Using the Python SDK (Boto3)

Boto3 is the AWS software development kit (SDK) for Python. It enables you to create, configure, and manage S3-compatible services, such as Cloud Temple object storage, directly from your Python applications.

Installation

The library installation is straightforward using pip, the Python package manager:

pip install boto3

Boto3 Client Configuration

To interact with the service, you must initialize a Boto3 client by providing your credentials and the specific endpoint for Cloud Temple.

import boto3
from botocore.config import Config

# Your credentials
access_key = 'YOUR_ACCESS_KEY'
secret_key = 'YOUR_SECRET_KEY'
endpoint_url = 'https://YOUR_NAMESPACE.s3.fr1.cloud-temple.com'
region = 'fr1'

# Bucket name for examples
bucket_name = 'mon-bucket-test'

Region

The region to use for Cloud Temple object storage is fr1. Make sure to properly set this value in the region_name parameter.

Signature specifics (SigV2 / SigV4)

Cloud Temple Object Storage, based on Dell ECS, supports two versions of AWS signing:

SigV4 (signature_version='s3v4'): recommended for metadata operations (LIST, HEAD)
SigV2 (signature_version='s3'): recommended for data operations (PUT, GET, DELETE)

Recommended hybrid approach

For optimal compatibility, it is recommended to use two Boto3 clients with different signing configurations depending on the operation type. This avoids XAmzContentSHA256Mismatch errors that can occur with SigV4 during write operations.

import boto3
from botocore.config import Config

# Client SigV2 for Data Operations (PUT/GET/DELETE)
config_v2 = Config(
    region_name='fr1',
    signature_version='s3',       # SigV2 (legacy)
    s3={'addressing_style': 'path'},
    retries={'max_attempts': 3, 'mode': 'adaptive'}
)

client_data = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=config_v2
)

# Client SigV4 for Metadata Operations (HEAD/LIST)
config_v4 = Config(
    region_name='fr1',
    signature_version='s3v4',
    s3={'addressing_style': 'path', 'payload_signing_enabled': False},
    retries={'max_attempts': 3, 'mode': 'adaptive'}
)

client_metadata = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=config_v4
)

Simplified Configuration

If you want to use a single client, prefer SigV2 (signature_version='s3') with path-style addressing, which works for most operations:

config = Config(
    region_name='fr1',
    signature_version='s3',
    s3={'addressing_style': 'path'}
)

s3 = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=config
)

1. List files in an S3 bucket

To list all objects in a bucket, you can use a paginator. This is the most robust method for traversing a large number of objects without risking API response limits.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import boto3
from botocore.config import Config

# Initialize S3 client (SigV4 recommended for LIST)
s3_client = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=Config(
        region_name='fr1',
        signature_version='s3v4',
        s3={'addressing_style': 'path'}
    )
)

paginator = s3_client.get_paginator('list_objects_v2')
operation_parameters = {'Bucket': bucket_name}

for page in paginator.paginate(**operation_parameters):
    for obj in page.get('Contents', []):
        print(f"{obj['Key']} - {obj['Size']} bytes")

2. Upload an object (Upload)

The put_object method allows you to upload data to a bucket. It is particularly useful when the data is generated or manipulated in memory.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import boto3
import json
from botocore.config import Config

# S3 Configuration (SigV2 recommended for PUT)
s3 = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=Config(
        region_name='fr1',
        signature_version='s3',
        s3={'addressing_style': 'path'}
    )
)

json_file = {
    "id": 12345,
    "nom": "Dupont",
    "prenom": "Claire",
    "email": "claire.dupont@example.com",
    "actif": True,
    "date_creation": "2024-10-15T08:30:00Z"
}

key = "users/dupont_claire.json"

# Convert to JSON string then to bytes
json_str = json.dumps(json_file, indent=2)
body_bytes = json_str.encode("utf-8")

result = s3.put_object(
    Bucket=bucket_name,
    Key=key,
    Body=body_bytes,
    ContentType="application/json"
)
print(f"Upload successful : {result['ResponseMetadata']['HTTPStatusCode']}")

Define permissions (ACL) during upload

You can specify access permissions (ACL) directly when calling put_object:

result = s3.put_object(
    Bucket=bucket_name,
    Key=key,
    Body=body_bytes,
    ContentType="application/json",
    ACL="public-read"  # Make the object publicly readable
)

3. Read an object (Download)

To retrieve the content of an object, use the get_object method. It returns an object of type StreamingBody that you can read to obtain the raw data.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import boto3
from botocore.config import Config

# Initialisation du client S3 (SigV2 recommandé pour GET)
s3 = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=Config(
        region_name='fr1',
        signature_version='s3',
        s3={'addressing_style': 'path'}
    )
)

response = s3.get_object(Bucket=bucket_name, Key='users/dupont_claire.json')
content = response['Body'].read().decode('utf-8')
print(content)

4. Delete an object

The delete_object method allows you to permanently delete an object from a bucket.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import boto3
from botocore.config import Config

# Initialize the S3 client (SigV2 recommended for DELETE)
s3 = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=Config(
        region_name='fr1',
        signature_version='s3',
        s3={'addressing_style': 'path'}
    )
)

key = "users/dupont_claire.json"
result = s3.delete_object(Bucket=bucket_name, Key=key)
print(f"Deletion: {result['ResponseMetadata']['HTTPStatusCode']}")

5. Make an object public (Read All)

If an object already exists, you can modify its permissions to make it publicly readable using the put_object_acl method.

Attention

Making an object public means that anyone with the URL can access it. Use this option with caution and in accordance with your security policy.

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import boto3
from botocore.config import Config

# Configuration S3
s3 = boto3.client(
    's3',
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key,
    endpoint_url=endpoint_url,
    config=Config(
        region_name='fr1',
        signature_version='s3',
        s3={'addressing_style': 'path'}
    )
)

key = "users/dupont_claire.json"
s3.put_object_acl(Bucket=bucket_name, Key=key, ACL="public-read")

public_url = f"{endpoint_url.rstrip('/')}/{bucket_name}/{key}"
print("Public URL:", public_url)

Troubleshooting

Error `XAmzContentSHA256Mismatch`

If you encounter this error during PUT operations:

An error occurred (XAmzContentSHA256Mismatch) when calling the PutObject operation

Cause: The Cloud Temple object storage infrastructure (Dell ECS) may have compatibility issues with the SHA-256 hash calculation used by SigV4 for write operations.

Solution: Use SigV2 signature for data operations:

config = Config(
    region_name='fr1',
    signature_version='s3',       # Use SigV2 instead of s3v4
    s3={'addressing_style': 'path'}
)

Error `SignatureDoesNotMatch`

If you receive this error, check the following:

That your access keys (access_key_id and secret_access_key) are correct
That the region is properly set to fr1
That you are using the correct endpoint (format: https://VOTRE_NAMESPACE.s3.fr1.cloud-temple.com)

Bucket not found

Make sure to use path-style addressing (s3={'addressing_style': 'path'}) in the configuration. Virtual-hosted style will be supported in S2 2026.

Installation​

Boto3 Client Configuration​

Signature specifics (SigV2 / SigV4)​

1. List files in an S3 bucket​

2. Upload an object (Upload)​

Define permissions (ACL) during upload​

3. Read an object (Download)​

4. Delete an object​

5. Make an object public (Read All)​

Troubleshooting​

Error XAmzContentSHA256Mismatch​

Error SignatureDoesNotMatch​

Bucket not found​