ChatGPT解决这个技术问题 Extra ChatGPT

AWS CloudFront access denied to S3 bucket

I am trying to setup CloudFront to serve static files hosted in my S3 bucket. I have setup distribution but I get AccessDenied when trying to browse to the CSS (/CSS/stlyle.css) file inside S3 bucket:

<Error>
    <Code>AccessDenied</Code>
    <Message>Access Denied</Message>
    <RequestId>E193C9CDF4319589</RequestId>
    <HostId>
xbU85maj87/jukYihXnADjXoa4j2AMLFx7t08vtWZ9SRVmU1Ijq6ry2RDAh4G1IGPIeZG9IbFZg=
    </HostId>
</Error>

I have set my CloudFront distribution to my S3 bucket and created new Origin Access Identity policy which was added automatically to the S3 bucket:

{
    "Version": "2008-10-17",
    "Id": "PolicyForCloudFrontPrivateContent",
    "Statement": [
        {
            "Sid": "1",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::cloudfront:user/CloudFront Origin Access Identity E21XQ8NAGWMBQQ"
            },
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::myhost.com.cdn/*"
        }
    ]
}

Did I miss something?

I want all my files in this S3 bucket be served via CloudFront...

*** UPDATE ***

This cloud front guide says:

By default, your Amazon S3 bucket and all of the objects in it are private—only the AWS account that created the bucket has permission to read or write the objects in it. If you want to allow anyone to access the objects in your Amazon S3 bucket using CloudFront URLs, you must grant public read permissions to the objects. (This is one of the most common mistakes when working with CloudFront and Amazon S3. You must explicitly grant privileges to each object in an Amazon S3 bucket.)

So based on this I have added new permissions to all objects inside S3 bucket to Everyone Read/Download. Now I can access files.

But now when I access the file like https://d3u61axijg36on.cloudfront.net/css/style.css this is being redirected to S3 URI and HTTP. How do I disable this?

The redirection might be temporary forums.aws.amazon.com/message.jspa?messageID=677452
I have come across that topic myself. Still waiting to see if this is the case...
Your update is the most confusing thing ever! Not a jab at you - more at amazon. Why must my objects be public if I have declared an OAI, changed the policy for that OAI, and have my bucket as private - as I want it to not be publicly accessible?! Makes no sense to me. Will they be private if the bucket is private and the objects are public?

J
John Rotenstein

To assist with your question, I recreated the situation via:

Created an Amazon S3 bucket with no Bucket Policy

Uploaded public.jpg and make it public via "Make Public"

Uploaded private.jpg and kept it private

Created an Amazon CloudFront web distribution: Origin Domain Name: Selected my S3 bucket from the list Restrict Bucket Access: Yes Origin Access Identity: Create a New Identity Grant Read Permissions on Bucket: Yes, Update Bucket Policy

Origin Domain Name: Selected my S3 bucket from the list

Restrict Bucket Access: Yes

Origin Access Identity: Create a New Identity

Grant Read Permissions on Bucket: Yes, Update Bucket Policy

I checked the bucket, and CloudFront had added a Bucket Policy similar to yours.

The distribution was marked as In Progress for a while. Once it said Enabled, I accessed the files via the xxx.cloudfront.net URL:

xxx.cloudfront.net/public.jpg redirected me to the S3 URL http://bucketname.s3.amazonaws.com/public.jpg. Yes, I could see the file, but it should not use a redirect.

xxx.cloudfront.net/private.jpg redirected me also, but I then received Access Denied because it is a private file in S3.

I then did some research and found that this is quite a common occurrence. Some people use a workaround by pointing their CloudFront distribution to the static hosted website URL, but this has the disadvantage that it will not work with the Origin Access Identity and I also suspect it won't receive the 'free S3 traffic to the edge' discount.

So, I waited overnight, tested it this morning and everything is working fine.

Bottom line: Even if it says ENABLED, things might take several hours (eg overnight) to get themselves right. It will then work as documented.


I have the same experience. Thanks for sharing this with us.
What wasn't obvious to me is that the cloudfront url does NOT include the bucket name. i.e. 'xxxx.cloudfront.net/:path' and NOT 'xxxx.cloudfront.net/bucketname/:path'
If I could have upvoted this twice, I would have. This stuff has cost me hours of my life.
@john-rotenstein I'm not following your answer. If I make objects in my bucket public, I can access them via my CloudFront distribution URL. That's great! But I can also then access them directly via the S3 URL. I implemented an OAI so that only my CloudFront dist can read my objects and no one else can. I don't understand the use of that if my objects are public. Have I misunderstood your answer? Can I provide any more information to clarify my question?
If you see that cloudfront is redirecting you to the s3 asset (check your url after entering the cloudfront domain), then this (waiting a few hours) is very likely the solution.
Z
Zohab Ali

I added index.html in Default Root Object under General tab of cloudFront Distribution Settings and it worked for me. As index.html was the root file for my project!


bless you. I was pulling my hair out on this, and adding the root object fixed it!
This is the thing to check if the S3 bucket endpoint is using API-style. Even though it says "Access Denied", it's not truly a permissions error - it's that you're asking for a resource that doesn't technically exist.
make sure it's index.html NOT /index.html
These are things that keep suggesting me "leave all this stuff and go farming".
Wow, life saver. And make sure to be patient. It took about 30 seconds for things to propagate or whatever needed to happen for it to work (which is why I originally thought setting this "optional" field wasn't a fix), but yes, telling it that the index.html is the Default Root Object fixed it for me.
u
uday reddy

Instead of choosing default s3 bucket for Origin Domain Name, please enter the <bucket-name>.s3-website.<region>.amazonaws.com as origin Domain Name(You can get this URL at Static website hosting property under S3 bucket properties).


S
Scott Jungwirth

In my case I was using multiple origins with "Path Pattern" Behaviors along with an Origin Path in my S3 bucket:

Bad setup:

CloudFront Behavior: /images/* -> My-S3-origin

My-S3-origin: Origin Path: /images

S3 files: /images/my-image.jpg

GET Request: /images/my-image.jpg -> 403

What was happening was the entire CloudFront GET request gets sent to the origin: /image/my-image.jpg prefixed by Origin Path: /images, so the request into S3 looks like /images/images/my-image.jpg which doesn't exist.

Solution

remove Origin Path.


S
Suresh

This can happen if you are using a bucket that has just been newly created.

According official reply here: AWS Forun link, you have to wait for a couple of hours after creating a new bucket before you can have cloud front distribution working on it correctly.

Solution is to temporarily work from one of your old buckets and switch to the new bucket a couple of hours later.


S
Sahar

I also got 403 from CloudFront but my issue was a bit different so sharing it here as it might help others.

Make sure the Origin Access Id you've defined as part of the bucket policy is the right one:

{
    "Version": "2008-10-17",
    "Id": "PolicyForCloudFrontPrivateContent",
    "Statement": [
        {
            "Sid": "1",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::cloudfront:user/CloudFront Origin Access Identity HERE_GOES_YOUR_ORIGIN_ACCESS_ID"
            },
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::deepen-frontend-prod/*"
        }
    ]
}

An Origin Access Identity Id roughly looks like E19F48VV5H01ZD.

https://i.stack.imgur.com/nZE90.png


H
Hosam Elzagh

add this in Bucket policy

    {
        "Sid": "2",
        "Effect": "Allow",
        "Principal": {
            "AWS": "arn:aws:iam::cloudfront:user/CloudFront Origin Access Identity ESDK2T2CSNT57"
        },
        "Action": "s3:GetObject",
        "Resource": "arn:aws:s3:::{Bucket name }/*"
    }

M
Mohit Mutha

Making the objects in the S3 bucket public-read is the fastest way to achieve this. However this is not recommended. Rather create a origin_access_identity under the s3 orgin policy and grant this identity access to the S3 bucket using a bucket policy. That way you can keep all the objects private.


I did this. Still having a 403 for root objects!
K
Kaushal Sachan

See 100% working solution...

It can be only solve at cloudFront panel by adding error response rule.

See Solution here - react router doesn't work in aws s3 bucket


m
mayur chavan

It worked for me once I added index.html to the Default Root Object under the General tab of cloudFront Distribution Settings.


it's a duplicate of other answear.
u
user1267675

As an AWS novice, I encountered a similar problem after following a tutorial to set up a public S3 bucket. I wanted my S3 bucket to be private and have CloudFront serve the files. I set up an OAI and turned on Block all public access, only to receive AccessDenied errors on all my CloudFront URLs.

My problem was that I had created two conflicting statements in my S3 bucket policy: one when I set up the public bucket following the tutorial, and the other that was generated automatically when I set up the OAI and selected Update bucket policy:

{
    "Version": "2012-10-17",
    "Id": "Policy1620442091089",
    "Statement": [
        {
            "Sid": "Stmt1620442087873",
            "Effect": "Allow",
            "Principal": "*",
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::my-s3-bucket/*"
        },
        {
            "Sid": "2",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::cloudfront:user/CloudFront Origin Access Identity XXXXXXXXXXXXXX"
            },
            "Action": "s3:GetObject",
            "Resource": "arn:aws:s3:::my-s3-bucket/*"
        }
    ]
}

Removing the top statement solved my AccessDenied errors. Hopefully this helps someone new to AWS. It wasn't immediately obvious to me to check for conflicting statements.


S
Sandeep Dixit

You can point cloud front to use s3 bucket or you can host your s3 bucket as a static website and point Cloud Front to a Static website URL.

When you are pointing CloudFront to s3 bucket I used OAI as prescribed here. It works for me.

https://i.stack.imgur.com/5Ko2Q.png

https://i.stack.imgur.com/X0D3y.png

https://i.stack.imgur.com/EJ8x1.png

However, please note that you cannot access the root without mentioning a root document.

append /existingfile.html to url and see whether you can access any document by full url. If not you need check your permissions issues mostly at s3 bucket. Follow above URL.

you need to provide a default root document (index.html or something like that).. Otherwise, your root will be denied access.

When you want to use s3 the static website. Do not choose S3 origin from the available list. Rather copy and paste S3 static root. In this case, the Origin has to be the exact URI given in the Static website hosting of the bucket. NOTE that This is different from the list very similar bucket name given in options while creating cloud front distribution.

Before creating a cloud front distribution make sure the website is working via this URL.

https://i.stack.imgur.com/OJbl1.png