comments

cloudinit/cmd/main.py

@@ -1151,6 +1151,11 @@ def main(sysv_args=None):
     setattr(args, "skip_log_setup", False)
     if not args.all_stages:
         return sub_main(args)
+    return all_stages(parser)
+
+
+def all_stages(parser):
+    """Run all stages in a single process using an ordering protocol."""
     LOG.info("Running cloud-init in single process mode.")
 
     # this _must_ be called before sd_notify is called otherwise netcat may
@@ -1167,7 +1172,7 @@ def main(sysv_args=None):
         # run local stage
         sub_main(args)
 
-    # wait for cloud-init.service to start
+    # wait for cloud-init-network.service to start
     with sync("network"):
         # skip re-setting up logger
         args = parser.parse_args(args=["init"])
@@ -1191,7 +1196,7 @@ def main(sysv_args=None):
         # run final stage
         sub_main(args)
 
-    # signal completion to cloud-init.service
+    # signal completion to cloud-init-main.service
     if sync.first_exception:
         socket.sd_notify(
             "STATUS=Completed with failure, first exception received: "

cloudinit/cmd/status.py

@@ -318,8 +318,9 @@ def systemd_failed(wait: bool) -> bool:
     for service in [
         "cloud-final.service",
         "cloud-config.service",
-        "cloud-init.service",
+        "cloud-init-network.service",
         "cloud-init-local.service",
+        "cloud-init-main.service",
     ]:
         try:
             stdout = query_systemctl(

cloudinit/config/cc_mounts.py

@@ -340,7 +340,8 @@ def handle(name: str, cfg: Config, cloud: Cloud, args: list) -> None:
     uses_systemd = cloud.distro.uses_systemd()
     if uses_systemd:
         def_mnt_opts = (
-            "defaults,nofail,x-systemd.after=cloud-init.service,_netdev"
+            "defaults,nofail,x-systemd.after=cloud-"
+            "init-network.service,_netdev"
         )
 
     defvals = [None, None, "auto", def_mnt_opts, "0", "2"]

cloudinit/config/schemas/schema-cloud-config-v1.json

@@ -2017,12 +2017,12 @@
         },
         "mount_default_fields": {
           "type": "array",
-          "description": "Default mount configuration for any mount entry with less than 6 options provided. When specified, 6 items are required and represent ``/etc/fstab`` entries. Default: ``defaults,nofail,x-systemd.after=cloud-init.service,_netdev``",
+          "description": "Default mount configuration for any mount entry with less than 6 options provided. When specified, 6 items are required and represent ``/etc/fstab`` entries. Default: ``defaults,nofail,x-systemd.after=cloud-init-network.service,_netdev``",
           "default": [
             null,
             null,
             "auto",
-            "defaults,nofail,x-systemd.after=cloud-init.service",
+            "defaults,nofail,x-systemd.after=cloud-init-network.service",
             "0",
             "2"
           ],

cloudinit/socket.py

@@ -98,11 +98,13 @@ def __enter__(self):
         Once the message has been received, enter the context.
         """
         if os.isatty(sys.stdin.fileno()):
-            LOG.info("Not syncing, stdin is a tty.")
+            LOG.info(
+                "Stdin is a tty, so skipping stage synchronization protocol"
+            )
             return
         sd_notify(
             "STATUS=Waiting on external services to "
-            f"complete ({self.stage} stage)"
+            f"complete before starting the {self.stage} stage."
         )
         start_time = time.monotonic()
         # block until init system sends us data
@@ -126,7 +128,7 @@ def __enter__(self):
             raise ValueError(f"Unexpected path to unix socket: {self.remote}")
 
         total = time.monotonic() - start_time
-        time_msg = f"took {total: .3f}s " if total > 0.1 else ""
+        time_msg = f"took {total: .3f}s to " if total > 0.01 else ""
         sd_notify(f"STATUS=Running ({self.stage} stage)")
         LOG.debug("sync(%s): synchronization %scomplete", self.stage, time_msg)
         return self
@@ -140,8 +142,8 @@ def __exit__(self, exc_type, exc_val, exc_tb):
             systemd_exit_code = "1"
             status = f"{repr(exc_val)} in {exc_tb.tb_frame}"
             message = (
-                'fatal error, run "systemctl cloud-init.service" for more '
-                "details"
+                'fatal error, run "systemctl status cloud-init-main.service" '
+                'and "cloud-init status --long" for more details'
             )
             if not self.first_exception:
                 self.first_exception = message

systemd/cloud-config.service.tmpl

@@ -10,6 +10,13 @@ ConditionEnvironment=!KERNEL_CMDLINE=cloud-init=disabled
 
 [Service]
 Type=oneshot
+# This service is a shim which preserves systemd ordering while allowing a
+# single Python process to run cloud-init's logic. This works by communicating
+# with the cloud-init process over a unix socket to tell the process that this
+# stage can start, and then wait on a return socket until the cloud-init
+# process has completed this stage. The output from the return socket is piped
+# into a shell so that the process can send a completion message (defaults to
+# "done", otherwise includes an error message) and an exit code to systemd.
 ExecStart=sh -c 'echo "start" | nc.openbsd -Uu -W1 /run/cloud-init/share/config.sock -s /run/cloud-init/share/config-return.sock | sh'
 RemainAfterExit=yes
 TimeoutSec=0

systemd/cloud-final.service.tmpl

@@ -15,6 +15,13 @@ ConditionEnvironment=!KERNEL_CMDLINE=cloud-init=disabled
 
 [Service]
 Type=oneshot
+# This service is a shim which preserves systemd ordering while allowing a
+# single Python process to run cloud-init's logic. This works by communicating
+# with the cloud-init process over a unix socket to tell the process that this
+# stage can start, and then wait on a return socket until the cloud-init
+# process has completed this stage. The output from the return socket is piped
+# into a shell so that the process can send a completion message (defaults to
+# "done", otherwise includes an error message) and an exit code to systemd.
 ExecStart=sh -c 'echo "start" | nc.openbsd -Uu -W1 /run/cloud-init/share/final.sock -s /run/cloud-init/share/final-return.sock | sh'
 RemainAfterExit=yes
 TimeoutSec=0

systemd/cloud-init-local.service.tmpl

@@ -37,6 +37,13 @@ ExecStartPre=/bin/mkdir -p /run/cloud-init
 ExecStartPre=/sbin/restorecon /run/cloud-init
 ExecStartPre=/usr/bin/touch /run/cloud-init/enabled
 {% endif %}
+# This service is a shim which preserves systemd ordering while allowing a
+# single Python process to run cloud-init's logic. This works by communicating
+# with the cloud-init process over a unix socket to tell the process that this
+# stage can start, and then wait on a return socket until the cloud-init
+# process has completed this stage. The output from the return socket is piped
+# into a shell so that the process can send a completion message (defaults to
+# "done", otherwise includes an error message) and an exit code to systemd.
 ExecStart=sh -c 'echo "start" | nc.openbsd -Uu -W1 /run/cloud-init/share/local.sock -s /run/cloud-init/share/local-return.sock | sh'
 RemainAfterExit=yes
 TimeoutSec=0

systemd/cloud-network.service.tmpl → systemd/cloud-init-network.service.tmpl

@@ -46,6 +46,13 @@ ConditionEnvironment=!KERNEL_CMDLINE=cloud-init=disabled
 
 [Service]
 Type=oneshot
+# This service is a shim which preserves systemd ordering while allowing a
+# single Python process to run cloud-init's logic. This works by communicating
+# with the cloud-init process over a unix socket to tell the process that this
+# stage can start, and then wait on a return socket until the cloud-init
+# process has completed this stage. The output from the return socket is piped
+# into a shell so that the process can send a completion message (defaults to
+# "done", otherwise includes an error message) and an exit code to systemd.
 ExecStart=sh -c 'echo "start" | nc.openbsd -Uu -W1 /run/cloud-init/share/network.sock -s /run/cloud-init/share/network-return.sock | sh'
 RemainAfterExit=yes
 TimeoutSec=0

tests/integration_tests/assets/enable_coverage.py

@@ -2,7 +2,7 @@
 
 services = [
     "cloud-init-local.service",
-    "cloud-init.service",
+    "cloud-init-network.service",
     "cloud-config.service",
     "cloud-final.service",
 ]

tests/integration_tests/assets/enable_profile.py

@@ -2,7 +2,7 @@
 
 services = [
     "cloud-init-local.service",
-    "cloud-init.service",
+    "cloud-init-network.service",
     "cloud-config.service",
     "cloud-final.service",
 ]

tests/integration_tests/conftest.py

@@ -191,7 +191,7 @@ def _collect_profile(instance: IntegrationInstance, log_dir: Path):
             log_dir / "profile" / "local.stats",
         )
         instance.pull_file(
-            "/var/log/cloud-init.service.stats",
+            "/var/log/cloud-init-network.service.stats",
             log_dir / "profile" / "network.stats",
         )
         instance.pull_file(

tests/integration_tests/datasources/test_nocloud.py

@@ -162,7 +162,7 @@ def test_smbios_seed_network(self, client: IntegrationInstance):
                 """\
                 [Unit]
                 Description=Serve a local webserver
-                Before=cloud-init.service
+                Before=cloud-init-network.service
                 Wants=cloud-init-local.service
                 DefaultDependencies=no
                 After=systemd-networkd-wait-online.service
@@ -353,7 +353,7 @@ def _boot_with_cmdline(
                 # and NoCloud operates in network timeframe
                 After=systemd-networkd-wait-online.service
                 After=networking.service
-                Before=cloud-init.service
+                Before=cloud-init-network.service
 
                 [Service]
                 Type=exec

tests/unittests/test_all_stages.py

@@ -202,6 +202,7 @@ def test_all_stages_exception(tmp_path):
             1 / 0  # pylint: disable=W0104
 
         assert (
-            b'echo \'fatal error, run "systemctl cloud-init.service" for '
+            b"echo 'fatal error, run \"systemctl status cloud-init-main."
+            b'service" and "cloud-init status --long" for '
             b"more details'; exit 1;" == first.receive()
         )