Docker Desktop 4.85.0 on macOS 26.6.2 — Engine becomes unresponsive again after reboot #778

Bug report

Bug description

Docker Desktop on my Apple Silicon Mac becomes unresponsive. The Docker Engine Unix socket exists, but requests to /_ping time out without an HTTP response.

A macOS reboot temporarily restored Engine responsiveness, but the problem returned approximately 100 minutes later.

The Docker VM contains PostgreSQL/PostGIS and Keycloak development resources. I have not established a verified restorable backup of their current data.

Because data preservation is critical, I have not reset Docker Desktop, deleted volumes, pruned resources, reinstalled Docker, or removed Docker.raw.

Steps and observed behavior

1. Docker Desktop became unresponsive.
2. On October 21, a normal docker desktop restart --timeout 45 failed with context deadline exceeded.
3. On October 22, a request to the existing Engine socket timed out.
4. On October 23, I manually rebooted macOS.
5. At 14:02 JST, GET /_ping returned HTTP 200 and OK in 0.263 seconds.
6. At approximately 15:41 JST, the same endpoint stopped responding. A single request timed out after 5.008 seconds with zero response bytes.

Expected behavior: Docker Engine remains responsive after Docker Desktop starts.

Actual behavior: The Engine becomes unresponsive again without an identified cause.

Environment

• Platform: macOS
• macOS: 26.6.2
• Architecture: Apple Silicon arm64
• RAM: 8 GiB
• Docker Desktop: 4.85.0
• Docker Engine: 29.6.2 (last confirmed while responsive)
• Docker context: desktop-linux

Storage observations

• Docker.raw apparent size: approximately 228.27 GiB
• Docker.raw allocated size: approximately 28.68 GiB
• Host free capacity during the latest audit: approximately 18.51 GiB
• Previously observed: 21 images, 14 volumes, 4 stopped containers
• Docker guest filesystem free capacity: unknown

I cannot establish whether storage pressure is the cause.

Diagnostics ID

Not available at this time. Docker Desktop cannot be used reliably, and a diagnostic upload has not been completed.

Request

Could you help identify the cause of this recurring unresponsiveness and recommend safe, non-destructive diagnostic or recovery steps?

Please avoid recommending Factory Reset, volume deletion, disk-image removal, or reinstallation without first addressing preservation of the existing VM data.

Platform

macOS

Version information

macOS: 26.6.2  
Architecture: Apple Silicon (arm64)  
RAM: 8 GiB  
Docker Desktop: 4.85.0  
Docker Engine: 29.6.2 (last confirmed on October 23, 2026, at 14:02 JST)  
Docker Context: desktop-linux

Current Docker Engine status: UNRESPONSIVE

The Docker Desktop backend process is present, but the Engine does not respond through its existing local Unix socket.

On October 23, 2026, at approximately 15:41 JST, a GET /_ping request timed out after 5.008 seconds with zero response bytes.

I cannot run docker version or docker info successfully while the Engine is unresponsive. The Engine version above was confirmed earlier, when the Engine temporarily recovered after a macOS reboot.

Diagnostics ID

No response