Using our Taverna Server Amazon Machine Image (AMI)
These instructions assume that you have experience in running instances within the Amazon Web Services EC2 infrastructure. For more information on this topic please see the AWS documentation:
The following instructions describe managing your Taverna Server instances using the AWS Management Console but the details are also applicable to the command-line tools or other APIs.
The AMI is located in the EU Region and has the following details:
|Region||EU West (Ireland)|
As there are nearly 8000 AMIs listed in the EU region the easiest way to find our image is to put either the AMI ID or our owner ID in the search box of the AWS Console.
Running the AMI
The AMI can be started in the same way as any other but there are a couple of things to bare in mind when doing so:
This AMI cannot be run satisfactorily on Micro or Small instance types (t1.micro or m1.small). We have seen very odd behaviour on such instances including the whole application just stopping (with no log output) during a workflow run or even just after a seemingly successful run. Some small and lightweight workflows may be run on the Medium instance type (m1.medium) but due to the high memory and CPU requirements of the application we find that at least a Large instance (m1.large) should be used.
For high numbers of concurrent users, or for workflows which run local tools or contain many local workers you may find that even larger instance types are needed. Due to the fact that Taverna Server is built on top of proven server software, such as Apache Tomcat, it is well able to scale up as the instance size increases.
Kernel and RAM disk
The defaults selected by the AWS Console should be used.
Firewall settings (Security Group)
Taverna Server uses the standard Tomcat ports in its default configuration. As a minimum you should allow port 8080 through your firewall (along with port 22 for SSH if you want to log in yourself, of course). If you configure your Taverna Server to use HTTPS then you should open port 8443 as well. If you turn off normal HTTP operation you may close port 8080.
This AMI is based on an official Ubuntu AMI so the user account name is "ubuntu". So logging in with SSH will look something like this:
my-aws-key.pem is the user credential that you have already registered with AWS or created while starting the instance and
ec2-W-X-Y-Z.eu-west-1.compute.amazonaws.com is the public DNS name of your instance.
Using the Taverna Server
The Taverna Server on the image is ready to use.
The various endpoints that Taverna Server presents will, of course, depend on what the DNS entry of your instance is but they all have a common root and by default use port 8080 as explained above. Using
taverna.example.org as an example server name the following endpoints are provided by our AMI:
|Taverna Server root|
|Taverna Server SOAP service listing|
|Taverna Server WSDL||http://taverna.example.org:8080/taverna-server/soap?wsdl|
|Taverna Server root REST interface||http://taverna.example.org:8080/taverna-server/rest|
|Taverna Server root REST WADL||http://taverna.example.org:8080/taverna-server/rest?_wadl|
|Taverna Server admin REST interface||/admin|
|Taverna Server admin REST WADL||admin?_wadl /|
Username and password
There is a default user set up:
This user is both a regular taverna user and an administrator user. For details about these user types, how to change the password and how to set up more users please see the Taverna Server installation and configuration documentation.
Updates to the AMI
We will provide updated versions of Taverna Server, and apply bug-fixes, when we release them and we will create updated AMIs. When we do this we will also apply any OS updates that have been released as well. The AMI details table at the top of this page will be updated to reflect any new AMI that we make available.
Between such releases please ensure that any OS updates are applied to your running instances periodically.
For more detailed information on managing a Taverna Server instance, for using it via a client library or via the REST or SOAP API please see the Taverna Server documentation main page.