GV

GeezerVison

The World Wide Web, restored to 800 pixels of glory.

Home|About|Gaming cabinet|Archives|How it works|Starter sites

On this site

  • Sign in
  • ActionsCI/CD and automation
  • CodespacesCloud development environments
  • All Features
  • Documentation
  • Training
  • Certifications
  • Build with Copilot
  • Customize Copilot
  • Integrations
  • Trending repos
  • GitHub Sponsors
  • Developer articles
  • Enterprises

This view

  • Original page
  • New address

GeezerVison is showing github.com — fix: Pre-create S3A event log dir before SparkContext init by abhijeet-dhumal · Pull Request #6317 · feast-dev/feast

fix: Pre-create S3A event log dir before SparkContext init #6317

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ntkathole merged 5 commits into feast-dev:master from abhijeet-dhumal:fix/spark-s3a-event-log-init
Apr 27, 2026

Uh oh!

There was an error while loading. Please reload this page.
Merged

fix: Pre-create S3A event log dir before SparkContext init #6317

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
bc97a4c
fix(spark): Pre-create S3A event log dir before SparkContext init to …
abhijeet-dhumal Apr 22, 2026
c586577
fix(spark): handle bucket-root S3A paths, read credentials from spark…
abhijeet-dhumal Apr 24, 2026
c7b74db
fix(spark): use AWS_ENDPOINT_URL, support path-style addressing, fix …
abhijeet-dhumal Apr 27, 2026
46094c4
fix: allow sample aws secret and fix linting issue
abhijeet-dhumal Apr 27, 2026
70215e2
fix: fix linting issue
abhijeet-dhumal Apr 27, 2026
File filter
Filter by extension
Clear filters
Conversations
Failed to load comments.
Loading

Uh oh!

There was an error while loading. Please reload this page.
Jump to
Jump to file
Failed to load files.
Loading

Uh oh!

There was an error while loading. Please reload this page.
Diff view
Diff view

There are no files selected for viewing

99 changes: 99 additions & 0 deletions 99 sdk/python/feast/infra/compute_engines/spark/utils.py
Show comments
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
import logging
import os
from typing import Dict, Iterable, Literal, Optional

import pandas as pd
Expand All @@ -9,6 +11,102 @@
from feast.infra.common.serde import SerializedArtifacts
from feast.utils import _convert_arrow_to_proto, _run_pyarrow_field_mapping

try:
import boto3
from botocore.client import Config as BotoConfig
except ImportError:
boto3 = None # type: ignore[assignment]
BotoConfig = None # type: ignore[assignment,misc]

logger = logging.getLogger(__name__)


def _ensure_s3a_event_log_dir(spark_config: Dict[str, str]) -> None:
"""Pre-create the S3A event log prefix before SparkContext initialisation.

Spark's EventLogFileWriter.requireLogBaseDirAsDirectory() is called inside
SparkContext.__init__ and crashes if the S3A path doesn't exist yet (S3 has no
real directories, so an empty prefix returns a 404). This function writes a
zero-byte placeholder so the prefix exists before SparkContext is built.

This is only attempted when:
- spark.eventLog.enabled == "true"
- spark.eventLog.dir starts with "s3a://"
Failures are non-fatal: Spark will surface its own error if the dir is still missing.
"""
if spark_config.get("spark.eventLog.enabled", "false").lower() != "true":
return
event_dir = spark_config.get("spark.eventLog.dir", "")
if not event_dir.startswith("s3a://"):
return

path = event_dir[len("s3a://") :]
bucket, _, prefix = path.partition("/")
prefix = prefix.rstrip("/")
prefix = (prefix + "/") if prefix else prefix
placeholder_key = prefix + ".keep"

endpoint = spark_config.get(
"spark.hadoop.fs.s3a.endpoint",
os.environ.get("AWS_ENDPOINT_URL", ""),
)
access_key = spark_config.get(
"spark.hadoop.fs.s3a.access.key",
os.environ.get("AWS_ACCESS_KEY_ID", ""),
)
secret_key = spark_config.get(
"spark.hadoop.fs.s3a.secret.key",
os.environ.get("AWS_SECRET_ACCESS_KEY", ""),
)
session_token = (
spark_config.get(
"spark.hadoop.fs.s3a.session.token",
os.environ.get("AWS_SESSION_TOKEN", ""),
)
or None
)

try:
if boto3 is None:
raise ImportError("boto3 is not installed")

addressing_style = (
"path"
if spark_config.get(
"spark.hadoop.fs.s3a.path.style.access", "false"
).lower()
== "true"
else "auto"
)

s3 = boto3.client(
"s3",
endpoint_url=endpoint if endpoint else None,
aws_access_key_id=access_key or None,
aws_secret_access_key=secret_key or None,
aws_session_token=session_token,
config=BotoConfig(
signature_version="s3v4",
s3={"addressing_style": addressing_style},
),
)
resp = s3.list_objects_v2(Bucket=bucket, Prefix=prefix, MaxKeys=1)
if resp.get("KeyCount", 0) == 0:
s3.put_object(Bucket=bucket, Key=placeholder_key, Body=b"")
logger.debug(
"Created S3A event log dir placeholder: s3a://%s/%s",
bucket,
placeholder_key,
)
except Exception as exc:
logger.warning(
"Could not pre-create S3A event log dir s3a://%s/%s — "
"SparkContext may fail if the path still doesn't exist: %s",
bucket,
prefix,
exc,
)


def get_or_create_new_spark_session(
spark_config: Optional[Dict[str, str]] = None,
Expand All @@ -17,6 +115,7 @@ def get_or_create_new_spark_session(
if not spark_session:
spark_builder = SparkSession.builder
if spark_config:
_ensure_s3a_event_log_dir(spark_config)
spark_builder = spark_builder.config(
conf=SparkConf().setAll([(k, v) for k, v in spark_config.items()])
)
Expand Down
274 changes: 274 additions & 0 deletions 274 sdk/python/tests/component/spark/test_spark_utils.py
Show comments
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,274 @@
from unittest.mock import MagicMock, patch

from feast.infra.compute_engines.spark.utils import _ensure_s3a_event_log_dir

BOTO3_PATH = "feast.infra.compute_engines.spark.utils.boto3"
BOTOCONFIG_PATH = "feast.infra.compute_engines.spark.utils.BotoConfig"


def _base_conf(event_log_dir: str) -> dict:
return {
"spark.eventLog.enabled": "true",
"spark.eventLog.dir": event_log_dir,
"spark.hadoop.fs.s3a.endpoint": "http://minio:9000",
}


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_creates_placeholder_when_empty(mock_boto3):
"""S3A prefix doesn't exist -> placeholder object is written."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 0}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/spark-events/"))

s3.list_objects_v2.assert_called_once_with(
Bucket="my-bucket", Prefix="spark-events/", MaxKeys=1
)
s3.put_object.assert_called_once_with(
Bucket="my-bucket", Key="spark-events/.keep", Body=b""
)


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_skips_when_prefix_exists(mock_boto3):
"""S3A prefix already has objects -> no placeholder written."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 3}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/spark-events/"))

s3.put_object.assert_not_called()


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_noop_when_event_log_disabled(mock_boto3):
"""spark.eventLog.enabled != true -> boto3 never called."""
_ensure_s3a_event_log_dir(
{"spark.eventLog.enabled": "false", "spark.eventLog.dir": "s3a://b/p/"}
)
mock_boto3.client.assert_not_called()


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_noop_for_non_s3a_path(mock_boto3):
"""Non-S3A paths (hdfs://, file://, etc.) are left untouched."""
_ensure_s3a_event_log_dir(
{"spark.eventLog.enabled": "true", "spark.eventLog.dir": "hdfs:///spark-logs"}
)
mock_boto3.client.assert_not_called()


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_non_fatal_on_s3_error(mock_boto3):
"""boto3 errors are swallowed -> SparkContext will surface its own error."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.side_effect = Exception("connection refused")

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/spark-events/"))


# ---------------------------------------------------------------------------
# Bucket-root edge cases (s3a://bucket, s3a://bucket/)
# ---------------------------------------------------------------------------


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_bucket_root_no_trailing_slash(mock_boto3):
"""s3a://bucket (no path) -> .keep at bucket root, not /.keep."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 0}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket"))

s3.list_objects_v2.assert_called_once_with(Bucket="my-bucket", Prefix="", MaxKeys=1)
s3.put_object.assert_called_once_with(Bucket="my-bucket", Key=".keep", Body=b"")


@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_bucket_root_trailing_slash(mock_boto3):
"""s3a://bucket/ (trailing slash, empty prefix) -> .keep at bucket root."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 0}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/"))

s3.list_objects_v2.assert_called_once_with(Bucket="my-bucket", Prefix="", MaxKeys=1)
s3.put_object.assert_called_once_with(Bucket="my-bucket", Key=".keep", Body=b"")


# ---------------------------------------------------------------------------
# Credentials from spark config / env var fallback
# ---------------------------------------------------------------------------


@patch.dict(
"os.environ",
{
"AWS_ACCESS_KEY_ID": "env-ak",
"AWS_SECRET_ACCESS_KEY": "env-sk", # pragma: allowlist secret
"AWS_SESSION_TOKEN": "env-st",
},
)
@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_uses_spark_config_credentials(mock_boto3):
"""Credentials in spark config take precedence over env vars."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

conf = {
**_base_conf("s3a://my-bucket/logs/"),
"spark.hadoop.fs.s3a.access.key": "spark-ak",
"spark.hadoop.fs.s3a.secret.key": "spark-sk", # pragma: allowlist secret
"spark.hadoop.fs.s3a.session.token": "spark-st",
}
_ensure_s3a_event_log_dir(conf)

mock_boto3.client.assert_called_once()
kw = mock_boto3.client.call_args.kwargs
assert kw["aws_access_key_id"] == "spark-ak"
assert kw["aws_secret_access_key"] == "spark-sk" # pragma: allowlist secret
assert kw["aws_session_token"] == "spark-st"


@patch.dict(
"os.environ",
{
"AWS_ACCESS_KEY_ID": "env-ak",
"AWS_SECRET_ACCESS_KEY": "env-sk", # pragma: allowlist secret
"AWS_SESSION_TOKEN": "env-st",
},
)
@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_falls_back_to_env_credentials(mock_boto3):
"""Without spark config keys, env vars are used."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/logs/"))

mock_boto3.client.assert_called_once()
kw = mock_boto3.client.call_args.kwargs
assert kw["aws_access_key_id"] == "env-ak"
assert kw["aws_secret_access_key"] == "env-sk" # pragma: allowlist secret
assert kw["aws_session_token"] == "env-st"


@patch.dict("os.environ", {}, clear=True)
@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_no_credentials_passes_none(mock_boto3):
"""No credentials anywhere -> None passed to boto3 (anonymous / instance role)."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

conf = {
"spark.eventLog.enabled": "true",
"spark.eventLog.dir": "s3a://my-bucket/logs/",
}
_ensure_s3a_event_log_dir(conf)

mock_boto3.client.assert_called_once()
kw = mock_boto3.client.call_args.kwargs
assert kw["aws_access_key_id"] is None
assert kw["aws_secret_access_key"] is None
assert kw["aws_session_token"] is None


# ---------------------------------------------------------------------------
# Path-style addressing (MinIO / S3-compatible)
# ---------------------------------------------------------------------------


@patch(BOTOCONFIG_PATH)
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_path_style_when_enabled(mock_boto3, mock_config_cls):
"""spark.hadoop.fs.s3a.path.style.access=true -> addressing_style='path'."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

conf = {
**_base_conf("s3a://my-bucket/logs/"),
"spark.hadoop.fs.s3a.path.style.access": "true",
}
_ensure_s3a_event_log_dir(conf)

mock_config_cls.assert_called_once()
config_kwargs = mock_config_cls.call_args
assert config_kwargs.kwargs["s3"] == {"addressing_style": "path"}


@patch(BOTOCONFIG_PATH)
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_virtual_hosted_style_by_default(
mock_boto3, mock_config_cls
):
"""No path.style.access config -> addressing_style='auto'."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/logs/"))

mock_config_cls.assert_called_once()
config_kwargs = mock_config_cls.call_args
assert config_kwargs.kwargs["s3"] == {"addressing_style": "auto"}


# ---------------------------------------------------------------------------
# Endpoint env var fallback (AWS_ENDPOINT_URL)
# ---------------------------------------------------------------------------


@patch.dict("os.environ", {"AWS_ENDPOINT_URL": "http://localhost:9000"}, clear=True)
@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_endpoint_from_env(mock_boto3):
"""AWS_ENDPOINT_URL env var is used when spark config has no endpoint."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

conf = {
"spark.eventLog.enabled": "true",
"spark.eventLog.dir": "s3a://my-bucket/logs/",
}
_ensure_s3a_event_log_dir(conf)

mock_boto3.client.assert_called_once()
kw = mock_boto3.client.call_args.kwargs
assert kw["endpoint_url"] == "http://localhost:9000"


@patch.dict("os.environ", {"AWS_ENDPOINT_URL": "http://env-endpoint:9000"}, clear=True)
@patch(BOTOCONFIG_PATH, MagicMock())
@patch(BOTO3_PATH)
def test_ensure_s3a_event_log_dir_spark_endpoint_over_env(mock_boto3):
"""spark.hadoop.fs.s3a.endpoint takes precedence over AWS_ENDPOINT_URL."""
s3 = MagicMock()
mock_boto3.client.return_value = s3
s3.list_objects_v2.return_value = {"KeyCount": 1}

_ensure_s3a_event_log_dir(_base_conf("s3a://my-bucket/logs/"))

mock_boto3.client.assert_called_once()
kw = mock_boto3.client.call_args.kwargs
assert kw["endpoint_url"] == "http://minio:9000"
Loading
Oops, something went wrong.

Uh oh!

There was an error while loading. Please reload this page.
Add this suggestion to a batch that can be applied as a single commit. This suggestion is invalid because no changes were made to the code. Suggestions cannot be applied while the pull request is closed. Suggestions cannot be applied while viewing a subset of changes. Only one suggestion per line can be applied in a batch. Add this suggestion to a batch that can be applied as a single commit. Applying suggestions on deleted lines is not supported. You must change the existing code in this line in order to create a valid suggestion. Outdated suggestions cannot be applied. This suggestion has been applied or marked resolved. Suggestions cannot be applied from pending reviews. Suggestions cannot be applied on multi-line comments. Suggestions cannot be applied while the pull request is queued to merge. Suggestion cannot be applied right now. Please check back later.
You are visitor 01402503 — GeezerVison does not store the pages you look up.
Best viewed with Netscape Navigator 4.7 or Internet Explorer 6 at 800×600. No cookies required.