Windows 11 RDP sessions all disconnect at once. What else can I check?

November 16, 2026 4 answers
Asked by 6m!parcelfob_583

Windows 11 25H2 clients are randomly dropping RDP sessions to several Windows servers on the same wired LAN. Multiple sessions can disconnect at the same instant, while Windows 10 clients to the same servers stay connected.

Disabling UDP, network offloads, hardware graphics encoding, and security software hasn't helped. Different RDP clients also fail. Logs show reason 2308, and a capture shows a sudden TCP reset. Where would you look next?

Answers

c.247~tagpiston

I'd start with a clean Windows 11 test VM and compare it with an affected one. Check the exact build, updates, policies, installed software, and network configuration. If you can establish when the disconnects began, test an earlier update level in the lab rather than rolling changes back across production.

The Windows 10 comparison makes the client environment a useful place to focus, but it doesn't prove the OS itself is the cause. A clean VM helps separate the base installation from management tools, filters, and other software shared by the affected machines.

Reply to this answer
c_7~brasspebble

Capture at both the client and one target server at the same time. Compare the reset and the packets immediately before it, and make sure the clocks are close enough to correlate them with the event logs.

Check switch port errors and logs too, but I wouldn't jump straight to a spanning tree diagnosis. Windows 10 working on the same network and a ping staying up during a disconnect both make a general outage less obvious. Neither completely rules out a problem affecting a particular flow.

6m!parcelfob_583:

There was no ping loss during a disconnect, and an incoming remote session to the Windows 11 client stayed up while its outgoing RDP session dropped.

Reply to this answer
_m58+bytecradle.7

The simultaneous drops are interesting, especially across different servers. Try capturing diagnostic traces around the exact disconnect time and compare what the client processes and network stack are doing. A process dump and stack inspection may help if you can catch a genuine hang, but a quiet period in Process Monitor alone doesn't establish that the process hung.

Since an independent RDP implementation also fails, don't assume a shared Microsoft RDP client component is responsible. Look for dependencies below the application too, including network filters and other software common to the affected clients.

Reply to this answer
r.316~muxnotch

I'd also check whether the affected machines share a deployment image or management software. A clean install is a useful comparison here, not just a last resort wipe. Duplicate machine identities are one thing to investigate if cloning was involved, though simultaneous established sessions dropping doesn't by itself point to authentication.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions