Home / server / Questions

Questions[eucalyptus](server)

Martin Hope
João Pereira
Asked: 2012-03-04 11:01:07 +0800 CST
  • 2

I'm beginning with Eucalyptus, with some S3 and EC2 background. I have an app for which I use a personal library to interface with S3 and store/retrieve files from there. Now I want to make my app compatible with Walrus too.

This is my setup:

Ubuntu 10.04 server UEC:

  • Just installed it on a fresh VM, selecting all components but the NC, as I only need Walrus
  • After boot I verified that walrus was running at 10.211.55.18:8773/services/Walrus
  • Go to 10.211.55.18:8443 web portal and changed the admin password
  • From the same place I retrieved the query interface credentials (id+secret)

I was able to interact with Walrus through Cyberduck and Cloudberry without problems, I've created then a bucket 'mybucket' and an object 'myobject', but cant make a plain request with a programming HTTP library, neither using curl with generated authorization.

If I go to the browser and insert: http://10.211.55.18:8773/services/Walrus/mybucket/myfile I get a pretty XML message with an authorization error (as expected).

But when trying to get that object, I've done the following signing process (pseudo code):

digest = new_openssl_sha1_digest

date = "Sat, 03 Mar 2012 18:13:54 GMT" (now UTC datetime)

description = "GET\n\n\nSat, 03 Mar 2012 18:13:54 GMT\n/mybucket/myfile"

signature = new_base_64_HMAC(digest, secret_key, description)

headers = { 
            Date: date,  
            Authorization: "AWS <access_key_id>:<signature>"
          }

So then I end up with these request URL and headers:

Path: http://10.211.55.18:8773/services/Walrus/mybucket/myfile

Headers: {"date"=>"Sat, 03 Mar 2012 18:13:54 GMT", "Authorization"=>"AWS WKy3rMzOWPouVOxK1p3Ar1C2uRBwa2XXXXXX:U/fpoVUCbN64h8iRICp9JZzKWgE="}

But when I issue a GET request as above I only get a tiny response saying:

<headers>
Status: 400 Unauthorized
Content-Type: text/plain; charset=UTF-8
<body>
Failure: 400 Bad Request

I cannot find what I'm I doing wrong, I'm following the S3 signing process. I need to understand this as I need to implement a plugin to Walrus for my app, so I cannot use third party tools.

Hope that someone can help, thanks.

authorization request cloud-storage eucalyptus
Martin Hope
Aleksandr Makhov
Asked: 2011-09-27 06:12:03 +0800 CST
  • 2

I installed UEC with the help of this manual. And and I had not any problems until I tried to enter the section Store in Web-interface. At this point, get error messages:

Error 60: server certificate verification failed. CAfile: /etc/ssl/certs/ca-certificates.crt CRLfile: none

What could be the problem?

ubuntu cloud cloud-storage eucalyptus
Martin Hope
pallaire
Asked: 2011-06-10 08:49:49 +0800 CST
  • 2

I have downloaded the Amazon AWS SDK for C#, I have no problem accessing the EC2 part of our private cloud running Eucalyptus, I can list, Images, Instances, Zones ...

This is working fine :

AmazonEC2 ec2 = AWSClientFactory.CreateAmazonEC2Client("abcdefghijklmnopqrstuvwxyz1234567890", "abcdefghijklmnopqrstuvwxyz1234567890", new AmazonEC2Config().WithServiceURL("http://10.140.54.12:8773/services/Eucalyptus"));

DescribeInstancesRequest ec2Request = new DescribeInstancesRequest();
try
{
    DescribeInstancesResponse ec2Response = ec2.DescribeInstances(ec2Request);
    int numInstances = 0;
    numInstances = ec2Response.DescribeInstancesResult.Reservation.Count;
    textBoxInstancesLog.AppendText("You have " + numInstances + " running instances");
    textBoxInstancesLog.AppendText(ec2Response.ToString());
}
catch (Exception ex)
{
    MessageBox.Show(ex.ToString());
}

But I need to access the Walrus (S3) part of our Cloud. This is how I try to access the Walrus, the code is almost identical, but with this call I will get an exception.

This is not working:

AmazonS3 s3 = AWSClientFactory.CreateAmazonS3Client("abcdefghijklmnopqrstuvwxyz1234567890", "abcdefghijklmnopqrstuvwxyz1234567890", new AmazonS3Config().WithServiceURL("http://10.140.54.12:8773/services/Walrus"));
ListBucketsRequest s3Request = new ListBucketsRequest();
try
{
    ListBucketsResponse s3Response = s3.ListBuckets(s3Request);
    textBoxS3Log.AppendText(s3Response.ToString());
}
catch (Exception ex)
{
    MessageBox.Show(ex.ToString());
}

I will receive this exception : 

System.Net.WebException: The remote name could not be resolved: 'http'
   at Amazon.S3.AmazonS3Client.processRequestError(String actionName, HttpWebRequest request, WebException we, HttpWebResponse errorResponse, String requestAddr, WebHeaderCollection& respHdrs, Type t, Exception& cause)
   at Amazon.S3.AmazonS3Client.handleHttpWebErrorResponse(S3Request userRequest, WebException we, HttpWebRequest request, HttpWebResponse httpResponse, Exception& cause, HttpStatusCode& statusCode)
   at Amazon.S3.AmazonS3Client.getResponseCallback[T](IAsyncResult result)
   at Amazon.S3.AmazonS3Client.endOperation[T](IAsyncResult result)
   at Amazon.S3.AmazonS3Client.ListBuckets(ListBucketsRequest request)
   at IAASClient.FormMain.buttonS3Test_Click(Object sender, EventArgs e) in X:\work\IAASClient\FormMain.cs:line 107

From Eucalyptus site :

Eucalyptus implements an IaaS (Infrastructure as a Service) private cloud that is accessible via an API compatible with Amazon EC2 and Amazon S3

What am I missing ?

Note: The same code work flawlessly with Amazon S3, the problem is to access Eucalyptus Walrus.

amazon-s3 eucalyptus
Martin Hope
itgorilla
Asked: 2011-05-20 08:22:20 +0800 CST
  • 2

I've read a couple of articles that Eucalyptus doesn't support bootable EBS yet. This is a problem since you can't make a backup form an instance like you a able on Amazon cloud or Rack Space Cloud.

If you ever reboot the physical Ecalyptus sever that's running the node controller the instance is gone along with all your settings. Are there any alternatives to Eucalyptus or is just the only game in town when it comes to open source cloud?

cloud-computing cloud-storage eucalyptus ubuntu-enterprise-cloud
Martin Hope
Patrick.C
Asked: 2011-05-01 22:33:35 +0800 CST
  • 3

I'm new in using cloud computing technologies. There is a popular cloud software called Eucalyptus.

I found that the way to start a VM in Eucalyptus is very different from Virtualbox. In Virtualbox, we can create a virtual disk file, install any OS into the disk file, and then we can use the disk file as a virtual harddisk to start a VM. The whole process is straight forward and easy to understand.

On the other hand, I found that the way to start a VM in Eucalyptus is quite complicated. First, we need to create raw disk file and install the guest OS. Then, we need to split the VM image into kernel image, ramdisk image, and disk image. We also need to perform the so called "bundling" process on those image files before we can use it.

I don't understand why Eucalyptus make it so difficult to start a VM. Why can't it use the Virtualbox method which is much easier?

May I know what is the purpose of splitting a VM into kernel image, ramdisk image, and disk image? If the VM is a Windows VM, then how are we going to split it?

Why can't we use the raw disk file directly? What is the purpose of bundling an image?

virtualization kvm-virtualization cloud-computing virtual-machines eucalyptus