Graceful shutdown of any systemd distro unmounts shared /mnt/wsl mounts in ALL distros (breaks Docker Desktop WSL integration VM-wide)

[sh41]

Windows Version

Microsoft Windows [Version 10.0.26200.8875]

WSL Version

2.7.10.0

Are you using WSL 1 or WSL 2?

• WSL 2
• WSL 1

Kernel Version

Linux version 6.18.33.2-microsoft-standard-WSL2 (root@f1bbfb02316b) (gcc (GCC) 13.2.0, GNU ld (GNU Binutils) 2.41) #1 SMP PREEMPT_DYNAMIC Thu Jun 18 21:54:43 UTC 2026

Distro Version

Ubuntu 24.04 & 26.04

Other Software

Docker Desktop (Windows): Version 4.82.0 (233772)

Repro Steps

1. Start Docker Desktop with WSL integration enabled for distro A (e.g. Ubuntu-24.04). Verify in A:

$ findmnt /mnt/wsl/docker-desktop/cli-tools
/mnt/wsl/docker-desktop/cli-tools /dev/loop0 iso9660 ro,relatime,...
$ docker --version
Docker version 29.6.1, ...

2. Start any other systemd-enabled distro B, then terminate it:

wsl --terminate B

2. (WSL's idle auto-termination of B would also trigger this. No explicit terminate needed.)
3. In distro A (and in docker-desktop itself): /mnt/wsl/docker-desktop/cli-tools is gone, /dev/loop0 is detached, docker is broken. Docker Desktop's UI still reports everything healthy; its backend log shows nothing.

VM kernel log (dmesg from docker-desktop) at the moment of failure — distro B (ec3ed5cb... = its root ext4) shutting down, immediately followed by the shared mounts dying VM-wide:

[1951.663026] Exception: Operation canceled @p9io.cpp:258 (AcceptAsync)
[1957.137509] systemd-journald[43]: Received SIGTERM from PID 1 (systemd-shutdow).
[1957.337165] EXT4-fs (sdh): unmounting filesystem ec3ed5cb-9466-40be-810a-a5223fd6cd3f.

After this instant: losetup -a inside docker-desktop shows loop0 (docker-wsl-cli.iso) detached; findmnt /mnt/wsl/docker-desktop/cli-tools fails in every distro. The Docker engine itself is untouched (its data disk /dev/sde stays mounted; running containers keep running), only the shared-mount state is destroyed.

Mount propagation state (all distros):

$ grep '/mnt/wsl ' /proc/self/mountinfo
359 367 0:34 / /mnt/wsl rw,relatime shared:1 - tmpfs none rw
$ grep cli-tools /proc/self/mountinfo
978 359 7:0 / /mnt/wsl/docker-desktop/cli-tools ro,relatime shared:568 - iso9660 /dev/loop0 ...

shared:* peer groups spanning all distros are probably why one distro's umount reaches every other one.

When a read of the ISO is in flight during the event, the kernel additionally logs I/O error, dev loop0, sector ... op READ etc. Just noise from the original umount.

Expected Behavior

Terminating/shutting down one distro should not remove shared mounts in other distros.

Actual Behavior

Terminating/shutting down one distro removes shared mounts in other distros.

Diagnostic Logs

Emailed

Workaround

A per-distro systemd unit that flips /mnt/wsl to slave propagation only at shutdown, so the dying distro's unmounts stop propagating outward:

# /etc/systemd/system/wsl-mnt-guard.service
[Unit]
Description=Protect shared /mnt/wsl mounts from shutdown unmount propagation

[Service]
Type=oneshot
RemainAfterExit=yes
ExecStart=/bin/true
ExecStop=/bin/sh -c "mount --make-rslave /mnt/wsl 2>/dev/null || true"

[Install]
WantedBy=multi-user.target

With this enabled in the systemd distros, the repro above no longer breaks anything: cli-tools stays mounted everywhere across distro terminations. Runtime propagation must stay shared. Docker Desktop relies on distro to VM propagation for its bind-mount publishing, which is why the fix happens only in ExecStop.

Possibly linked to:

• WSLInterop binfmt entry disappears when a systemd-enabled WSL distro terminates rancher-sandbox/rancher-desktop#10049
• After Windows reboot, links to docker in /mnt/wsl/docker-desktop/cli-tools disappear docker/for-win#11308
• /mnt/wsl/docker-desktop/cli-tools/ is empty docker/for-win#12578
• WSL2: CLI tools got lost from integration. Cannot restore them. docker/for-win#6441
• Command 'docker' could not be found in this WSL 2 distro error. docker/for-win#13088
• [WSL2][systemd][interop] Unable to Execute Windows Binary when systemd enabled #8843.

[github-actions]

Logs are required for review from WSL team

If this a feature request, please reply with '/feature'. If this is a question, reply with '/question'.
Otherwise, please attach logs by following the instructions below, your issue will not be reviewed unless they are added. These logs will help us understand what is going on in your machine.

How to collect WSL logs

Download and execute collect-wsl-logs.ps1 in an administrative powershell prompt:

Invoke-WebRequest -UseBasicParsing "https://raw.githubusercontent.com/microsoft/WSL/master/diagnostics/collect-wsl-logs.ps1" -OutFile collect-wsl-logs.ps1
Set-ExecutionPolicy Bypass -Scope Process -Force
.\collect-wsl-logs.ps1

The script will output the path of the log file once done.

If this is a networking issue, please use .\collect-wsl-logs.ps1 -LogProfile networking instead, following the instructions in Collect WSL logs for networking issues

Once completed please upload the output files to this GitHub issue.

See Collect WSL logs (recommended method).

If you choose to email these logs instead of attaching them to the bug, please send them to wsl-gh-logs@microsoft.com with the GitHub issue number in the subject, and include a link to your GitHub issue comment in the message body, and reply with '/emailed-logs'.

[sh41]

/emailed-logs

[github-actions]

Diagnostic information

Found '/emailed-logs', adding tag 'emailed-logs'

[jimb-ku]

i use rancher desktop instead of docker and a systemd debian distro for development work. I mount vhdx to /mnt/wsl/data to share files between debian and rancher desktop. in more recent versions of WSL when the debian distro is stopped /mnt/wsl/data gets unmounted. I've found that this behavior starts in WSL 2.6.2 and i still see it in 2.9.4. in WSL 2.6.1 after shutting down debian the files in /mnt/wsl/data remain available.