From 5e1337e52440c2be059e37c57e62711900b48523 Mon Sep 17 00:00:00 2001
From: emil <emil@oblachno.fyi>
Date: Fri, 14 Aug 2026 20:30:58 +0200
Subject: [PATCH] DEVX-155: fix: use /dev/shm/docker.sock socket for local
 dockerd

The existing dockerd holds /var/run/docker.sock and can't be killed
from inside the container (different PID namespace). Use a new socket
path /dev/shm/docker.sock and data root /dev/shm/docker (16G tmpfs).

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
---
 src/devx/molecule/start_docker.py | 26 +++++++++++++-------------
 tests/unit/test_start_docker.py   |  4 ++--
 2 files changed, 15 insertions(+), 15 deletions(-)

diff --git a/src/devx/molecule/start_docker.py b/src/devx/molecule/start_docker.py
index 9913177..dcba628 100644
--- a/src/devx/molecule/start_docker.py
+++ b/src/devx/molecule/start_docker.py
@@ -23,6 +23,7 @@ Usage::
 
 from __future__ import annotations
 
+import contextlib
 import glob
 import os
 import shutil
@@ -223,24 +224,23 @@ def start_docker_daemon(timeout: int = DEFAULT_TIMEOUT) -> bool:
 
     click.echo(_("Host Docker not available, starting local dockerd..."))
 
-    # Kill any existing dockerd that's using the default data root on the
-    # CI container's overlay (38G, often 100% full). The new dockerd will
-    # use /dev/shm/docker (16G tmpfs) which has plenty of space.
-    subprocess.run(  # nosec B603 B607
-        ["pkill", "-f", "dockerd"],
-        capture_output=True,
-        check=False,
-    )
-    time.sleep(2)
+    # The existing dockerd (started by the CI image) uses the default data
+    # root on the CI container's overlay (38G, often 100% full). We can't
+    # kill it (different PID namespace) or bind to the same socket. Use a
+    # new socket path and data root on /dev/shm (16G tmpfs).
+    local_sock = "/dev/shm/docker.sock"  # nosec B108
+    docker_data_root = "/dev/shm/docker"  # nosec B108
 
-    # Reset DOCKER_HOST to host socket for local dockerd
-    os.environ["DOCKER_HOST"] = f"unix://{DOCKER_SOCK}"
+    # Remove stale socket if present
+    with contextlib.suppress(OSError):
+        os.unlink(local_sock)
+
+    os.environ["DOCKER_HOST"] = f"unix://{local_sock}"
 
     # Start local dockerd (requires privileged container)
     # Use /dev/shm/docker as data root — the CI container's overlay (38G)
     # is often 100% full, causing image pull failures. /dev/shm is a 16G
     # tmpfs with plenty of space for molecule test images.
-    docker_data_root = "/dev/shm/docker"  # nosec B108
     log_file = tempfile.NamedTemporaryFile(  # noqa: SIM115
         mode="w", suffix="dockerd.log", delete=False
     )
@@ -253,7 +253,7 @@ def start_docker_daemon(timeout: int = DEFAULT_TIMEOUT) -> bool:
             "--data-root",
             docker_data_root,
             "-H",
-            f"unix://{DOCKER_SOCK}",
+            f"unix://{local_sock}",
         ],
         stdout=log_file,
         stderr=subprocess.STDOUT,
diff --git a/tests/unit/test_start_docker.py b/tests/unit/test_start_docker.py
index 72be872..e2f4a18 100644
--- a/tests/unit/test_start_docker.py
+++ b/tests/unit/test_start_docker.py
@@ -257,7 +257,7 @@ class TestStartDockerDaemon:
         with patch("builtins.open", mock_open(read_data="dockerd error log")):
             assert start_docker_daemon(timeout=3) is False
         mock_popen.assert_called_once()
-        assert mock_sleep.call_count == 4  # 1 pkill + 3 timeout
+        assert mock_sleep.call_count == 3
 
     @patch("devx.molecule.start_docker.glob.glob", return_value=[])
     @patch("devx.molecule.start_docker._diagnose_socket")
@@ -311,7 +311,7 @@ class TestStartDockerDaemon:
         mock_ready.side_effect = [False, False, True]
         assert start_docker_daemon(timeout=5) is True
         assert mock_popen.call_count == 1
-        assert mock_sleep.call_count == 2  # 1 pkill + 1 timeout
+        assert mock_sleep.call_count == 1
 
     @patch("devx.molecule.start_docker._diagnose_socket")
     @patch("devx.molecule.start_docker._get_docker_free_bytes", return_value=100 * 1024**3)