What you quoted sounds like a classic “Windows update + WSL + VPN” networking regression, just with a new twist (mirrored mode + ARP on the VPN virtual NIC). I can’t verify those specific KBs or dates because my knowledge stops in late 2024, but the behavior and explanation Microsoft gives are technically plausible.
Based on what you posted, here’s what it
likely boils down to and what you can realistically do right now.
What’s actually happening (in plain terms)
- WSL in mirrored mode networking gives the Linux side an IP on the same network as Windows, to improve VPN and LAN integration (IPv6, multicast, etc.).
- Your Windows host can still reach corporate resources over VPN.
- Inside WSL, connections to the same targets fail with "No route to host".
- Microsoft’s explanation: the VPN’s virtual adapter isn’t responding properly to ARP requests coming from the WSL side in mirrored mode, so WSL can’t resolve MAC addresses for those remote endpoints. No ARP = no layer‑3 connectivity = “No route to host.”
- Reportedly affects OpenVPN and enterprise VPNs like Cisco Secure Client; home users are mostly unaffected because they rarely use this enterprise-style setup.
If this issue is real/ongoing, what can you do?
Microsoft (per your quote) hasn’t published an official workaround yet, so anything below is general guidance based on how WSL/VPN breakage is usually handled. Treat these as
options to test, not guaranteed fixes.
1. Temporarily avoid or roll back the problematic updates (enterprise best practice)
If you manage devices in an org:
- Check the official Windows release health dashboard for your Windows 11 version to confirm the issue and track any official mitigations.
- On affected test machines, you can:
- Uninstall the specific October 2025 preview update (KB5067036) and/or the cumulative update (KB5072033) that introduced the issue, if that’s allowed by your policy.
- Delay or block those updates via WSUS/Intune/Group Policy on machines that require WSL + VPN for daily work, until Microsoft lists a fix.
- Always test rollback on a small pilot group before doing it org‑wide.
2. Switch WSL away from mirrored mode (back to NAT)
Since the bug is specifically tied to
mirrored mode networking, one practical workaround is to make WSL use classic NAT networking again so that ARP behavior on the VPN virtual NIC is bypassed from the Linux side.
Conceptually, this means:
- WSL gets a private IP on a virtual NAT network.
- Windows acts as a router/NAT for WSL traffic.
- The VPN only needs to work correctly on the Windows side; WSL traffic is just “normal” outbound traffic from the host, which typically survives these ARP quirks.
How to do this will depend on how WSL networking mode is configured on your systems (for example via `.wslconfig` in your user profile or machine‑wide settings). If you’ve explicitly enabled mirrored mode, you can:
- Remove or comment out the mirrored mode setting.
- Or explicitly set networking back to NAT mode (if that option exists on the WSL build you’re running).
Because I don’t have access to your exact WSL version or Microsoft’s 2025 docs, I’d recommend:
- Checking the official WSL documentation and the WSL GitHub repo for:
- Networking mode settings (mirrored vs NAT)
- Any open issues specifically mentioning these KBs + "mirrored mode" + "No route to host"
3. Use Windows as a “jump host” instead of connecting directly from WSL
If you
must keep mirrored mode (or can’t change system settings), an ugly but often workable workaround is:
- Connect VPN on Windows as usual.
- From WSL, don’t try to hit corporate resources directly.
- Instead, SSH or RDP into a host that is reachable from Windows over the VPN (e.g. a dev/terminal server) and do your Linux work there, or:
- Use port forwarding from Windows into WSL (for development servers) and access them via browser/clients on Windows rather than from inside WSL.
This doesn’t “fix” the bug, but it can keep people productive until Microsoft ships a patch.
4. Test alternative VPN clients/profiles (if your org allows it)
Since the core issue is how the VPN’s
virtual interface handles ARP in this mirrored scenario:
- If your environment supports multiple VPN solutions (e.g. OpenVPN vs WireGuard vs Cisco), test whether all of them are affected.
- Sometimes a different driver mode or adapter type (TAP vs TUN, different version of the client) changes the ARP behavior enough to avoid the bug.
- Only do this if it’s allowed by your corporate policy/security team.
5. For home users
As the Microsoft quote suggests:
- If you’re on Windows Home/Pro and only connect to consumer VPNs (privacy VPNs, streaming, etc.), you’re unlikely to hit this.
- If you do see “No route to host” inside WSL after these updates:
- Try switching WSL networking back to NAT as above.
- If that doesn’t help, consider uninstalling the specific update as a test and see if WSL connectivity returns.
How to track this going forward
Since everything in your quote is for 2025 (beyond my knowledge window), the best sources of
reliable info will be:
- The official Windows release health page for your exact Windows 11 build.
- The WSL GitHub repository (Issues section) – search for:
- The KB numbers (KB5067036, KB5072033)
- "mirrored mode", "VPN", "ARP", "No route to host"
- Your organization’s internal IT / networking team – they may already have a documented mitigation (GPO to block a certain update, standard .wslconfig, etc.).
If you share how you’re currently using WSL (version, networking mode, VPN type/client), I can suggest a more concrete “try this first” sequence tailored to that setup.