Release: Amazon EC2 on 2010-07-12

Release Notes>Amazon EC2>Release: Amazon EC2 on 2010 07 12
Support for the new cluster compute instance type (cc1.4xlarge), and the ability to specify the IP address for an instance launched in an Amazon VPC subnet.

Details

Submitted By: cheriaws
Release Date: July 12, 2010 12:00 AM GMT
Latest Version: 2010-06-15
Latest WSDL: http://ec2.amazonaws.com/doc/2010-06-15/AmazonEC2.wsdl
Created On: July 13, 2010 1:39 AM GMT
Last Updated: July 15, 2010 5:30 PM GMT

New Features

FeatureDescription
New Instance Type Amazon EC2 now supports a new instance type: Cluster Compute (cc1.4xlarge). This type is available only for Linux/UNIX in the US-East (Northern Virginia) Region.

For more information about the new instance type, go to:

New API Version With this release, Amazon EC2 has a new API version (2010-06-15). The WSDL is at http://ec2.amazonaws.com/doc/2010-06-15/AmazonEC2.wsdl.
Placement Groups We've introduced the placement group, which is a logical grouping of instances. You create a cluster placement group (a placement group that has its strategy=cluster) for the purposes of bringing together cluster compute instances into a cluster. There are three new API actions for managing placement groups: CreatePlacementGroup, DescribePlacementGroups, and DeletePlacementGroup.

The API tools have corresponding new commands: ec2-add-placement-group, ec2-describe-placement-groups, and ec2-delete-placement-group. To get the latest version of the API tools, go to Amazon EC2 API Tools.

To launch cluster compute instances into a cluster, just specify the cluster placement group when you launch the instances. The RunInstances action and corresponding ec2-run-instances command both now accept a placement group value. When you describe an instance, its placement group is included in the results.

Important: We recommend that before you launch any cluster compute instances into a cluster placement group, you first estimate the total number of instances you'll need in that cluster, and then launch all of them at once with a single RunInstances call. If you launch only a few and then try adding to the placement group later, you increase your chances of getting an "Insufficient capacity" error.

The AWS Management Console now also supports placement groups. There's a new page in the console where you can create, view, and delete your placement groups. You can configure the Instances page in the console to include a column that displays the instance's placement group. The console's launch wizard lets you specify a placement group when you launch a cluster compute instance.

You're not required to specify a placement group when you launch a cluster compute instance. Note that if you run cluster compute instances in a cluster placement group, Reserved Instance requests and Auto Scaling are not available for the instances.

AMI Virtualization Type You must launch all cluster compute instances from an HVM-based, Amazon EBS-backed AMI. This is the first instance type with this requirement. Currently only Linux/UNIX AMIs are supported. We've provided a CentOS reference AMI: ami-7ea24a17. You can customize an instance of the reference AMI and then create your own image.

The cluster compute AMIs use hardware-assisted virtualization (HVM), in contrast with the other Linux/UNIX AMIs that Amazon EC2 offers, which use paravirtualization (PVM). Starting with the 2010-06-15 API version, all AMIs are labeled with a virtualization type, (hvm or paravirtual). When you search for an AMI in the console, or describe an image through the API or API tools, the image's virtualization type is included in the results.

Note: The reference AMI was originally created using CentOS 5.4. Since that time, CentOS has released 5.5. If you invoke "yum update", the AMI is automatically updated to 5.5.

Amazon VPC IP Address Designation Before this release, when you launched an instance into a VPC, we assigned an available IP address from the subnet to the instance. With the 2010-06-15 API version, you can now optionally specify which of the subnet's available IP addresses you want to assign the instance. If the address you specify is already in use, the service returns an InvalidIPAddress.InUse client error.

The ec2-run-instances command now accepts a --private-ip-address parameter. For example:

> ec2-run-instances ami-1a2b3c4d --subnet subnet-5d6c7b8a --private-ip-address 10.0.0.25 ...

For more information, go to ec2-run-instances in the Amazon Elastic Compute Cloud Command Line Reference. To get the latest version of the API tools, go to Amazon EC2 API Tools.

The RunInstances action in the Query API now takes a PrivateIpAddress parameter. The RunInstances action in the SOAP API now takes a privateIpAddress element. For more information, go to RunInstances (for the Query API) and RunInstances (for the SOAP API) in the API reference.

Known Issues

IssueDescription
Current Limitations for Cluster Compute Instances Following are current limitations of cluster compute instances:
  • They support Linux/UNIX operating systems, but currently not Microsoft Windows Server
  • They are available only in the US-East (Northern Virginia) Region
  • Spot Instance requests for cluster compute instances are not supported
  • They are not available for use with Amazon Virtual Private Cloud
  • They are not available for use with Amazon DevPay
Current Limitations for Cluster Placement Groups Following are current limitations of cluster placement groups:
  • You can't currently use Auto Scaling to launch cluster compute instances into a cluster placement group
  • Reserved Instances are not currently available within a cluster placement group
Setting the TCP_MAXSEG Socket Option The 2.6.18-164.15.1 kernel used in the CentOS 5.4 reference AMI (ami-7ea24a17) contains a known issue where setting the TCP_MAXSEG socket option on TCP sockets to certain values (known values include 1500) causes the kernel to generate TCP Segmentation Offload (TSO) packets of invalid size that are then dropped by the NIC driver, resulting in very low network throughput. Don't set this socket option and let the kernel handle segmentation and Path Maximum Transmission Unit (PMTU) discovery using the default settings.
Query Version of ModifyInstanceAttribute The Query API ModifyInstanceAttribute request does not allow you to modify the block-device-mappings, and you can't modify the user-data attribute if it already contains data.
Paid AMIs Backed by Amazon EBS Amazon EBS-backed AMIs are not currently supported by Amazon DevPay.
Windows AMI launch times Windows AMIs take longer to launch than Linux/UNIX instances due to larger AMI sizes and multiple reboots.
Windows AMI sizes Installing software on Amazon S3-backed Windows AMIs can cause them to become large and easily reach the 10 GB limit. Before bundling, check the size of the C:\ volume.
Limitation on drive mapping There are limitations on devices available for storage attachment. For more information, go to How to Attach the Volume to an Instance in the Amazon Elastic Compute Cloud User Guide.
Instance limit New users are limited to a maximum of 20 concurrent instances, but many of our customers use hundreds or thousands of instances. If you need a higher limit, go to http://aws.amazon.com/contact-us/ec2-request.
©2014, Amazon Web Services, Inc. or its affiliates. All rights reserved.