in the stop script is incorrect, the systemd script was somewhat malformed (line-breaks), at least when you copy paste from the PDF and the "node" command is not present in the /usr/bin directory. I created a symlink to the actual bin and it worked great from then on), but was able to get through them quickly.
Installing the agent on a linux box (actually the MeshCentral AWS instance itself) went w/o issues as well as the Windows agent install.. BUT:
Question:
Whew.. sorry for bombarding you with all these questions/notes. I really like your work and would love to make this my standard access tool.
Hi. Thanks for the questions. I will do my best to answer everything, but I am sure I will miss things. In any case, here goes...
For any documentation issues, there is a document folder in GitHub that is in Microsoft Word format. If you can make the right edits and send the document back, that would be wonderful. Do put change tracking on.
The MeshAgent can be run in "interactive mode" when you just run it temporary and when you close the window it disconnects or it can be installed in the background, in that case it currently does not have any icons. If you opt to enable use consent features on the server, you can prompt to ask for permission to take control, notify when connecting and show a privacy bar at the top of the screen.
Let me know what features your looking for that you don't see. Server v0.4.5-k is the latest one right now, the agents are auto-updated when you update the server, so they are sync'ed to the server's version. You have the latest installed, you can go to "My Server" and click "Check for updates" to see your version and update your server if needed.
For MeshCentral Mini Router, you need "ClickOnce" support in your browser. Try IE or Edge that have it or you will need an add-in for Chrome and FireFox.
As for your own SSH/RDP client, if they are well know ones, let me know what they are and I will add them to the MeshCentral Router. As longs as I can call them with command line parameters to connect to a local port, that should work.
Let me know if this helps and thanks against for the questions.
Ylian
Thank You so much for your reply ! I'll take a crack at the documentation and send it back for your review.
{
"action": "route",
"localPort": 1234,
"remoteName": "ip-<SSH Server instance IP/hostname>.internal",
"remoteNodeId": "node//<nodeID>",
"remoteTarget": "",
"remotePort": 22,
"username": "admin",
"password": "",
"serverId": "<generated Server ID>",
"serverHttpsHash": "<generated HTTPS hash>>",
"debugLevel": 0,
"serverUrl": "wss://<MeshCentral Instance hostname>:443/meshrelay.ashx"
}
If I try to connect to this route as I would usually use an SSH tunnel (localhost:1234), it doesn't work.
- Agent icon : I do strongly feel any running application, in general, should have an icon in the task bar. For this specific agent, I'd think, even if confirmation is not required, the icon should indicate a connected session when active, and have a quit/stop service option. (I'd be happy to create an application icon for this if none exist yet.
The Mesh agent runs as a service like many other applications/services within a system, i highly doubt you have an icon sitting in the tray many of the other services running in the background doing all kinds of activities. In my usecase, the agent allows for interactive and non-interactive access into systems based on security rights and groups. This is much more preferred then having yet another icon sitting in the tray to confuse end users. You may also want to read up on MeshCommander which is created for the sole purpose of AMT driven system management of devices in organizations. You will notice that AMT does not provide notifications on when it is being accessed. Again, a preferred functionality in large organizations.
Again, the functionality of notifying a user is already present within the Consent features, it works well to notify end users when they are required to accept a session and is well logged within "Events" for auditing purposes.
I agree with penguinthingie in regards to the system tray icon. While it might be a nice option for some I would definitely not want the default behavior for it to utilize a system tray icon. BTW, if it were something that could be made optional it would also be important (at least from perspective) for the icon to be customizable so that we could brand it with our own company logo.