Amazon Web Services Blog

A large part of a successful move is finding a good neighborhood. MSN has a good article that helps you figure this out, and one of the best tips that I've used myself is this one:

Visit a neighborhood on your own at different times of day and night. Talk to neighbors. Visit nearby schools and local businesses. Subscribe to the local paper. Small local papers can be chock-full of information that gives you a feel for the neighborhood or community. If you depend on public transportation, find out how accessible it is in this area. Drive to and from the house from several different directions, so you see both the scenic and not-so-scenic routes.

What are your best tips for finding a good neighborhood? Leave your thoughts in the comments. — Wendy Boswell

I’ve recently added a new section to my site with Automotive Parts, but I think that Automotive is turning out to be one of the more difficult products to sell. E-commerce is never easy, but finding the right formula to sell auto parts is getting to be difficult. A critique would be greatly appreciated.

I think that a junk-yard could sell parts online easier they what I’ve got going on, and that’s just weak. I’ve got everything categorized, into different sections, like oils and fluids, or windshield wipers, which everyone needs. At the very least I do have Google Adsense running on the pages, so typically if people don’t find what they

are looking for, they might even click through to a competitor’s ad. Gotta love Adsense. In any case, I’m going to keep on working on it. Maybe I can re-arrange the look, or limit the categories. What gets me though is that there are even brand name parts that most car mechanics would know and trust. I guess most of them still buy from places like Napa and Discount Auto. That’ll soon change as more mechanics start to place orders online, and next day delivery gets less expensive.

Hello,  I'm having some difficulty signing up and hope someone can help.   I created my AWS account and siged up for S3 (following the "Getting Started" guide).  When I try to sign up for EC2, I end up at a completely blank page.  I'm using FireFox.  I tried with IE6 also and instead of the blank page, I got a 500: Internal server error page.

Specific steps:  From the EC2 page, I clicked on "Sign up for Web Service", which takes me to a sign-in page.  After signing in, I end up with the balnk page (or the 500 error using IE).

Thank you for your help!

I'm very new to the webservices and i tryed around on AWS Zone with the ItemSearch "Home Scratch Pads ECS - DE ItemSearch"


My question is:
Why can i get some results if I use "SOAP" -> "View Response". Otherwise with the generated c# code, with the same parameters left in the form, i get no results in my application.

There is the message: "System.Exception: We did not find any matches for your request."

Any idea what the difference between the both are? Maybe do I use another service from my "local" Application? I use: http://webservices.amazon.com/AWSECommerceService/2005-03-23/AWSECommerceService.wsdl as noted in the generated source code.

Thanks for any hint!
Patrik

PS: If I do a SOAP Request through AWS Zone, there is a parameter in the url = awszone_baseUrl=http://ecs.amazonaws.de/onca/xml . This will send the request - I think - to the german server. How can ich set this server (url/parameter) if I do a request in my VisualStudio c# project? (Because i think, this is the reason of the two different results)

Most clients today send  the following header when making a request:

Accept-Encoding: gzip,deflate

This means that the client will accept content that is compressed and will decompress it (on the client side).

I have uploaded a compressed html page (gzip compression) and added the header " Content-Encoding: gzip " in the response. When I request this page with a browser, S3 sends the compressed page with the header indicating that the response is compressded. As expected it works. The page is successully rendered.

So far, so good. Then I wondered what would happen if the client did not support compressed content? What if the client did not sent " Accept-Encoding: gzip,deflate "?

I did the test and S3 sends exactly the same compressed response. Therefore the client could not render the page.

I would have hoped that S3 would have noticed that the response is gzip and would have decompressed it (and removed the Content-Encoding header).

Dear Amazon, could you imprement this feature please?
It would help me a lot. You already do https so decompressing a few responses should not be too much for you I hope.

Hi, I'm developing a system with C# / .NET forms which allows backup of files to S3.  But it also sends usage information to a second server using .NET remoting via a HttpChannel. 

Now access to the usage-server via .NET remoting works fine until I call any S3 method via the web-service.  Any sub-sequent calls to the usage server using remoting fails with a:

Exception has been thrown by the target of an invocation
Specified cast is not valid

If I call S3 in isolation it's fine, if I call my usage-server via remoting it's fine.  It just seems like one is somehow affecting the other. 

Has anyone else experienced anything similar?  It has got me slightly flumoxed.

Paul

I just wanted to share some SQS pricing feedback with you guys.  We'd really like to use SQS to provide transactions around S3 updates that touch multiple resources, but are concerned that the $0.10/1k messages price is too high (for our market, anyways).

If we had 50,000 users making 10 transactions a day, that turns out to be over $18k/year in just the SQS transaction costs, not even including bandwidth fees. For some applications I'm sure that's okay, but it's definitely a problem for us. Consequently, we've elected to not implement transactions and instead handle problems at the application level as they come up.  Which is not fun, as you might imagine!  If the price was $0.10 for 100k messages we would be a lot happier. :-)

The AIM documentation gives a url and format and method for submitting batch refunds:
https://secure.amazon.com/exec/panama/seller-admin/catalog-upload/batch-refund

Great. Then they give you a tool to check the status of bulk refunds:

https://secure.amazon.com/exec/panama/seller-admin/manual-reports/get-report-status
ReportName:BatchRefund|NumberOfReports:10

Here's where it gets dumb.

When you upload a batch refund successfully, all you get back is "SUCCESS".

When you query for the status of bulk refunds, you get:

<Reports>
<Report>reportstarttime=10-18-2004:10-50-56 reportendtime=10-18-2004:10-50-56 reportid=12345456 </Report>
<Report>reportstarttime=10-11-2004:15-02-21 reportendtime=10-11-2004:15-02-21 reportid=23664721 </Report>
</Reports>

Nevermind the fact that it's difficult to match your upload to a reportid (could've it have been returned as part of a succesful upload?) - but what do you do with this information?

What's 'the status of my refund batch? How do I find out?

Hello,

Unfortunaly Amazon is only providing an RPM. You have to cross a small boisé to get the things work. Here is my notes:

First you need rsync, openssl, ruby, libopenssl-ruby and rpm (all of these are apt-gettable).

Then here is how I've extracted the files from the can:

 $ rpm2cpio ec2-ami-tools.noarch.rpm > /tmp/ami.cpio
 $ mkdir -p usr/local/aes
 $ mkdir -p usr/lib/site_ruby/aes/amiutil
 $ mkdir -p etc/aes/amiutil

Then spread it into your system:

 $ cd /
 $ cpio -i < /tmp/ami.cpio

Debianize the lib path:

 $ ln -s /usr/lib/site_ruby/aes/ /usr/lib/ruby/1.8/

Fix the image.rb, see:

http://developer.amazonwebservices.com/connect/thread.jspa?messageID=42883

Now you should be able to use the ec2-* bundling tool.

Extra stuff! We were bundling from a vserver ( http://www.linux-vserver.org), that require special capability (mknod and mount).

You need CAP_MKNOD in your bcapabilities files (you have to restart your vservers). SECURE_MOUNT is also required, once your vserver is started, run:

 $ vattribute --set --xid 185 --ccap SECURE_MOUNT

If you don't have the loop device, you'll have to create it:

 $ mkdir /var/lib/vservers/subversion/dev/loop
 $ mknod /var/lib/vservers/subversion/dev/loop/0 b 7 0

Did not try the image yet! So, I'll keep you updated on my progress.

Regards,

--
Samuel T. Cossette
Poly9 Group, Inc.
http://www.poly9.com/
Message was edited by: samueltc

With either a modified Ruby API or Firefox, I can't seem to get persistent connections.  Here is an example from FF:
GET /ServEdge_pub/s3sync/README.txt HTTP/1.1
Host: s3.amazonaws.com
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.7) Gecko/20060909 Firefox/1.5.0.7
Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip,deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 300
Connection: keep-alive

HTTP/1.1 200 OK
x-amz-id-2: tk/vY1eJAR3rI7jsEUiVIQsDlPYNJ9QFC7bYN2Yt/D6JObnqzG8UwxphJThsFsTE
x-amz-request-id: 21F99A6524B072F2
Date: Sat, 30 Sep 2006 01:45:59 GMT
x-amz-meta-group: 0
x-amz-meta-owner: 0
x-amz-meta-permissions: 33188
Last-Modified: Fri, 29 Sep 2006 06:08:44 GMT
ETag: "036067dab59825484743b37148382c93"
Content-Type: text/plain
Content-Length: 7324
Connection: close
Server: AmazonS3


Am I just missing something simple?
Thanks!