Hi @billowy-yak-5387, I ran into this exact same issue on v1.24.0!

The root cause is a healthcheck bug: in

/etc/init.d/docker

, it runs

curl --url http://./_ping

, but newer versions of

curl

reject the

http://.

format with

URL rejected: Bad hostname

. Because the healthcheck fails, Rancher Desktop assumes Docker is dead and drops the Hyper-V socket connection.

Here is the fix to resolve it permanently:

1. Open PowerShell and run this one-liner to create a persistent provisioning script:

powershell

$dir = "$env:LOCALAPPDATA\rancher-desktop\provisioning"

if (!(Test-Path $dir)) { New-Item -ItemType Directory -Path $dir }

[System.IO.File]::WriteAllText("$dir\fix-docker-healthcheck.start", "#!/bin/sh`nsed -i 's|http://./_ping|http://localhost/_ping|' /etc/init.d/docker`n", [System.Text.UTF8Encoding]::new($false))

2. Completely quit Rancher Desktop (right-click the tray icon → Quit).

3. Restart Rancher Desktop.

This changes

http://./_ping

to

<http://localhost/_ping>

, which returns

200 OK

. The Hyper-V socket timeout will completely stop happening!