[runtime env] Clarify error message about where to install `smart_ope…

…n` for remote URI (#32110)

At least two users reported encountering

ImportError(
                            "You must `pip install smart_open` and "
                            "`pip install boto3` to fetch URIs in s3 "
                            "bucket. "
and trying to fix it by specifying them in the pip field of runtime_env, which won't work because the runtime_env setup code doesn't run inside the runtime_env. This PR clarifies the error message to say that they must be preinstalled on the cluster, and adds a note to the docs.

doc/source/ray-core/handling-dependencies.rst

@@ -572,7 +572,8 @@ Currently, three types of remote URIs are supported for hosting ``working_dir``
 
  - ``runtime_env = {"working_dir": "gs:https://example_bucket/example_file.zip"}``
 
-
+Note that the ``smart_open``, ``boto3``, and ``google-cloud-storage`` packages are not installed by default, and it is not sufficient to specify them in the ``pip`` section of your ``runtime_env``.
+The relevant packages must already be installed on all nodes of the cluster when Ray starts.
 
 Hosting a Dependency on a Remote Git Provider: Step-by-Step Guide
 -----------------------------------------------------------------

python/ray/_private/runtime_env/packaging.py

@@ -649,6 +649,11 @@ async def download_and_unpack_package(
  elif protocol in Protocol.remote_protocols():
  # Download package from remote URI
  tp = None
+ install_warning = (
+ "Note that these must be preinstalled "
+ "on all nodes in the Ray cluster; it is not "
+ "sufficient to install them in the runtime_env."
+ )
 
  if protocol == Protocol.S3:
  try:
@@ -658,7 +663,7 @@ async def download_and_unpack_package(
  raise ImportError(
  "You must `pip install smart_open` and "
  "`pip install boto3` to fetch URIs in s3 "
- "bucket."
+ "bucket. " + install_warning
  )
  tp = {"client": boto3.client("s3")}
  elif protocol == Protocol.GS:
@@ -670,6 +675,7 @@ async def download_and_unpack_package(
  "You must `pip install smart_open` and "
  "`pip install google-cloud-storage` "
  "to fetch URIs in Google Cloud Storage bucket."
+ + install_warning
  )
  elif protocol == Protocol.FILE:
  pkg_uri = pkg_uri[len("file:https://") :]
@@ -683,7 +689,8 @@ def open_file(uri, mode, *, transport_params=None):
  except ImportError:
  raise ImportError(
  "You must `pip install smart_open` "
- f"to fetch {protocol.value.upper()} URIs."
+ f"to fetch {protocol.value.upper()} URIs. "
+ + install_warning
  )
 
  with open_file(pkg_uri, "rb", transport_params=tp) as package_zip: