问题
I'm trying to upload my static website to my s3 bucket. I've managed to accomplish this. I've created my_bucket and then create a sub_bucket named test_folder and in that I uploaded all my css, html, js files.
It would look like this now:
my_bucket/
test_folder
index.html
And I was able to view my index.html, horray! :D. But my question is in setting up the index document, since the index.html is located to a sub_bucket: test_folder/index.html when I try to save it, gave me The
IndexDocument Suffix is not well formed
Is it possible to link the index.html in a sub_bucket? If yes, how? If not, is there an alternative way to achieve this? I have here a screenshot link
回答1:
Though often used for different purposes, the index document was originally intended, conceptually, to be the "index" (directory listing or other content summary) of all of the files within each folder, so this configuration parameter specifies the index document to return for each folder in the entire bucket, if such a document exists within the folder... this isn't a single configuration "thing" for the bucket as a whole.
If your attempted configuration had been accepted by S3, it would have had the following impact, assuming a bucket name of "example.com":
browser address bar file (object) actually returned
--------------------------- ------------------------------------
http://example.com example.com/test/index.html
http://example.com/help example.com/help/test/index.html
http://example.com/foo/test example.com/foo/test/test/index.html
It seems very unlikely that this is what you actually intended.
However, but that is how index documents work... they are conceptually intended to be related to the other things at each level of the directory hierarchy, which of course could be the actual listing of files, or could be an "index" in a much more broad and vague and general sense of any "page," such as a landing page you want a visitor to see when they go to a specific "directory" on your site, which of course, in the modern web is not typically conceptualized as a "directory" but rather simply as a "page."
So the index document has to be immediately under the same / delimiter and can't contain an additional / within its own specification.
The index document for example.com has to be stored in example.com/index.html (assuming "index.html" is your chosen index filename) -- it has to be stored within the "directory" that it indexes, just like on a conventional web server, where, in some configurations, the web server will actually display a directory listing of files, with the "index" page replacing that directory listing in cases where the "index" page actually exists. Of course, S3 doesn't have default directory listing page functionality.
http://docs.aws.amazon.com/AmazonS3/latest/dev/IndexDocumentSupport.html
In contrast to the index document, the error document, if you configure it, is a global configuration that is used regardless of where, within the bucket, the 404 occurs, so slashes are supported in that entry. The AWS console prompts are light on hints as to the nature of the two entries, which are so different in their behavior that they arguably should be more separated, visually.
You'll note that "sub-bucket" isn't an actual term, for what you are describing, which is an object with delimiters in its key (path), which gives the appearance of being nested under a directory or folder.
For clarity, I have used the words "folder" and "directory" very casually all throughout this answer, with the conventional meaning... but for technical accuracy, I'll mention that S3 objects are not really stored internally in a hierarchical fashion "in directories." It appears this way, and for practical purposes, it works that way; however, it's actually the case that the / character, while close to being just another character in the object key, although it gets some special treatment as a delimiter because of its conventional use as a directory delimiter.Unlike some more conventional filesystems, the number of "files in each directory" does not pose any performance concern with S3 and doesn't need to be managed in the same way as is needed in a conventional filesystem when a large number of files exists, since S3 internally hashes the key ("path") of each object for its internal storage partitioning logic.
回答2:
I also had this issue. Based on Michael's answer, I developed this work-around:
In the Amazon S3 bucket properties, set the Index Document: value to something arbitrary (I used 'index.htm' and didn't have a index.htm file present in the 'root' of the bucket. Then set the Error Document: value to wherever you want the user sent to (test_folder/index.html in your case).
Remember, any 404 error message on your site will be sent to the specified page. I know this is a horrible work around, but it seems to work for me. Any other suggestions gratefully accepted!
回答3:
I have been trying to do the same thing (serve static from s3) and reading this post made me realize that you need to place the index file (e.g. 'index.html') in the bucket outside of any folders.
I had my index document in a templates folder. I had read the documentation but I was so frustrated that I couldn't understand what it was really saying. Sometimes, the really stupid mistakes are the most difficult to troubleshoot. Hope this helps somebody.
回答4:
try to fill index.html in the "Index Document" cell.
回答5:
The alternate way is If you are using(or use) cloudfront with S3, you can have
DefaultRootObject: test_folder/index.html
This way your cloudfront url https://yourdomain/ will return test_folder/index.html
来源：https://stackoverflow.com/questions/20264928/s3-static-website-index-document