Hi Databricks Community,

I'm trying to read an index from OpenSearch or write a DataFrame into an OpenSearch index using the native Spark OpenSearch connector:

host = dbutils.secrets.get(scope="opensearch", key="host")
port = dbutils.secrets.get(scope="opensearch", key="port")
index = "index_name"

def save_to_open_search(df):
    df.write.format("org.opensearch.spark.sql") \
        .option("opensearch.nodes", host) \
        .option("opensearch.port", port) \
        .option("opensearch.nodes.resolve.hostname", "false") \
        .option("opensearch.nodes.wan.only", "true") \
        .option("opensearch.net.ssl", "true") \
        .option("opensearch.aws.sigv4.enabled", "true") \
        .option("opensearch.aws.sigv4.region", "us-west-2") \
        .option("opensearch.batch.size.entries", "200") \
        .option("opensearch.mapping.id", "id") \
        .option("opensearch.write.operation", "upsert") \
        .option("opensearch.resource", index) \
        .mode("append") \
        .save()

However, I’m encountering the following error:

Py4JJavaError: An error occurred while calling o456.save.
org.opensearch.hadoop.OpenSearchHadoopIllegalArgumentException: Cannot detect OpenSearch version - 
typically this happens if the network/OpenSearch cluster is not accessible or when targeting a WAN/Cloud instance without the proper setting

Environment Details

• Databricks Workspace: Running on AWS
• Databricks Runtime: 15.4 LTS (Apache Spark 3.5.0, Scala 2.12)
• OpenSearch Spark Connector: org.opensearch.client:opensearch-spark-30_2.12:1.3.0
• Cluster Type: Dedicated

What I’ve Tried

I ruled out network access issues, because I am able to successfully connect to OpenSearch on the same cluster using opensearch-py==2.8.0:

import boto3
from opensearchpy import (
    AWSV4SignerAuth,
    OpenSearch,
    RequestsHttpConnection,
)

def create_os_client(host, port, use_ssl=True, region="us-west-2") -> OpenSearch:
    """
    Function to create an OpenSearch client.
    """
    http_auth = AWSV4SignerAuth(boto3.Session().get_credentials(), region, "es")
    return OpenSearch(
        hosts=[{"host": host, "port": port, "use_ssl": use_ssl}],
        http_auth=http_auth,
        connection_class=RequestsHttpConnection,
        http_compress=True,
    )

os_client = create_os_client(
    host=dbutils.secrets.get(scope="opensearch", key="host"), 
    port=dbutils.secrets.get(scope="opensearch", key="port"),
)

Even though I can connect using opensearch-py, I specifically want to use the native Spark OpenSearch connector (org.opensearch.spark.sql).

Has anyone encountered a similar issue? Any suggestions on how to resolve this?
I’d really appreciate any insights or workarounds!