pulumi inside docker container with local backend

[llo-man]

nice product!
i wil try it inside docker desktop container.
i have create image based upon "FROM pulumi/pulumi-python:3.256.0-debian-amd64".

the scenario is very simple:

"
export PULUMI_RUNTIME_VIRTUALENV=venv
export PULUMI_RUNTIME_TOOLCHAIN=pip
export PULUMI_CONFIG_PASSPHRASE="xxx"
export PULUMI_DISABLE_AUTOMATIC_PLUGIN_ACQUISITION=true
printenv
mkdir 001-cmd
cd 001-cmd
mkdir state
pulumi login file://./state
pulumi new /pulumi/templates/python --name "001-cmd" --description "test de création basique" -y --force
pulumi up
"
i encounter grpc error.

where am i wrong?

error first lines:

diagnostics:
pulumi:pulumi:Stack (001-cmd-dev):
Traceback (most recent call last):
File "/ciril/tests/001-cmd/venv/lib/python3.13/site-packages/pulumi/runtime/settings.py", line 549, in _load_monitor_feature_support
deployment_info = await asyncio.get_running_loop().run_in_executor(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
...<4 lines>...
)
^
File "/usr/local/lib/python3.13/concurrent/futures/thread.py", line 59, in run
result = self.fn(*self.args, **self.kwargs)
File "/ciril/tests/001-cmd/venv/lib/python3.13/site-packages/pulumi/runtime/_instrumentation.py", line 117, in wrapper
return fn(*args, **kwargs)
File "/ciril/tests/001-cmd/venv/lib/python3.13/site-packages/pulumi/runtime/settings.py", line 552, in
lambda: SETTINGS.monitor.GetDeploymentInfo(empty_pb2.Empty())
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^
File "/ciril/tests/001-cmd/venv/lib/python3.13/site-packages/grpc/_channel.py", line 1168, in call
return _end_unary_response_blocking(state, call, False, None)
File "/ciril/tests/001-cmd/venv/lib/python3.13/site-packages/grpc/_channel.py", line 999, in _end_unary_response_blocking
raise _InactiveRpcError(state) # pytype: disable=not-instantiable
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
grpc._channel._InactiveRpcError: <_InactiveRpcError of RPC that terminated with:
status = StatusCode.UNAVAILABLE
details = "failed to connect to all addresses; last error: UNKNOWN: ipv4:127.0.0.1:36669: Failed to connect to remote host: Connection refused"
debug_error_string = "UNAVAILABLE:failed to connect to all addresses; last error: UNKNOWN: ipv4:127.0.0.1:36669: Failed to connect to remote host: Connection refused"
>

thanks

[tgummerer]

I'm guessing you're running your docker container with "host" networking. This will make it so localhost is the hosts localhost instead of inside the container. However pulumi uses RPCs to communicate with its plugins, so that networking mode won't work with pulumi running inside the docker container. Any other docker networking mode should allow pulumi to work correctly.

[llo-man]

thanks.
i will try without "host".
when i use terraform, i must set it to drive plugin. so i can deploy resource in other docker container. that's why i try this with pulumi

[llo-man]

removing "--net=host" will work perfectly.
thanks
the final goal wiil be a DIY postgresql inside another docker container

really thanks

[llo-man]

hi
this work perfectly. now i use "--add-host=host.docker.internal:host-gateway". i can deploy a container with "pulumi_docker" driver. i also can deploy a kind cluster with "terraform kind provider" after add package.
but i encouter a pb. the kubeconfig file set "127.0.0.1" and a certificates key associated. then with a pulumi kubernetes provider, it failed upon openapi v2: connection refused.
for deeper note, i can access kind cluster outside docker using the kubeconfig file => windows shell kubectl command. where am i wrong?

kind provider generate these token inside kubeconfig file:

"- cluster:
certificate-authority-data: xxxxx
server: https://127.0.0.1:43161"

kubernetes works only with:

"- cluster:
server: https://host.docker.internal:43161
insecure-skip-tls-verify: true"

windows shell kubectl command works with:

"- cluster:
server: https://localhost:43161
insecure-skip-tls-verify: true"