Amazon Web Services为什么无法解析我的CloudFront失效批处理请求XML文档？

Question

I'm trying to invalidate some objects I uploaded to Amazon CloudFront using their invalidation API.

As per their documentation, I'm sending a POST request with an XML document as its content. The XML document specifies the paths to invalidate.

The error I'm getting back from Amazon is:

<Error>
    <Type>Sender</Type>
    <Code>MalformedInput</Code>
    <Message>Could not parse XML</Message>
</Error>

As far as I can see, my XML document matches their documentation.

Their documentation:

• http://docs.amazonwebservices.com/AmazonCloudFront/latest/APIReference/InvalidationBatchDatatype.html

My XML document:

<InvalidationBatch xmlns="http://cloudfront.amazonaws.com/doc/2010-11-01/">
    <Path>/-o-replace.css</Path>
    <Path>/-o-set-link-source.css</Path>

    ...16 other path elements, each containing only letters, numbers, hyphens and periods

    <CallerReference>fixing-accidental-setting-of-gzip-header</CallerReference>
</InvalidationBatch>

I've tried including the XML prolog as they do in their response, keeping and removing whitespace, and leaving out the xmlns attribute, all to no effect.

I'm sending the POST request manually using Python. Here's the Python code used to send it. I've confirmed that the file contents are getting read in correctly.

from httplib import HTTPSConnection
from datetime import datetime
from hashlib import sha1
import hmac

conneck = HTTPSConnection('cloudfront.amazonaws.com')

invalidation_file = file('invalidation.xml')
invalidation = unicode(invalidation_file.read()).encode('utf-8')

now_as_string = datetime.now().strftime('%a, %d %b %Y %H:%M:%S GMT')
signature = hmac.new('MY_SECRET_KEY', now_as_string, sha1).digest().encode('base64')

conneck.request('POST', '/2010-11-01/distribution/MY_DISTRIBUTION_ID/invalidation', invalidation, {
     'Content-Type': 'text/xml',
    'Authorization': 'AWS MY_ACCESS_KEY_ID' + ':' + signature,
       'x-amz-date': now_as_string,
})

response = conneck.getresponse()

Can anyone see what I'm doing wrong?

Answer 1

Not an expert with CF APIs, but I think you are doing the base64 encoding wrong, see example:

>>> 'xyz'.encode('base64')
'eHl6\n'
>>> base64.b64encode('xyz')
'eHl6'