requests not being signed after switching to a cdn backed by an s3 bucket

[dacrowlah]

I had this working just fine when I used my-bucket-name.s3.amazonaws.com/ as the default storage for my site. Last night I set up CloudFlare to proxy the requests through. In doing this you do the following:

• create a bucket name that matches the CNAME you want to serve through (cdn.mysite.com) so that the url would be cdn.mysite.com.s3.amazonaws......
• in the CloudFlare DNS panel, you setup a CNAME for cdn.mysite.com to redirect to cdn.mysite.com.s3.amazonaws.com
• presto, magic works.

however, when it was using my-bucket-name.s3, it django was automatically signing all of the requests for the content, which I want it to do... but when I switched to the cdn, it stopped, and I had to open the bucket up to public viewing, which I do not want.

[dacrowlah]

FYI: I'm using Python 3.4, django 1.7.1

[jschneier]

Hi @dacrowlah and thanks for the report! What version of 3.4 are you using and do you have a traceback? I'm wondering if this is related to #13 which references an upstream boto bug. Depending on the traceback switching your CallingFormat might fix this issue. More information would be great!

Also your boto version would be helpful as well.

[dacrowlah]

I haven't seen any tracebacks/exceptions. Pardon, but this is my first
django app, and I haven't seen a log file anywhere (tho I imagine its there
somewhere). I'm running in debug mode, and dont see anything.

On Thu, Jan 22, 2015 at 11:12 AM, Josh Schneier notifications@github.com
wrote:

Hi @dacrowlah https://github.com/dacrowlah and thanks for the report!
What version of 3.4 are you using and do you have a traceback? I'm
wondering if this is related to #13
#13 which
references an upstream boto bug. Depending on the traceback switching your
CallingFormat might fix this issue. More information would be great!

—
Reply to this email directly or view it on GitHub
#18 (comment)
.

[jschneier]

Dupe #456 I did not understand the question at the time.