MidVision RapidDeploy for WAS ND Edition on AWS

Amazon EC2 Installation guide

Instructions for Early Fixpack versions for WAS ND:

If you are using AWS Images containing the following fixpack versions of IBM WebSphere Application Server Network Deployment Edition, please follow these installation instructions.

7.0.0.37, 7.0.0.39, 8.0.0.7, 8.0.0.10, 8.5.0.2.

For all other WebSphere Network Deployment images on any other fixpack versions, continue to follow the instructions on this page.

1. Overview

This guide will provide step by step instructions how to start using the installed products on your AWS EC2 instance.

This AMI contains a ready-to-run MidVision RapidDeploy for WAS ND image at the selected version. In order to use this image, you need to launch it with your selected type, and log in via SSH to create required profile(s) and activate (start) the WebSphere servers.

Visit the MidVision RapidDeploy for WAS ND page on the AWS Marketplace
Launch from AWS

2. Launching an instance

In order to launch a MidVision RapidDeploy for WAS ND instance, a few settings need to be configured on the AWS console as follows. The instructions for launching an instance differ depending on where you launch from. Initially you will launch the instance from the AWS Marketplace.

You can also view the YouTube video which will take you through the launch and configuration steps for this AMI.

Follow the step-by-step YouTube video to launch and configure your IBM WebSphere ND instance.

3. Initial login and setup

4. Accessing the Command Line & Web Consoles

4.1 Access the IBM WebSphere Application Server Admin Console

4.2 Subsequent logins via SSH to the instance

When logging into the instance after the first time, you will be asked if you need to reassign the bound hostnames in the IBM WebSphere configuration. You will only need to do this if you have not used the default (localhost) configuration and have not used elastic IP addresses in bespoke configuration, and you have stopped and started your instance.

In this situation, your instance will probably get a new IP address, whereas the IBM WebSphere configuration still references the old host name of the instance in the deployment manager and node agent profile configurations.

In order to fix this issue, you can run the setuphost.sh script provided in the midvision user home directory, which will search and replace for the old host name occurrences in the configuration files. See the section below for detailed instructions running setuphost.sh

4.3 Access the RapidDeploy Web Console

RapidDeploy server and agent will start up automatically when you start your instance. You can access the web console on:

http://[publicip]:9090/MidVision

Note: make sure you have port 9090 open in your Security group when trying to access RapidDeploy web console. For the first time, you can set the password for the default username ” mvadmin “. 

5. Maintaining the installation

The following sections cover some of the activities you might want to perform after the instance has been started.
  • midvision: This is the default user, which you can log in as. It is permitted to use all SUDO rights. To switch to the root user, type “sudo su“.
  • root:      This is the superuser in linux systems. You can log in as any other user without using passwords. E.g: “su ec2-user“, “su midvision“
  • ec2-user:  This user does not have SUDO rights. If you want to switch back to root user, type “exit“, this will take you back to the previous user session.

The IBM WebSphere Application Server Network Deployment Admin Console is initially configured so that you can log in to WebSphere console as “wasadmin” using your ‘[instance-id]‘ as password.

You can further configure WebSphere security in the normal way according to your requirements. However you should pay attention to some AWS specifics covered on this page.

You can use the managewas.sh in the midvision home directory to stop and start profiles (Deployment Manager and/or Node Agent) on your instance. The script takes the argument ‘stop‘ or ‘start’. It will detect all profiles and perform the action on each profile in turn. For example:

sudo /home/midvision/managewas.sh stop

You can open the necessary ports quickly using the setup wizard.  To re-run this section of the wizard, from the midvision home directory run:

sudo /home/midvision/openwasports.sh

There is an option to open all ports used by IBM WebSphere services, by entering ‘process‘ input. Alternatively, you can choose to open all ports on the instance by entering ‘all‘ input. If you are creating a distributed cell over multiple nodes, we strongly recommend you don’t try to estimate which ports should be opened, but select the ‘process’ option, which will open all the required ports to ensure node agent synchronisation will work correctly. Port opening options:

  • <port number> – Open this port. Multiple ports can be entered, separated by pressing the ‘[return]’ key.
  •  ‘process’           – Open all WebSphere ports for all WebSphere processes detected on the instance.
  • ‘all’                    – Open all ports on the instance (effectively remove the firewall).
  • ‘exit’                  – Exit the port opening script.

You can re-run the first run setup wizard at any time by issuing the following command:

sudo /home/midvision/firstrunsetup.sh

RHEL instances are shipped with a firewall by default to protect your machine. For security reasons, the instance is only accessible via SSH (port 22) at first, so further ports can be opened on the firewall as needed. You will need to open all the ports in this internal firewall, which you have open in your Security group. There is a script placed in the user home of midvision (/home/midvision), which is also the starting location when logged in. You will need to be the root user to run this script. Example usage:

[midvision@ ] sudo ./open-firewall.sh 9090
Open firewall port 9090 iptables: Saving firewall rules to /etc/sysconfig/iptables:[ OK ]
 

WE RECOMMEND USING THE DEFAULT CONFIGURATION OR AN ELASTIC IP ADDRESS DURING BESPOKE CONFIGURATION TO AVOID PERFORMING THESE OPERATIONS ON EACH RESTART

If you restart your instance(s) and are using a bespoke configuration and have not assigned elastic IP addresses, your instance(s) will probably get new IP address(es), whereas the IBM WebSphere configuration still references the old host name of the instance in the deployment manager and node agent profile configurations. In order to fix this issue, you can run the setuphost.sh script provided in the midvision user home directory, which will search and replace for the old host name occurrences in the configuration files. See the section below for detailed instructions running setuphost.sh

sudo /home/midvision/setuphost.sh

Example changing IP addresses in a distributed system

Please see below for the example steps for the case where a Deployment Manager and remote Node Agent instance are both stopped and restarted without an elastic IP address: Take a note of the four IP addresses: the original deployment manager host (referred as oldDmgrHost), the original node agent host (ref oldNodeagentHost), the current deployment manager host (ref newDmgrHost), the current node agent host (refnewNodeAgentHost)

  • Record [oldDmgrHost] and [oldNodeagentHost] addresses
  • Start DMGR instance – get a [newDmgrHost] instance address
  • Login to Dmgr instance – You will be prompted for [oldDmgrHost] and [newDmgrHost] host to replace. Answer and run.
  • Start NodeAgent instance – get a [newNodeAgentHost] instance address
  • Login to NodeAgent instance – You will be prompted for [oldNodeagentHost] and [newNodeAgentHost] host to replace. Answer and run.
  • On Dmgr host , run sudo ./setuphost.sh, entering the [oldNodeagentHost] and [newNodeAgentHost]
  • On NodeAgent host, run ./setuphost.sh, entering the [oldDmgrHost] and [newDmgrHost]
  • Restart the NodeAgent
  • In the AdminConsole, run a full resynchronise from System Administration -> Nodes -> Select the nodes and click “Full resynchronise“

There are a few scripts and other files in midvision and root users home directory, which will need to remain unchanged in order to keep the provided scripts working. Files are placed under /root/WAS_install are required to create the IBM WebSphere profiles. There are a set of scripts in midvision home directory. Those scripts are executed when logging into the instance controlled by .bash_profile, but you can also execute them manually whenever they are needed. There are some hidden files used as well,

.firstrun indicates that the setup wizard has already ran once,

.dontask indicates that the user will not be prompted again for running the setuphost script (if chosen not to ask again),

.host file is needed to determine the previously bound hostname to WebSphere service.

To restart RapidDeploy manually from the filesystem, use the rapiddeploy linux service. The RapidDeploy version is 5.0.4. RapidDeploy home is located at /var/rd/midvision Example usage:

[midvision@ ] sudo service rapiddeploystart start 
[midvision@ ] sudo service rapiddeploystart stop

The RapidDeploy server uses an in-memory database, so your RapidDeploy framework needs to be shut down properly to save your work. This happens when the instance is stopping or when calling the stop service manually from the command line. Note that on instance restart, unsaved data will be lost.

For security reasons, you will need to change the default password (default value is ‘[instance-id]‘) for user mvadmin. This will be requested the first time only when you log in to the RapidDeploy server.

6. Troubleshooting

6.1 Session loss during setup

If you lose your SSH connection to the target instance during the first run setup script execution (e.g. as a result of a network problem), we advise you to delete and recreate the EC2 instance and run the script again.

6.2 Profile startup exceptions

If the profile fails to start with the startmanager.sh or startnode.sh commands, check that you are running as the root user, either by being in a root shell, or by prepending ‘sudo’ to the start command.

6.3 Cannot access the IBM WebSphere admin console (ISC)

Check that  the default server is running, and you have correctly opened all the required ports on the firewall, and that your instance was created using a security group definition that allows TCP access to the instance on the required ports.

6.4 java.net.BindException: Cannot assign requested address: bind

We experienced this error on ec2-instances with no public-dns. When you set numeric public ip you need to add some changes to your network configuration. 

[root@ip-172-30-0-134 midvision]# nmcli con mod "System eth0" ipv4.addresses [numeric-public-ip]
[root@ip-172-30-0-134 midvision]# nmcli con mod "System eth0" +ipv4.dns [numeric-public-ip]
[root@ip-172-30-0-134 midvision]# nmcli con up "System eth0"

6.5 Contacting MidVision support

Please visit our support website.

7. Reference

7.1 Creating a Deployment manager profile

You will be prompted to enter the following values, or accept the defaults:

  • dmgrHostName: the host name binding for the deployment manager profile. This resource has a default value of the current public host name of the instance. You are strongly recommended to assign an elastic IP address for the instance before proceeding with the deployment manager profile creation.
  • cellID: the id of the WebSphere cell containing the profile
  • startingPort: the WebSphere service ports will be incremented from this number, assigning each unused ports to the WebSphere services.
  • profileName: the name of the profile to create

7.2 Creating a Node agent profile

You will be prompted to enter the following values, or accept the defaults:

  • nodeAgentHostName: the host name binding for the Node Agent profile. This resource has a default value of the current public host name of the instance. It is strongly recommended to assign an elastic IP address for the instance before proceeding with the node agent profile creation.

  • nodeAgentStartingPort: the WebSphere service ports will be incremented from this number, assigning each unused port to the WebSphere services.

  • nodeAgentCellID: the id of the WebSphere cell containing the profile

  • nodeAgentAdminUserName: the username to authenticate against the deployment manager (if required)

  • nodeAgentAdminPassword: the password to authenticate against the deployment manager (if required)

  • nodeAgentProfileName: the name of the profile

  • dmgrHostName: the host name the deployment manager is running at. This resource has a default value of the current public host name of the instance. If necessary,change this to the hostname, IP or DNS of the remote Deployment Manager instance. Note that if you use a remote deployment manager host, you will need to make sure all the required ports are open on both machines.

  • dmgrPort: the deployment manager SOAP connector port to connect to for the purposes of federation and subsequent synchronisation.