1
0
Fork 0
mirror of https://github.com/ansible-collections/community.general.git synced 2024-09-14 20:13:21 +02:00

[cloud] Convert s3 module to use boto3 (#21529)

* replace boto with boto3 for the s3 module

make s3 pep8 and remove from legacy files

fix s3 unit tests

* fix indentation

* s3 module - if we can't create an MD5 sum return None and always upload file

* remove Location.DEFAULT which isn't used in boto3 and tidy up the docs

* pep8

* s3: remove default: null, empty aliases, and required: false from documentation

fix incorrectly documented defaults

* Porting s3 to boto3. Simplify some logic and remove unused imports

* Fix s3 module variables

* Fix a typo in s3 module and remove from pep8 legacy files

* s3: add pagination for listing objects.

Fix logic and use head_object instead of get_object for efficiency.

Fix typo in unit test.

* Fix pagination to maintain backwards compatibility.

Fix incorrect conditional.

Remove redundant variable assignment.

Fix s3 list_object pagination to return all pages

* Use the revised List Objects API as recommended.

* Wrap call to paginated_list in a try/except

Also remembered to allow marker/prefix/max_keys to modify what keys are listed

* Simplify argument
This commit is contained in:
Sloane Hertel 2017-08-11 09:34:33 -04:00 committed by Ryan Brown
parent 7d1308b0d8
commit 1de91a9aa0
4 changed files with 289 additions and 311 deletions

View file

@ -24,128 +24,100 @@ DOCUMENTATION = '''
module: s3 module: s3
short_description: manage objects in S3. short_description: manage objects in S3.
description: description:
- This module allows the user to manage S3 buckets and the objects within them. Includes support for creating and deleting both objects and buckets, - This module allows the user to manage S3 buckets and the objects within them. Includes support for creating and
retrieving objects as files or strings and generating download links. This module has a dependency on python-boto. deleting both objects and buckets, retrieving objects as files or strings and generating download links.
This module has a dependency on boto3 and botocore.
version_added: "1.1" version_added: "1.1"
options: options:
aws_access_key: aws_access_key:
description: description:
- AWS access key id. If not set then the value of the AWS_ACCESS_KEY environment variable is used. - AWS access key id. If not set then the value of the AWS_ACCESS_KEY environment variable is used.
required: false
default: null
aliases: [ 'ec2_access_key', 'access_key' ] aliases: [ 'ec2_access_key', 'access_key' ]
aws_secret_key: aws_secret_key:
description: description:
- AWS secret key. If not set then the value of the AWS_SECRET_KEY environment variable is used. - AWS secret key. If not set then the value of the AWS_SECRET_KEY environment variable is used.
required: false
default: null
aliases: ['ec2_secret_key', 'secret_key'] aliases: ['ec2_secret_key', 'secret_key']
bucket: bucket:
description: description:
- Bucket name. - Bucket name.
required: true required: true
default: null
aliases: []
dest: dest:
description: description:
- The destination file path when downloading an object/key with a GET operation. - The destination file path when downloading an object/key with a GET operation.
required: false
aliases: []
version_added: "1.3" version_added: "1.3"
encrypt: encrypt:
description: description:
- When set for PUT mode, asks for server-side encryption - When set for PUT mode, asks for server-side encryption.
required: false default: True
default: no
version_added: "2.0" version_added: "2.0"
expiration: expiration:
description: description:
- Time limit (in seconds) for the URL generated and returned by S3/Walrus when performing a mode=put or mode=geturl operation. - Time limit (in seconds) for the URL generated and returned by S3/Walrus when performing a mode=put or mode=geturl operation.
required: false
default: 600 default: 600
aliases: []
headers: headers:
description: description:
- Custom headers for PUT operation, as a dictionary of 'key=value' and 'key=value,key=value'. - Custom headers for PUT operation, as a dictionary of 'key=value' and 'key=value,key=value'.
required: false
default: null
version_added: "2.0" version_added: "2.0"
marker: marker:
description: description:
- Specifies the key to start with when using list mode. Object keys are returned in alphabetical order, starting with key after the marker in order. - Specifies the key to start with when using list mode. Object keys are returned in alphabetical order, starting with key after the marker in order.
required: false
default: null
version_added: "2.0" version_added: "2.0"
max_keys: max_keys:
description: description:
- Max number of results to return in list mode, set this if you want to retrieve fewer than the default 1000 keys. - Max number of results to return in list mode, set this if you want to retrieve fewer than the default 1000 keys.
required: false
default: 1000 default: 1000
version_added: "2.0" version_added: "2.0"
metadata: metadata:
description: description:
- Metadata for PUT operation, as a dictionary of 'key=value' and 'key=value,key=value'. - Metadata for PUT operation, as a dictionary of 'key=value' and 'key=value,key=value'.
required: false
default: null
version_added: "1.6" version_added: "1.6"
mode: mode:
description: description:
- Switches the module behaviour between put (upload), get (download), geturl (return download url, Ansible 1.3+), - Switches the module behaviour between put (upload), get (download), geturl (return download url, Ansible 1.3+),
getstr (download object as string (1.3+)), list (list keys, Ansible 2.0+), create (bucket), delete (bucket), and delobj (delete object, Ansible 2.0+). getstr (download object as string (1.3+)), list (list keys, Ansible 2.0+), create (bucket), delete (bucket),
and delobj (delete object, Ansible 2.0+).
required: true required: true
choices: ['get', 'put', 'delete', 'create', 'geturl', 'getstr', 'delobj', 'list'] choices: ['get', 'put', 'delete', 'create', 'geturl', 'getstr', 'delobj', 'list']
object: object:
description: description:
- Keyname of the object inside the bucket. Can be used to create "virtual directories", see examples. - Keyname of the object inside the bucket. Can be used to create "virtual directories", see examples.
required: false
default: null
permission: permission:
description: description:
- This option lets the user set the canned permissions on the object/bucket that are created. - This option lets the user set the canned permissions on the object/bucket that are created.
The permissions that can be set are 'private', 'public-read', 'public-read-write', 'authenticated-read'. Multiple permissions can be The permissions that can be set are 'private', 'public-read', 'public-read-write', 'authenticated-read' for a bucket or
specified as a list. 'private', 'public-read', 'public-read-write', 'aws-exec-read', 'authenticated-read', 'bucket-owner-read',
required: false 'bucket-owner-full-control' for an object. Multiple permissions can be specified as a list.
default: private default: private
version_added: "2.0" version_added: "2.0"
prefix: prefix:
description: description:
- Limits the response to keys that begin with the specified prefix for list mode - Limits the response to keys that begin with the specified prefix for list mode
required: false default: ""
default: null
version_added: "2.0" version_added: "2.0"
version: version:
description: description:
- Version ID of the object inside the bucket. Can be used to get a specific version of a file if versioning is enabled in the target bucket. - Version ID of the object inside the bucket. Can be used to get a specific version of a file if versioning is enabled in the target bucket.
required: false
default: null
aliases: []
version_added: "2.0" version_added: "2.0"
overwrite: overwrite:
description: description:
- Force overwrite either locally on the filesystem or remotely with the object/key. Used with PUT and GET operations. - Force overwrite either locally on the filesystem or remotely with the object/key. Used with PUT and GET operations.
Boolean or one of [always, never, different], true is equal to 'always' and false is equal to 'never', new in 2.0 Boolean or one of [always, never, different], true is equal to 'always' and false is equal to 'never', new in 2.0
required: false
default: 'always' default: 'always'
version_added: "1.2" version_added: "1.2"
region: region:
description: description:
- > - "AWS region to create the bucket in. If not set then the value of the AWS_REGION and EC2_REGION environment variables
AWS region to create the bucket in. If not set then the value of the AWS_REGION and EC2_REGION environment variables are checked, are checked, followed by the aws_region and ec2_region settings in the Boto config file. If none of those are set the
followed by the aws_region and ec2_region settings in the Boto config file. If none of those are set the region defaults to the region defaults to the S3 Location: US Standard. Prior to ansible 1.8 this parameter could be specified but had no effect."
S3 Location: US Standard. Prior to ansible 1.8 this parameter could be specified but had no effect.
required: false
default: null
version_added: "1.8" version_added: "1.8"
retries: retries:
description: description:
- On recoverable failure, how many times to retry before actually failing. - On recoverable failure, how many times to retry before actually failing.
required: false
default: 0 default: 0
version_added: "2.0" version_added: "2.0"
s3_url: s3_url:
description: description:
- S3 URL endpoint for usage with Ceph, Eucalypus, fakes3, etc. Otherwise assumes AWS - S3 URL endpoint for usage with Ceph, Eucalypus, fakes3, etc. Otherwise assumes AWS
default: null
aliases: [ S3_URL ] aliases: [ S3_URL ]
rgw: rgw:
description: description:
@ -155,22 +127,18 @@ options:
src: src:
description: description:
- The source file path when performing a PUT operation. - The source file path when performing a PUT operation.
required: false
default: null
aliases: []
version_added: "1.3" version_added: "1.3"
ignore_nonexistent_bucket: ignore_nonexistent_bucket:
description: description:
- > - "Overrides initial bucket lookups in case bucket or iam policies are restrictive. Example: a user may have the
Overrides initial bucket lookups in case bucket or iam policies are restrictive. Example: a user may have the GetObject permission but no other GetObject permission but no other permissions. In this case using the option mode: get will fail without specifying
permissions. In this case using the option mode: get will fail without specifying ignore_nonexistent_bucket: True. ignore_nonexistent_bucket: True."
default: false
aliases: []
version_added: "2.3" version_added: "2.3"
requirements: [ "boto" ] requirements: [ "boto3", "botocore" ]
author: author:
- "Lester Wade (@lwade)" - "Lester Wade (@lwade)"
- "Sloane Hertel (@s-hertel)"
extends_documentation_fragment: aws extends_documentation_fragment: aws
''' '''
@ -272,131 +240,150 @@ import os
import traceback import traceback
from ansible.module_utils.six.moves.urllib.parse import urlparse from ansible.module_utils.six.moves.urllib.parse import urlparse
from ssl import SSLError from ssl import SSLError
from ansible.module_utils.basic import AnsibleModule, to_text, to_native
from ansible.module_utils.ec2 import ec2_argument_spec, camel_dict_to_snake_dict, get_aws_connection_info, boto3_conn, HAS_BOTO3
try: try:
import boto import botocore
import boto.ec2
from boto.s3.connection import Location
from boto.s3.connection import OrdinaryCallingFormat
from boto.s3.connection import S3Connection
from boto.s3.acl import CannedACLStrings
HAS_BOTO = True
except ImportError: except ImportError:
HAS_BOTO = False pass # will be detected by imported HAS_BOTO3
def key_check(module, s3, bucket, obj, version=None, validate=True): def key_check(module, s3, bucket, obj, version=None, validate=True):
exists = True
try: try:
bucket = s3.lookup(bucket, validate=validate) if version:
key_check = bucket.get_key(obj, version_id=version) s3.head_object(Bucket=bucket, Key=obj, VersionId=version)
except s3.provider.storage_response_error as e:
if version is not None and e.status == 400: # If a specified version doesn't exist a 400 is returned.
key_check = None
else: else:
module.fail_json(msg=str(e)) s3.head_object(Bucket=bucket, Key=obj)
if key_check: except botocore.exceptions.ClientError as e:
return True # if a client error is thrown, check if it's a 404 error
else: # if it's a 404 error, then the object does not exist
return False error_code = int(e.response['Error']['Code'])
if error_code == 404:
exists = False
elif error_code == 403 and validate is False:
pass
else:
module.fail_json(msg="Failed while looking up object (during key check) %s." % obj,
exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
return exists
def keysum(module, s3, bucket, obj, version=None, validate=True):
bucket = s3.lookup(bucket, validate=validate) def keysum(module, s3, bucket, obj, version=None):
key_check = bucket.get_key(obj, version_id=version) if version:
key_check = s3.head_object(Bucket=bucket, Key=obj, VersionId=version)
else:
key_check = s3.head_object(Bucket=bucket, Key=obj)
if not key_check: if not key_check:
return None return None
md5_remote = key_check.etag[1:-1] md5_remote = key_check['ETag'][1:-1]
etag_multipart = '-' in md5_remote # Check for multipart, etag is not md5 if '-' in md5_remote: # Check for multipart, etag is not md5
if etag_multipart is True: return None
module.fail_json(msg="Files uploaded with multipart of s3 are not supported with checksum, unable to compute checksum.")
return md5_remote return md5_remote
def bucket_check(module, s3, bucket, validate=True): def bucket_check(module, s3, bucket, validate=True):
exists = True
try: try:
result = s3.lookup(bucket, validate=validate) s3.head_bucket(Bucket=bucket)
except s3.provider.storage_response_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg="Failed while looking up bucket (during bucket_check) %s: %s" % (bucket, e), # If a client error is thrown, then check that it was a 404 error.
exception=traceback.format_exc()) # If it was a 404 error, then the bucket does not exist.
return bool(result) error_code = int(e.response['Error']['Code'])
if error_code == 404:
exists = False
elif error_code == 403 and validate is False:
pass
else:
module.fail_json(msg="Failed while looking up bucket (during bucket_check) %s." % bucket,
exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
except botocore.exceptions.EndpointConnectionError as e:
module.fail_json(msg="Invalid endpoint provided: %s" % to_text(e), exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
return exists
def create_bucket(module, s3, bucket, location=None): def create_bucket(module, s3, bucket, location=None):
if module.check_mode: if module.check_mode:
module.exit_json(msg="PUT operation skipped - running in check mode", changed=True) module.exit_json(msg="PUT operation skipped - running in check mode", changed=True)
if location is None: configuration = {}
location = Location.DEFAULT if location not in ('us-east-1', None):
configuration['LocationConstraint'] = location
try: try:
bucket = s3.create_bucket(bucket, location=location) if len(configuration) > 0:
s3.create_bucket(Bucket=bucket, CreateBucketConfiguration=configuration)
else:
s3.create_bucket(Bucket=bucket)
for acl in module.params.get('permission'): for acl in module.params.get('permission'):
bucket.set_acl(acl) s3.put_bucket_acl(ACL=acl, Bucket=bucket)
except s3.provider.storage_response_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg="Failed while creating bucket or setting acl (check that you have CreateBucket and PutBucketAcl permission) %s: %s" % (bucket, e), module.fail_json(msg="Failed while creating bucket or setting acl (check that you have CreateBucket and PutBucketAcl permission).",
exception=traceback.format_exc()) exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
if bucket: if bucket:
return True return True
def get_bucket(module, s3, bucket):
def paginated_list(s3, **pagination_params):
pg = s3.get_paginator('list_objects_v2')
for page in pg.paginate(**pagination_params):
for data in page.get('Contents', {}):
yield data['Key']
def list_keys(module, s3, bucket, prefix, marker, max_keys):
pagination_params = {'Bucket': bucket}
for param_name, param_value in (('Prefix', prefix), ('StartAfter', marker), ('MaxKeys', max_keys)):
pagination_params[param_name] = param_value
try: try:
return s3.lookup(bucket) keys = [key for key in paginated_list(s3, **pagination_params)]
except s3.provider.storage_response_error as e: module.exit_json(msg="LIST operation complete", s3_keys=keys)
module.fail_json(msg="Failed while getting bucket %s: %s" % (bucket, e), except botocore.exceptions.ClientError as e:
exception=traceback.format_exc()) module.fail_json(msg="Failed while listing the keys in the bucket {0}".format(bucket),
exception=traceback.format_exc(),
**camel_dict_to_snake_dict(e.response))
def list_keys(module, bucket_object, prefix, marker, max_keys):
all_keys = bucket_object.get_all_keys(prefix=prefix, marker=marker, max_keys=max_keys)
keys = [x.key for x in all_keys]
module.exit_json(msg="LIST operation complete", s3_keys=keys)
def delete_bucket(module, s3, bucket): def delete_bucket(module, s3, bucket):
if module.check_mode: if module.check_mode:
module.exit_json(msg="DELETE operation skipped - running in check mode", changed=True) module.exit_json(msg="DELETE operation skipped - running in check mode", changed=True)
try: try:
bucket = s3.lookup(bucket, validate=False) exists = bucket_check(module, s3, bucket)
bucket_contents = bucket.list() if exists is False:
bucket.delete_keys([key.name for key in bucket_contents])
except s3.provider.storage_response_error as e:
if e.status == 404:
# bucket doesn't appear to exist
return False return False
elif e.status == 403: # if there are contents then we need to delete them before we can delete the bucket
# bucket appears to exist but user doesn't have list bucket permission; may still be able to delete bucket keys = [{'Key': key} for key in paginated_list(s3, Bucket=bucket)]
pass if keys:
else: s3.delete_objects(Bucket=bucket, Delete={'Objects': keys})
module.fail_json(msg=str(e), exception=traceback.format_exc()) s3.delete_bucket(Bucket=bucket)
try:
bucket.delete()
return True return True
except s3.provider.storage_response_error as e: except botocore.exceptions.ClientError as e:
if e.status == 403: module.fail_json(msg="Failed while deleting bucket %s.", exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
module.exit_json(msg="Unable to complete DELETE operation. Check you have have s3:DeleteBucket "
"permission. Error: {0}.".format(e.message),
exception=traceback.format_exc())
elif e.status == 409:
module.exit_json(msg="Unable to complete DELETE operation. It appears there are contents in the "
"bucket that you don't have permission to delete. Error: {0}.".format(e.message),
exception=traceback.format_exc())
else:
module.fail_json(msg=str(e), exception=traceback.format_exc())
def delete_key(module, s3, bucket, obj, validate=True):
def delete_key(module, s3, bucket, obj):
if module.check_mode: if module.check_mode:
module.exit_json(msg="DELETE operation skipped - running in check mode", changed=True) module.exit_json(msg="DELETE operation skipped - running in check mode", changed=True)
try: try:
bucket = s3.lookup(bucket, validate=validate) s3.delete_object(Bucket=bucket, Key=obj)
bucket.delete_key(obj) module.exit_json(msg="Object deleted from bucket %s." % (bucket), changed=True)
module.exit_json(msg="Object deleted from bucket %s"%bucket, changed=True) except botocore.exceptions.ClientError as e:
except s3.provider.storage_response_error as e: module.fail_json(msg="Failed while trying to delete %s." % obj, exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
module.fail_json(msg= str(e))
def create_dirkey(module, s3, bucket, obj, validate=True):
def create_dirkey(module, s3, bucket, obj):
if module.check_mode: if module.check_mode:
module.exit_json(msg="PUT operation skipped - running in check mode", changed=True) module.exit_json(msg="PUT operation skipped - running in check mode", changed=True)
try: try:
bucket = s3.lookup(bucket, validate=validate) bucket = s3.Bucket(bucket)
key = bucket.new_key(obj) key = bucket.new_key(obj)
key.set_contents_from_string('') key.set_contents_from_string('')
for acl in module.params.get('permission'):
s3.put_object_acl(ACL=acl, Bucket=bucket, Key=obj)
module.exit_json(msg="Virtual directory %s created in bucket %s" % (obj, bucket.name), changed=True) module.exit_json(msg="Virtual directory %s created in bucket %s" % (obj, bucket.name), changed=True)
except s3.provider.storage_response_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg= str(e)) module.fail_json(msg="Failed while creating object %s." % obj, exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
def path_check(path): def path_check(path):
if os.path.exists(path): if os.path.exists(path):
@ -405,63 +392,80 @@ def path_check(path):
return False return False
def upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers, validate=True): def upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers):
if module.check_mode: if module.check_mode:
module.exit_json(msg="PUT operation skipped - running in check mode", changed=True) module.exit_json(msg="PUT operation skipped - running in check mode", changed=True)
try: try:
bucket = s3.lookup(bucket, validate=validate)
key = bucket.new_key(obj)
if metadata: if metadata:
for meta_key in metadata.keys(): extra = {'Metadata': dict(metadata)}
key.set_metadata(meta_key, metadata[meta_key]) s3.upload_file(Filename=src, Bucket=bucket, Key=obj, ExtraArgs=extra)
else:
key.set_contents_from_filename(src, encrypt_key=encrypt, headers=headers) s3.upload_file(Filename=src, Bucket=bucket, Key=obj)
for acl in module.params.get('permission'): for acl in module.params.get('permission'):
key.set_acl(acl) s3.put_object_acl(ACL=acl, Bucket=bucket, Key=obj)
url = key.generate_url(expiry) url = s3.generate_presigned_url(ClientMethod='put_object',
Params={'Bucket': bucket, 'Key': obj},
ExpiresIn=expiry)
module.exit_json(msg="PUT operation complete", url=url, changed=True) module.exit_json(msg="PUT operation complete", url=url, changed=True)
except s3.provider.storage_copy_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg= str(e)) module.fail_json(msg="Unable to complete PUT operation.", exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
def download_s3file(module, s3, bucket, obj, dest, retries, version=None, validate=True):
def download_s3file(module, s3, bucket, obj, dest, retries, version=None):
if module.check_mode: if module.check_mode:
module.exit_json(msg="GET operation skipped - running in check mode", changed=True) module.exit_json(msg="GET operation skipped - running in check mode", changed=True)
# retries is the number of loops; range/xrange needs to be one # retries is the number of loops; range/xrange needs to be one
# more to get that count of loops. # more to get that count of loops.
bucket = s3.lookup(bucket, validate=validate) try:
key = bucket.get_key(obj, version_id=version) if version:
key = s3.get_object(Bucket=bucket, Key=obj, VersionId=version)
else:
key = s3.get_object(Bucket=bucket, Key=obj)
except botocore.exceptions.ClientError as e:
if e.response['Error']['Code'] != "404":
module.fail_json(msg="Could not find the key %s." % obj, exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
for x in range(0, retries + 1): for x in range(0, retries + 1):
try: try:
key.get_contents_to_filename(dest) s3.download_file(bucket, obj, dest)
module.exit_json(msg="GET operation complete", changed=True) module.exit_json(msg="GET operation complete", changed=True)
except s3.provider.storage_copy_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg= str(e))
except SSLError as e:
# actually fail on last pass through the loop. # actually fail on last pass through the loop.
if x >= retries: if x >= retries:
module.fail_json(msg="s3 download failed; %s" % e) module.fail_json(msg="Failed while downloading %s." % obj, exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
# otherwise, try again, this may be a transient timeout. # otherwise, try again, this may be a transient timeout.
pass pass
except SSLError as e: # will ClientError catch SSLError?
# actually fail on last pass through the loop.
if x >= retries:
module.fail_json(msg="s3 download failed: %s." % e, exception=traceback.format_exc())
# otherwise, try again, this may be a transient timeout.
pass
def download_s3str(module, s3, bucket, obj, version=None, validate=True): def download_s3str(module, s3, bucket, obj, version=None, validate=True):
if module.check_mode: if module.check_mode:
module.exit_json(msg="GET operation skipped - running in check mode", changed=True) module.exit_json(msg="GET operation skipped - running in check mode", changed=True)
try: try:
bucket = s3.lookup(bucket, validate=validate) if version:
key = bucket.get_key(obj, version_id=version) contents = to_native(s3.get_object(Bucket=bucket, Key=obj, VersionId=version)["Body"].read())
contents = key.get_contents_as_string() else:
contents = to_native(s3.get_object(Bucket=bucket, Key=obj)["Body"].read())
module.exit_json(msg="GET operation complete", contents=contents, changed=True) module.exit_json(msg="GET operation complete", contents=contents, changed=True)
except s3.provider.storage_copy_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg= str(e)) module.fail_json(msg="Failed while getting contents of object %s as a string." % obj,
exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
def get_download_url(module, s3, bucket, obj, expiry, changed=True, validate=True):
def get_download_url(module, s3, bucket, obj, expiry, changed=True):
try: try:
bucket = s3.lookup(bucket, validate=validate) url = s3.generate_presigned_url(ClientMethod='get_object',
key = bucket.lookup(obj) Params={'Bucket': bucket, 'Key': obj},
url = key.generate_url(expiry) ExpiresIn=expiry)
module.exit_json(msg="Download url:", url=url, expiry=expiry, changed=changed) module.exit_json(msg="Download url:", url=url, expiry=expiry, changed=changed)
except s3.provider.storage_response_error as e: except botocore.exceptions.ClientError as e:
module.fail_json(msg= str(e)) module.fail_json(msg="Failed while getting download url.", exception=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
def is_fakes3(s3_url): def is_fakes3(s3_url):
""" Return True if s3_url has scheme fakes3:// """ """ Return True if s3_url has scheme fakes3:// """
@ -470,6 +474,7 @@ def is_fakes3(s3_url):
else: else:
return False return False
def is_walrus(s3_url): def is_walrus(s3_url):
""" Return True if it's Walrus endpoint, not S3 """ Return True if it's Walrus endpoint, not S3
@ -481,28 +486,51 @@ def is_walrus(s3_url):
return False return False
def get_s3_connection(module, aws_connect_kwargs, location, rgw, s3_url):
if s3_url and rgw: # TODO - test this
rgw = urlparse(s3_url)
params = dict(module=module, conn_type='client', resource='s3', use_ssl=rgw.scheme == 'https', region=location, endpoint=s3_url, **aws_connect_kwargs)
elif is_fakes3(s3_url):
for kw in ['is_secure', 'host', 'port'] and list(aws_connect_kwargs.keys()):
del aws_connect_kwargs[kw]
fakes3 = urlparse(s3_url)
if fakes3.scheme == 'fakes3s':
protocol = "https"
else:
protocol = "http"
params = dict(service_name='s3', endpoint_url="%s://%s:%s" % (protocol, fakes3.hostname, to_text(fakes3.port)),
use_ssl=fakes3.scheme == 'fakes3s', region_name=None, **aws_connect_kwargs)
elif is_walrus(s3_url):
walrus = urlparse(s3_url).hostname
params = dict(module=module, conn_type='client', resource='s3', region=location, endpoint=walrus, **aws_connect_kwargs)
else:
params = dict(module=module, conn_type='client', resource='s3', region=location, endpoint=s3_url, **aws_connect_kwargs)
return boto3_conn(**params)
def main(): def main():
argument_spec = ec2_argument_spec() argument_spec = ec2_argument_spec()
argument_spec.update(dict( argument_spec.update(
bucket = dict(required=True), dict(
dest = dict(default=None), bucket=dict(required=True),
encrypt = dict(default=True, type='bool'), dest=dict(default=None),
expiry = dict(default=600, aliases=['expiration']), encrypt=dict(default=True, type='bool'),
headers = dict(type='dict'), expiry=dict(default=600, type='int', aliases=['expiration']),
marker = dict(default=None), headers=dict(type='dict'),
max_keys = dict(default=1000), marker=dict(default=""),
metadata = dict(type='dict'), max_keys=dict(default=1000, type='int'),
mode = dict(choices=['get', 'put', 'delete', 'create', 'geturl', 'getstr', 'delobj', 'list'], required=True), metadata=dict(type='dict'),
object = dict(), mode=dict(choices=['get', 'put', 'delete', 'create', 'geturl', 'getstr', 'delobj', 'list'], required=True),
permission = dict(type='list', default=['private']), object=dict(),
version = dict(default=None), permission=dict(type='list', default=['private']),
overwrite = dict(aliases=['force'], default='always'), version=dict(default=None),
prefix = dict(default=None), overwrite=dict(aliases=['force'], default='always'),
retries = dict(aliases=['retry'], type='int', default=0), prefix=dict(default=""),
s3_url = dict(aliases=['S3_URL']), retries=dict(aliases=['retry'], type='int', default=0),
rgw = dict(default='no', type='bool'), s3_url=dict(aliases=['S3_URL']),
src = dict(), rgw=dict(default='no', type='bool'),
ignore_nonexistent_bucket = dict(default=False, type='bool') src=dict(),
ignore_nonexistent_bucket=dict(default=False, type='bool')
), ),
) )
module = AnsibleModule( module = AnsibleModule(
@ -510,12 +538,12 @@ def main():
supports_check_mode=True, supports_check_mode=True,
) )
if not HAS_BOTO: if not HAS_BOTO3:
module.fail_json(msg='boto required for this module') module.fail_json(msg='boto3 and botocore required for this module')
bucket = module.params.get('bucket') bucket = module.params.get('bucket')
encrypt = module.params.get('encrypt') encrypt = module.params.get('encrypt')
expiry = int(module.params['expiry']) expiry = module.params.get('expiry')
dest = module.params.get('dest', '') dest = module.params.get('dest', '')
headers = module.params.get('headers') headers = module.params.get('headers')
marker = module.params.get('marker') marker = module.params.get('marker')
@ -535,9 +563,8 @@ def main():
if dest: if dest:
dest = os.path.expanduser(dest) dest = os.path.expanduser(dest)
for acl in module.params.get('permission'): object_canned_acl = ["private", "public-read", "public-read-write", "aws-exec-read", "authenticated-read", "bucket-owner-read", "bucket-owner-full-control"]
if acl not in CannedACLStrings: bucket_canned_acl = ["private", "public-read", "public-read-write", "authenticated-read"]
module.fail_json(msg='Unknown permission specified: %s' % str(acl))
if overwrite not in ['always', 'never', 'different']: if overwrite not in ['always', 'never', 'different']:
if module.boolean(overwrite): if module.boolean(overwrite):
@ -545,11 +572,11 @@ def main():
else: else:
overwrite = 'never' overwrite = 'never'
region, ec2_url, aws_connect_kwargs = get_aws_connection_info(module) region, ec2_url, aws_connect_kwargs = get_aws_connection_info(module, boto3=True)
if region in ('us-east-1', '', None): if region in ('us-east-1', '', None):
# S3ism for the US Standard region # default to US Standard region
location = Location.DEFAULT location = 'us-east-1'
else: else:
# Boto uses symbolic names for locations but region strings will # Boto uses symbolic names for locations but region strings will
# actually work fine for everything except us-east-1 (US Standard) # actually work fine for everything except us-east-1 (US Standard)
@ -570,113 +597,105 @@ def main():
if rgw and not s3_url: if rgw and not s3_url:
module.fail_json(msg='rgw flavour requires s3_url') module.fail_json(msg='rgw flavour requires s3_url')
# bucket names with .'s in them need to use the calling_format option,
# otherwise the connection will fail. See https://github.com/boto/boto/issues/2836
# for more details.
if '.' in bucket:
aws_connect_kwargs['calling_format'] = OrdinaryCallingFormat()
# Look at s3_url and tweak connection settings # Look at s3_url and tweak connection settings
# if connecting to RGW, Walrus or fakes3 # if connecting to RGW, Walrus or fakes3
for key in ['validate_certs', 'security_token', 'profile_name']:
aws_connect_kwargs.pop(key, None)
try: try:
s3 = get_s3_connection(aws_connect_kwargs, location, rgw, s3_url) s3 = get_s3_connection(module, aws_connect_kwargs, location, rgw, s3_url)
except (botocore.exceptions.NoCredentialsError, botocore.exceptions.ProfileNotFound) as e:
module.fail_json(msg="Can't authorize connection. Check your credentials and profile.",
exceptions=traceback.format_exc(), **camel_dict_to_snake_dict(e.response))
except boto.exception.NoAuthHandlerFound as e: validate = not ignore_nonexistent_bucket
module.fail_json(msg='No Authentication Handler found: %s ' % str(e))
except Exception as e:
module.fail_json(msg='Failed to connect to S3: %s' % str(e))
if s3 is None: # this should never happen # separate types of ACLs
module.fail_json(msg ='Unknown error, failed to create s3 connection, no information from boto.') bucket_acl = [acl for acl in module.params.get('permission') if acl in bucket_canned_acl]
object_acl = [acl for acl in module.params.get('permission') if acl in object_canned_acl]
error_acl = [acl for acl in module.params.get('permission') if acl not in bucket_canned_acl and acl not in object_canned_acl]
if error_acl:
module.fail_json(msg='Unknown permission specified: %s' % error_acl)
# First, we check to see if the bucket exists, we get "bucket" returned. # First, we check to see if the bucket exists, we get "bucket" returned.
bucketrtn = bucket_check(module, s3, bucket) bucketrtn = bucket_check(module, s3, bucket, validate=validate)
if not ignore_nonexistent_bucket: if validate and mode not in ('create', 'put', 'delete') and not bucketrtn:
validate = True module.fail_json(msg="Source bucket cannot be found.")
if mode not in ('create', 'put', 'delete') and not bucketrtn:
module.fail_json(msg="Source bucket cannot be found.")
else:
validate = False
# If our mode is a GET operation (download), go through the procedure as appropriate ... # If our mode is a GET operation (download), go through the procedure as appropriate ...
if mode == 'get': if mode == 'get':
# Next, we check to see if the key in the bucket exists. If it exists, it also returns key_matches md5sum check. # Next, we check to see if the key in the bucket exists. If it exists, it also returns key_matches md5sum check.
keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate) keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate)
if keyrtn is False: if keyrtn is False:
if version is not None: module.fail_json(msg="Key %s with version id %s does not exist." % (obj, version))
module.fail_json(msg="Key %s with version id %s does not exist."% (obj, version))
else:
module.fail_json(msg="Key %s or source bucket %s does not exist."% (obj, bucket))
# If the destination path doesn't exist or overwrite is True, no need to do the md5um etag check, so just download. # If the destination path doesn't exist or overwrite is True, no need to do the md5um etag check, so just download.
pathrtn = path_check(dest)
# Compare the remote MD5 sum of the object with the local dest md5sum, if it already exists. # Compare the remote MD5 sum of the object with the local dest md5sum, if it already exists.
if pathrtn is True: if path_check(dest):
md5_remote = keysum(module, s3, bucket, obj, version=version, validate=validate) # Determine if the remote and local object are identical
md5_local = module.md5(dest) if keysum(module, s3, bucket, obj, version=version) == module.md5(dest):
if md5_local == md5_remote:
sum_matches = True sum_matches = True
if overwrite == 'always': if overwrite == 'always':
download_s3file(module, s3, bucket, obj, dest, retries, version=version, validate=validate) download_s3file(module, s3, bucket, obj, dest, retries, version=version)
else: else:
module.exit_json(msg="Local and remote object are identical, ignoring. Use overwrite=always parameter to force.", changed=False) module.exit_json(msg="Local and remote object are identical, ignoring. Use overwrite=always parameter to force.", changed=False)
else: else:
sum_matches = False sum_matches = False
if overwrite in ('always', 'different'): if overwrite in ('always', 'different'):
download_s3file(module, s3, bucket, obj, dest, retries, version=version, validate=validate) download_s3file(module, s3, bucket, obj, dest, retries, version=version)
else: else:
module.exit_json(msg="WARNING: Checksums do not match. Use overwrite parameter to force download.") module.exit_json(msg="WARNING: Checksums do not match. Use overwrite parameter to force download.")
else: else:
download_s3file(module, s3, bucket, obj, dest, retries, version=version, validate=validate) download_s3file(module, s3, bucket, obj, dest, retries, version=version)
# Firstly, if key_matches is TRUE and overwrite is not enabled, we EXIT with a helpful message.
if sum_matches and overwrite == 'never':
module.exit_json(msg="Local and remote object are identical, ignoring. Use overwrite parameter to force.", changed=False)
# if our mode is a PUT operation (upload), go through the procedure as appropriate ... # if our mode is a PUT operation (upload), go through the procedure as appropriate ...
if mode == 'put': if mode == 'put':
# Use this snippet to debug through conditionals: # if putting an object in a bucket yet to be created, acls for the bucket and/or the object may be specified
# module.exit_json(msg="Bucket return %s"%bucketrtn) # these were separated into the variables bucket_acl and object_acl above
# Lets check the src path. # Lets check the src path.
pathrtn = path_check(src) if not path_check(src):
if not pathrtn:
module.fail_json(msg="Local object for PUT does not exist") module.fail_json(msg="Local object for PUT does not exist")
# Lets check to see if bucket exists to get ground truth. # Lets check to see if bucket exists to get ground truth.
if bucketrtn: if bucketrtn:
keyrtn = key_check(module, s3, bucket, obj) keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate)
# Lets check key state. Does it exist and if it does, compute the etag md5sum. # Lets check key state. Does it exist and if it does, compute the etag md5sum.
if bucketrtn and keyrtn: if bucketrtn and keyrtn:
md5_remote = keysum(module, s3, bucket, obj) # Compare the local and remote object
md5_local = module.md5(src) if module.md5(src) == keysum(module, s3, bucket, obj):
if md5_local == md5_remote:
sum_matches = True sum_matches = True
if overwrite == 'always': if overwrite == 'always':
# only use valid object acls for the upload_s3file function
module.params['permission'] = object_acl
upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers) upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers)
else: else:
get_download_url(module, s3, bucket, obj, expiry, changed=False) get_download_url(module, s3, bucket, obj, expiry, changed=False)
else: else:
sum_matches = False sum_matches = False
if overwrite in ('always', 'different'): if overwrite in ('always', 'different'):
# only use valid object acls for the upload_s3file function
module.params['permission'] = object_acl
upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers) upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers)
else: else:
module.exit_json(msg="WARNING: Checksums do not match. Use overwrite parameter to force upload.") module.exit_json(msg="WARNING: Checksums do not match. Use overwrite parameter to force upload.")
# If neither exist (based on bucket existence), we can create both. # If neither exist (based on bucket existence), we can create both.
if pathrtn and not bucketrtn: if not bucketrtn:
# only use valid bucket acls for create_bucket function
module.params['permission'] = bucket_acl
create_bucket(module, s3, bucket, location) create_bucket(module, s3, bucket, location)
# only use valid object acls for the upload_s3file function
module.params['permission'] = object_acl
upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers) upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers)
# If bucket exists but key doesn't, just upload. # If bucket exists but key doesn't, just upload.
if bucketrtn and pathrtn and not keyrtn: if bucketrtn and not keyrtn:
# only use valid object acls for the upload_s3file function
module.params['permission'] = object_acl
upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers) upload_s3file(module, s3, bucket, obj, src, expiry, metadata, encrypt, headers)
# Delete an object from a bucket, not the entire bucket # Delete an object from a bucket, not the entire bucket
@ -684,39 +703,44 @@ def main():
if obj is None: if obj is None:
module.fail_json(msg="object parameter is required") module.fail_json(msg="object parameter is required")
if bucket: if bucket:
deletertn = delete_key(module, s3, bucket, obj, validate=validate) deletertn = delete_key(module, s3, bucket, obj)
if deletertn is True: if deletertn is True:
module.exit_json(msg="Object %s deleted from bucket %s." % (obj, bucket), changed=True) module.exit_json(msg="Object deleted from bucket %s." % bucket, changed=True)
else: else:
module.fail_json(msg="Bucket parameter is required.") module.fail_json(msg="Bucket parameter is required.")
# Delete an entire bucket, including all objects in the bucket # Delete an entire bucket, including all objects in the bucket
if mode == 'delete': if mode == 'delete':
if bucket: if bucket:
deletertn = delete_bucket(module, s3, bucket) deletertn = delete_bucket(module, s3, bucket)
message = "Bucket {0} and all keys have been deleted.".format(bucket) if deletertn is True:
module.exit_json(msg=message, changed=deletertn) module.exit_json(msg="Bucket %s and all keys have been deleted." % bucket, changed=True)
else: else:
module.fail_json(msg="Bucket parameter is required.") module.fail_json(msg="Bucket parameter is required.")
# Support for listing a set of keys # Support for listing a set of keys
if mode == 'list': if mode == 'list':
bucket_object = get_bucket(module, s3, bucket) exists = bucket_check(module, s3, bucket)
# If the bucket does not exist then bail out # If the bucket does not exist then bail out
if bucket_object is None: if not exists:
module.fail_json(msg="Target bucket (%s) cannot be found"% bucket) module.fail_json(msg="Target bucket (%s) cannot be found" % bucket)
list_keys(module, bucket_object, prefix, marker, max_keys) list_keys(module, s3, bucket, prefix, marker, max_keys)
# Need to research how to create directories without "populating" a key, so this should just do bucket creation for now. # Need to research how to create directories without "populating" a key, so this should just do bucket creation for now.
# WE SHOULD ENABLE SOME WAY OF CREATING AN EMPTY KEY TO CREATE "DIRECTORY" STRUCTURE, AWS CONSOLE DOES THIS. # WE SHOULD ENABLE SOME WAY OF CREATING AN EMPTY KEY TO CREATE "DIRECTORY" STRUCTURE, AWS CONSOLE DOES THIS.
if mode == 'create': if mode == 'create':
# if both creating a bucket and putting an object in it, acls for the bucket and/or the object may be specified
# these were separated above into the variables bucket_acl and object_acl
if bucket and not obj: if bucket and not obj:
if bucketrtn: if bucketrtn:
module.exit_json(msg="Bucket already exists.", changed=False) module.exit_json(msg="Bucket already exists.", changed=False)
else: else:
# only use valid bucket acls when creating the bucket
module.params['permission'] = bucket_acl
module.exit_json(msg="Bucket created successfully", changed=create_bucket(module, s3, bucket, location)) module.exit_json(msg="Bucket created successfully", changed=create_bucket(module, s3, bucket, location))
if bucket and obj: if bucket and obj:
if obj.endswith('/'): if obj.endswith('/'):
@ -724,13 +748,18 @@ def main():
else: else:
dirobj = obj + "/" dirobj = obj + "/"
if bucketrtn: if bucketrtn:
keyrtn = key_check(module, s3, bucket, dirobj) if key_check(module, s3, bucket, dirobj):
if keyrtn is True: module.exit_json(msg="Bucket %s and key %s already exists." % (bucket, obj), changed=False)
module.exit_json(msg="Bucket %s and key %s already exists."% (bucket, obj), changed=False)
else: else:
# setting valid object acls for the create_dirkey function
module.params['permission'] = object_acl
create_dirkey(module, s3, bucket, dirobj) create_dirkey(module, s3, bucket, dirobj)
else: else:
# only use valid bucket acls for the create_bucket function
module.params['permission'] = bucket_acl
created = create_bucket(module, s3, bucket, location) created = create_bucket(module, s3, bucket, location)
# only use valid object acls for the create_dirkey function
module.params['permission'] = object_acl
create_dirkey(module, s3, bucket, dirobj) create_dirkey(module, s3, bucket, dirobj)
# Support for grabbing the time-expired URL for an object in S3/Walrus. # Support for grabbing the time-expired URL for an object in S3/Walrus.
@ -738,9 +767,9 @@ def main():
if not bucket and not obj: if not bucket and not obj:
module.fail_json(msg="Bucket and Object parameters must be set") module.fail_json(msg="Bucket and Object parameters must be set")
keyrtn = key_check(module, s3, bucket, obj, validate=validate) keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate)
if keyrtn: if keyrtn:
get_download_url(module, s3, bucket, obj, expiry, validate=validate) get_download_url(module, s3, bucket, obj, expiry)
else: else:
module.fail_json(msg="Key %s does not exist." % obj) module.fail_json(msg="Key %s does not exist." % obj)
@ -748,7 +777,7 @@ def main():
if bucket and obj: if bucket and obj:
keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate) keyrtn = key_check(module, s3, bucket, obj, version=version, validate=validate)
if keyrtn: if keyrtn:
download_s3str(module, s3, bucket, obj, version=version, validate=validate) download_s3str(module, s3, bucket, obj, version=version)
elif version is not None: elif version is not None:
module.fail_json(msg="Key %s with version id %s does not exist." % (obj, version)) module.fail_json(msg="Key %s with version id %s does not exist." % (obj, version))
else: else:
@ -757,55 +786,5 @@ def main():
module.exit_json(failed=False) module.exit_json(failed=False)
def get_s3_connection(aws_connect_kwargs, location, rgw, s3_url):
if s3_url and rgw:
rgw = urlparse(s3_url)
# ensure none of the named arguments we will pass to boto.connect_s3
# are already present in aws_connect_kwargs
for kw in ['is_secure', 'host', 'port', 'calling_format']:
try:
del aws_connect_kwargs[kw]
except KeyError:
pass
s3 = boto.connect_s3(
is_secure=rgw.scheme == 'https',
host=rgw.hostname,
port=rgw.port,
calling_format=OrdinaryCallingFormat(),
**aws_connect_kwargs
)
elif is_fakes3(s3_url):
fakes3 = urlparse(s3_url)
# ensure none of the named arguments we will pass to S3Connection
# are already present in aws_connect_kwargs
for kw in ['is_secure', 'host', 'port', 'calling_format']:
try:
del aws_connect_kwargs[kw]
except KeyError:
pass
s3 = S3Connection(
is_secure=fakes3.scheme == 'fakes3s',
host=fakes3.hostname,
port=fakes3.port,
calling_format=OrdinaryCallingFormat(),
**aws_connect_kwargs
)
elif is_walrus(s3_url):
walrus = urlparse(s3_url).hostname
s3 = boto.connect_walrus(walrus, **aws_connect_kwargs)
else:
aws_connect_kwargs['is_secure'] = True
try:
s3 = connect_to_aws(boto.s3, location, **aws_connect_kwargs)
except AnsibleAWSError:
# use this as fallback because connect_to_region seems to fail in boto + non 'classic' aws accounts in some cases
s3 = boto.connect_s3(**aws_connect_kwargs)
return s3
# import module snippets
from ansible.module_utils.basic import *
from ansible.module_utils.ec2 import *
if __name__ == '__main__': if __name__ == '__main__':
main() main()

View file

@ -1,4 +1,3 @@
boto
boto3 boto3
placebo placebo
cryptography cryptography

View file

@ -55,7 +55,6 @@ lib/ansible/modules/cloud/amazon/rds_param_group.py
lib/ansible/modules/cloud/amazon/rds_subnet_group.py lib/ansible/modules/cloud/amazon/rds_subnet_group.py
lib/ansible/modules/cloud/amazon/redshift.py lib/ansible/modules/cloud/amazon/redshift.py
lib/ansible/modules/cloud/amazon/route53_health_check.py lib/ansible/modules/cloud/amazon/route53_health_check.py
lib/ansible/modules/cloud/amazon/s3.py
lib/ansible/modules/cloud/amazon/s3_lifecycle.py lib/ansible/modules/cloud/amazon/s3_lifecycle.py
lib/ansible/modules/cloud/amazon/s3_logging.py lib/ansible/modules/cloud/amazon/s3_logging.py
lib/ansible/modules/cloud/amazon/s3_website.py lib/ansible/modules/cloud/amazon/s3_website.py

View file

@ -1,10 +1,11 @@
import pytest import pytest
import unittest import unittest
import ansible.modules.cloud.amazon.s3 as s3 import ansible.modules.cloud.amazon.s3 as s3
from ansible.module_utils.six.moves.urllib.parse import urlparse from ansible.module_utils.six.moves.urllib.parse import urlparse
boto = pytest.importorskip("boto") boto3 = pytest.importorskip("boto3")
class TestUrlparse(unittest.TestCase): class TestUrlparse(unittest.TestCase):
@ -32,5 +33,5 @@ class TestUrlparse(unittest.TestCase):
location = None location = None
rgw = True rgw = True
s3_url = "http://bla.blubb" s3_url = "http://bla.blubb"
actual = s3.get_s3_connection(aws_connect_kwargs, location, rgw, s3_url) actual = s3.get_s3_connection(None, aws_connect_kwargs, location, rgw, s3_url)
self.assertEqual("bla.blubb", actual.host) self.assertEqual(bool("bla.blubb" in str(actual._endpoint)), True)