Dart Ami

From Biowiki
Jump to: navigation, search

Running the DART AMI

Notes

At $0.10/hour, the Amazon EC2 provides a practical alternative to cluster computing for some applications. We've provided an AMI based on Amazon's Fedora 8 AMI. This includes a compiled DART install, as well as HMMo C and some other tools (GSIMULATOR, GCONSTRUCTOR, GFF tools). This page will give a brief intro on getting the instance running. For more information, see Amazon's documentation linked at the bottom of this page.

As of 4/1/09, this image is merged with the JBrowse AMI.

For a list of the most current AMIs provided by the Holmes lab, and the associated identifiers (ami-xxxxxxxx), see this page: Amazon Machine Image IDs

NB: To run more than the default 20 instances at a time, you will have to submit a request to Amazon through this webform.

Preparation

Install the Amazon EC2 API Tools. You will also need your EC2 private key and certificate, which you should have obtained when setting up your Amazon Web Services account. See the Getting Started Guide for help with this.

Running a DART-AMI instance

Generate a keypair:

ec2-add-keypair dartami-keypair

Save your private key to a local file. You will use this to log in to your instances.

Start an instance:

ec2-run-instances ami-xxxxxxxx -k dartami-keypair

In the above, replace ami-xxxxxxxx with the latest AMI ID from Amazon Machine Image IDs.

This will return an instance id that looks something like:

i-<8-character-identifier>

Authorize ssh communication with your instance

ec2-authorize default -p 22 -s <your-ip-range>

Get your instance's URL:

ec2-describe-instances i-<id>

Once this returns your URL (this may take a few minutes), you should be able to log in to your instance:

ssh -i <your-private-key-file> root@ec2-<stuff>.amazonaws.com

Once logged in, .bash_profile is set up so that the DART binaries and perl libraries will be in your path. Run your analysis and enjoy!

Amazon EC2 Documentation

Amazon's EC2 page

Some relevant links to documentation, etc

-- Lars Barquist - 13 Mar 2009