From jan.bromberger at gmx.de  Wed May 28 10:37:40 2008
From: jan.bromberger at gmx.de (Jan Bromberger)
Date: Wed, 28 May 2008 16:37:40 +0200
Subject: [s3-dev] Cannot access EU buckets,
	although specifying given endpoint as :server
Message-ID: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>

Hi everybody,

while I can access US buckets fine, I fail to access Europe buckets.

After I found out that I have to use subdomain instead of path to  
specify my bucket, I changed the :server parameter of  
establish_connection! accordingly, but still get the same error  
message response:

Can anybody tell what I'm doing wrong?

Cheers from Munich
Jan


AWS::S3::Base.establish_connection!(:server =>  
'platinnetzd.s3.amazonaws.com', :access_key_id =>  
'AAAAAAAA', :secret_access_key => 'BBBBBBBBBBBBB')
=> #<AWS::S3::Connection:0x1108140 @http=#<Net::HTTP  
platinnetzd.s3.amazonaws.com:80 open=false>,  
@options 
= 
{:access_key_id 
= 
 > 
"AAAAAAAA 
", :port 
= 
 > 
80 
, :secret_access_key 
= 
 > 
"BBBBBBBBBBBBB 
", :persistent=>true, :server=>"platinnetzd.s3.amazonaws.com"},  
@secret_access_key= BBBBBBBBBBBBB", @access_key_id="AAAAAAAA">

AWS::S3::Bucket.find('platinnetzd')
=> AWS::S3::PermanentRedirect: The bucket you are attempting to access  
must be addressed using the specified endpoint. Please send all future  
requests to this endpoint.

<?xml version="1.0" encoding="UTF-8"?>
<Error>
  <Code>PermanentRedirect</Code>
  <Message>The bucket you are attempting to access must be addressed  
using the specified endpoint. Please send all future requests to this  
endpoint.</Message>
  <RequestId>728FC98925062009</RequestId>
  <Endpoint>platinnetzd.s3.amazonaws.com</Endpoint>
  <HostId>iTBibbLybjDmq2bXCGZpZ9FjQjUq4/0GQ1A/axf 
+AbaVXsZ2b6n6xLbUq1r5cBhi</HostId>
  <Bucket>platinnetzd</Bucket>
</Error>

From hosiawak at gmail.com  Wed May 28 10:54:58 2008
From: hosiawak at gmail.com (Karol Hosiawa)
Date: Wed, 28 May 2008 15:54:58 +0100
Subject: [s3-dev] Cannot access EU buckets,
 although specifying given endpoint as :server
In-Reply-To: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
References: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
Message-ID: <20080528155458.6a114a78.hosiawak@gmail.com>

On Wed, 28 May 2008 16:37:40 +0200
Jan Bromberger <jan.bromberger at gmx.de> wrote:

> Hi everybody,
> 
> while I can access US buckets fine, I fail to access Europe buckets.
> 
> After I found out that I have to use subdomain instead of path to  
> specify my bucket, I changed the :server parameter of  
> establish_connection! accordingly, but still get the same error  
> message response:
> 
> Can anybody tell what I'm doing wrong?
> 
> Cheers from Munich
> Jan

Currently the aws/s3 Ruby library doesn't handle European buckets at
all. You can either use US buckets or write your own library.

More about aws/s3 problems I encountered and how I solved them:

http://www.blik.it/2008/5/15/aws-s3-eoferror-sysread-error-and-ruby-threads

Cheers from Sheffield, UK
-- 
Karol Hosiawa <hosiawak at gmail.com>

From jan.bromberger at gmx.de  Wed May 28 13:39:14 2008
From: jan.bromberger at gmx.de (Jan Bromberger)
Date: Wed, 28 May 2008 19:39:14 +0200
Subject: [s3-dev] Cannot access EU buckets,
	although specifying given endpoint as :server
In-Reply-To: <20080528155458.6a114a78.hosiawak@gmail.com>
References: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
	<20080528155458.6a114a78.hosiawak@gmail.com>
Message-ID: <CACA72F6-F763-46A5-99EC-4A3A72E23605@gmx.de>

On 28. 05.2008, at 16:54, Karol Hosiawa wrote:

> On Wed, 28 May 2008 16:37:40 +0200
> Jan Bromberger <jan.bromberger at gmx.de> wrote:
>
>> Hi everybody,
>>
>> while I can access US buckets fine, I fail to access Europe buckets.
>>
>> After I found out that I have to use subdomain instead of path to
>> specify my bucket, I changed the :server parameter of
>> establish_connection! accordingly, but still get the same error
>> message response:
>>
>> Can anybody tell what I'm doing wrong?
>>
>> Cheers from Munich
>> Jan
>
> Currently the aws/s3 Ruby library doesn't handle European buckets at
> all. You can either use US buckets or write your own library.
>
> More about aws/s3 problems I encountered and how I solved them:
>
> http://www.blik.it/2008/5/15/aws-s3-eoferror-sysread-error-and-ruby-threads
>
> Cheers from Sheffield, UK

There's not much to be found on the web on using European buckets with  
ruby.
I thought it was as easy as specifying the right server, so the plugin  
doesn't have to know how to handle it at all.
But I guess that's not the case.

Thanks for your response!

From mdchaney at michaelchaney.com  Wed May 28 15:03:08 2008
From: mdchaney at michaelchaney.com (Michael Chaney)
Date: Wed, 28 May 2008 14:03:08 -0500
Subject: [s3-dev] Cannot access EU buckets,
	although specifying given endpoint as :server
In-Reply-To: <CACA72F6-F763-46A5-99EC-4A3A72E23605@gmx.de>
References: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
	<20080528155458.6a114a78.hosiawak@gmail.com>
	<CACA72F6-F763-46A5-99EC-4A3A72E23605@gmx.de>
Message-ID: <3b456edd0805281203h2723279avc443c834a1c7f301@mail.gmail.com>

> There's not much to be found on the web on using European buckets with ruby.
> I thought it was as easy as specifying the right server, so the plugin
> doesn't have to know how to handle it at all.
> But I guess that's not the case.

What is the difference, besides the server name?

Michael
-- 
Michael Darrin Chaney, Sr.
mdchaney at michaelchaney.com
http://www.michaelchaney.com/

From jan.bromberger at gmx.de  Wed May 28 16:56:23 2008
From: jan.bromberger at gmx.de (Jan Bromberger)
Date: Wed, 28 May 2008 22:56:23 +0200
Subject: [s3-dev] Cannot access EU buckets,
	although specifying given endpoint as :server
In-Reply-To: <3b456edd0805281203h2723279avc443c834a1c7f301@mail.gmail.com>
References: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
	<20080528155458.6a114a78.hosiawak@gmail.com>
	<CACA72F6-F763-46A5-99EC-4A3A72E23605@gmx.de>
	<3b456edd0805281203h2723279avc443c834a1c7f301@mail.gmail.com>
Message-ID: <6D6AA41A-6CE6-453A-AF7C-0A61AC9F9319@gmx.de>

On 28. 05.2008, at 21:03, Michael Chaney wrote:

>> There's not much to be found on the web on using European buckets  
>> with ruby.
>> I thought it was as easy as specifying the right server, so the  
>> plugin
>> doesn't have to know how to handle it at all.
>> But I guess that's not the case.
>
> What is the difference, besides the server name?

Well, I still don't know. But I set the server name to the "endpoint"  
stated in the PermanentRedirect and it still doesn't work.

So there must be something else.

-- Jan

From jan.bromberger at gmx.de  Thu May 29 08:39:49 2008
From: jan.bromberger at gmx.de (Jan Bromberger)
Date: Thu, 29 May 2008 14:39:49 +0200
Subject: [s3-dev] Cannot access EU buckets,
	although specifying given endpoint as :server
In-Reply-To: <6D6AA41A-6CE6-453A-AF7C-0A61AC9F9319@gmx.de>
References: <16E6806D-D059-42F3-A04B-FE26A9BF8CDC@gmx.de>
	<20080528155458.6a114a78.hosiawak@gmail.com>
	<CACA72F6-F763-46A5-99EC-4A3A72E23605@gmx.de>
	<3b456edd0805281203h2723279avc443c834a1c7f301@mail.gmail.com>
	<6D6AA41A-6CE6-453A-AF7C-0A61AC9F9319@gmx.de>
Message-ID: <1C264BE5-DA9F-48D3-B83F-93EA9FA3C5E3@gmx.de>

On 28. 05.2008, at 22:56, Jan Bromberger wrote:

> On 28. 05.2008, at 21:03, Michael Chaney wrote:
>
>>> There's not much to be found on the web on using European buckets  
>>> with ruby.
>>> I thought it was as easy as specifying the right server, so the  
>>> plugin
>>> doesn't have to know how to handle it at all.
>>> But I guess that's not the case.
>>
>> What is the difference, besides the server name?
>
> Well, I still don't know. But I set the server name to the  
> "endpoint" stated in the PermanentRedirect and it still doesn't work.
>
> So there must be something else.
>
> -- Jan

It looks like even when you connect to the correct server, the gem  
still sends relative paths.
After I took out the bucket prefixing of paths, it started working.  
maybe something like this should be done:

# bucket.rb, lines 177ff
           def path(name, options = {})
             if name.is_a?(Hash)
               options = name
               name    = nil
             end
             "/#{bucket_name(name) unless bucket_name(name) ==  
connection.subdomain}#{RequestOptions.process(options).to_query_string}"
           end

But I have only started to dig into this, so this is only a first idea.

I wonder... why isn't there just a server like s3eu.amazonaws.com that  
takes pathbs like s3.amazonaws.com does for US buckets?
Or alternatively they could just live with a path if it is to the same  
bucket or a bucket in the same realm... this really sucks. ;-)

Cheers
Jan

From jan.bromberger at gmx.de  Thu May 29 09:07:27 2008
From: jan.bromberger at gmx.de (Jan Bromberger)
Date: Thu, 29 May 2008 15:07:27 +0200
Subject: [s3-dev] Connection.subdomain wrong pattern
Message-ID: <DA8CBE92-C186-4397-AB3F-6F9D2CFD7D70@gmx.de>

By the way, the subdomain pattern is wrong, as the period '.'  
character is actually allowed in bucket names.
See: http://docs.amazonwebservices.com/AmazonS3/2006-03-01/

# connection.rb, line 69
       def subdomain
         http.address[/^([^.]+).#{DEFAULT_HOST}$/, 1]
       end

should be:

         http.address[/^(.+)\.#{DEFAULT_HOST}$/, 1]

or something fancy like

         http.address[/^([a-z0-9][a-z0-9._-]{0;61}[a-z0-9]) 
\.#{DEFAULT_HOST}$/, 1]

(but what's the point in restricting this when the connect has  
succeeded...)

Also notice the \. instead of the . before DEFAULT_HOST.

Cheers
Jan


P.S.: I would like to report this, but RubyForge gives me a nice and  
warm

Error
Permission: User Not Found