Elastic Load Balancing (ELB) with a Java Web Application + Tomcat + Session Stickiness

Suppose you have a web application and you want to deploy it in Amazon cloud environment with load balance support. The whole process is pretty straight-forward and it generally doesn't take much time.

For this post, I'm using Apache Tomcat web server and I already have a war file from my HelloWorld application. 

Here is the Tomcat version I'm using:

I'm using two instances and I have extracted my tomcat zip file into /opt/ folder in each of those two instances. I have also placed HelloWorld.jar file into /opt/apache-tomcat-7.0.39/webapps folder.

Now, I will go to each of those two instances and will start tomcat server. After some minutes (or seconds) I should see my deployed web application is up and running. Which means, I can navigate to these URLs and able to see Log-In screen (initial page of my web app).

• http://ip.address.instance-1:8080/HelloWorld/login.jsp
• http://ip.address.instance-2:8080/HelloWorld/login.jsp

All of the above steps which I described so far, have nothing to do with Elastic Load Balancing (ELB). Just like everyone, I just deployed a web app in tomcat server. Before I start showing steps for ELB, I'm assuming your web application is also up and running and you can navigate through URLs separately.


Create Load Balancer

Step#1: On AWS EC2 console, click on the Load Balancer option under "Network & Security" section. If you do not have any ELB yet, you will see an empty list. Click on "Create Load Balancer" button.

Step#2: Write a name of your Load Balancer, this name will be used when it will create a default link. I'm also creating this Load Balancer inside my Virtual Private Cloud (VPC) that's why I'm selecting a specific VPC Id. By default, you might see only port 80 in the listerner configuration list, I have added port 8080 as my web app is running on port 8080. Add appropriate port based on your web application and click "Continue".

Step#3: This screen is dedicated to Health check configuration. Based on configuration, ELB will ping that path with that port to check the health condition and if it fails it will automatically remove your instances from the load balancer.

Since, Log-In is the default screen of my application (welcome page), so I'm using the path of Login screen as my ping path.

Step#4: Choose your Subnet id based on where you want to use your Load Balancer. For my case, subnet-2e961843 is my expected Subnet id.

Step#5: Next screen will ask you to select your security groups. I already have a security group for my VPC and I'm using it here too.

Step#6: In the "Add EC2 Instances" section, add the instances in where you already deployed Tomcat and your web application.

Step#7: This screen is for review purpose. Once you review it you can finally create your load balancer by clicking on the "Create" button.

Step#8: Once you create your balancer, it will redirect you to Load Balancer list and now you will see your newly created load balancer in the list. DNS Name column shows newly created DNS Name for your load balancer and you should be able to navigate it with proper port.

So for my case, I can navigate my load balancer by using:

http://helloworld-353060791.us-east-1.elb.amazonaws.com:8080/HelloWorld/login.jsp


Sticky Session:
Since you are using Tomcat with load balancer, it's pretty obvious that you might want to enable sticky session with session replication in Tomcat. My web application is a Spring MVC application and it uses Spring Security for all type of authorizations and authentications. If I directly go to the Log-In screen of my load balancer and try to authenticate, it might not work. It's expected as Tomcat gets confused when sending request and response in multiple instances. If I enable sticky session I will not face this issue.

You can do it with the help of AWS EC2 console. Open the Load Balancer screen and select your newly created load balancer.

If you look carefully at the port configuration part, you will see "Stickiness: Disabled" for all of your ports. By default, stickiness is disabled for all the ports you select for load balancer. Now click on the "edit" button of the port on where you want to enable stickiness. For my case, it will be port 8080. Once you click on the "edit" button, it will ask you how you want to enable session stickiness. You can either choose Load Balancer Generated Cookie Stickiness or Application Generated Cookie Stickiness. For my simple application, I have selected "Load Balancer Generated Cookie Stickiness" and I entered 86400 as my cookie expiration period which is a day in seconds.

After you enable it, you should be able to test your session stickiness. For my case, now I'm able to successfully authenticate to my application.

Some considerations: Sometimes you might see your load balancer is down or the link is not working or shows no page. In that case, best way to quickly test is to check each of the instance where tomcat is running and check whether you can access them individually (e.g. http://ip.address.instance-1:8080/HelloWorld/login.jsp). If you find that each of the instance is up and running, you can try removing them from your load balancer and add them again. Remember, "Status" section under "Description" tab of your load balancer does not get updated instantly. It takes some time and it waits for the result of the next health check. So wait few minutes until you see "Status: N of N instances in service".
.

That's pretty much it! This is the very basic AWS Load Balancer example with minimum configuration of Tomcat + Session Stickiness. Once its working for you, you can try other options (highly encouraged) and see how it works for you.


Note: For privacy purpose, I had to modify several lines on this post from my original post. So if you find something is not working or facing any issues, please do not hesitate to contact me :)

Configure Ganglia for multiple clusters in Unicast mode

In my previous post I talked about how to: Setting up Ganglia in CentOS environment. At that time, I used only a single cluster for the whole setup. But it's highly unlikely that you have only a single cluster in your development/production environment. Consider you have two clusters - 1. Storm 2. Kafka and you want to monitor all of these cluster nodes through a single Ganglia UI. You do not have to install Ganglia multiple times for that, you just need to configure your Ganglia. It would have been much easier if AWS supports multicast but as it doesn't support multicast, you need to do a work-around in unicast mode to achieve monitoring multiple clusters in one single Ganglia.

The idea behind this work-around is pretty straightforward. Suppose I have two clusters: cluster#1 - Storm and cluster#2 - Kafka and their respective IP addresses are:

10.0.0.194 - Storm Cluster (supervisor 1)

10.0.0.195 - Storm Cluster (supervisor 2)

10.0.0.196 - Storm Cluster (supervisor 3)

10.0.0.182 - Storm Cluster (nimbus)

10.0.0.249 - Kafka Cluster

10.0.0.250 - Kafka Cluster

10.0.0.251 - Kafka Cluster

10.0.0.33 - my client machine

What I am going to do is, I will configure each of the cluster to send collected data (gmond) to one of their specific node only and configure the gmetad daemon in a way that it can collects the data only from a designated node (gmond daemon) from each cluster. Ganglia will categorize each cluster data by their unique cluster name defined in gmond.conf file.

As you can see in the above figure that all Kakfa cluster's data is sending to one specific node - 10.0.0.249 and all Storm cluster's data is sending to one of its node - 10.0.0.182. Client machine (10.0.0.33) is running gmetad daemon and I will configure that daemon so that it can look for two data sources for two clusters where their source IP addresses will be 10.0.0.249 and 10.0.0.182 for Kafka and Storm respectively.

I'm assuming that you already setup your Ganglia and it's running as expected. So I am not going to discuess about what is gmond.conf and gmetad.conf files. In case if you have not setup yet, you might want to take a look at this post. 

This is my gmond.conf file (only the part which I modified) which I'm using for all Kafka hosts (this file is unique for each host per cluster):

And here is my gmond.conf file for all Storm hosts (this file is unique for each host per cluster):

You notice that I'm using unique host address for udp_send_channel for each cluster. Now, I need to tell my gmetad daemon to look for those two host address to collect data from. Here is my gmetad.conf file:
You are done! Now restart all gmond daemons and gmetad daemon and wait for few minutes.
Once you navigate to your Ganglia UI url you should be able to see your grid and list of your clusters in the drop-down.

You can dig further to see each of your host for each cluster:

There is another work-around which you can also try to get a better understanding of Ganglia. In that case you need to use separate port number for each cluster. Here, I'm distinguishing each cluster's data source per IP address, but in that work-around you can have a single IP address for all clusters but multiple port numbers. You can try that work-around as an exercise :).



Note: For privacy purpose, I had to modify several lines on this post from my original post. So if you find something is not working or facing any issues, please do not hesitate to contact me.

Setup a Storm cluster on Amazon EC2

Storm - Real-time  Hadoop! Yes, you can call it in that way. As you know, Hadoop provides a set of general primitives for doing batch processing, Storm also provides a set of primitives for doing real-time computation. It's a very powerful tool and pretty straight forward to setup a Storm cluster. If you want to setup a Strom cluster on Amazon EC2, you should try Nathan's storm-deploy project first which will auto deploy a Storm cluster on EC2 based on your configurations. But if you want to manually deploy a Storm cluster, you can follow these steps (if you want more detailed information, you can also follow original documentation of Storm):

Let me show you my machine's current configuration first:

• Machine type: m1.large (for supervisor) and m1.small (for nimbus)
• OS: 64-bit CentOS 6.3
• JRE Version: 1.6.0_43
• JDK Version: 1.6.0_43
• Python Version: 2.6.6

For this tutorial, I am going to setup a 3-node Storm cluster. IP addresses of each hosts and my targetted configurations is:

10.0.0.194 - StormSupervisor1

10.0.0.195 - StormSupervisor2

10.0.0.196 - StormSupervisor3

10.0.0.182 - StormNimbus

Storm depends on Zookeeper for coordinating the cluster. I have already installed Zookeeper in each of those above hosts. Now apply each of the following steps on all Supervisor and Nimbus nodes:

A. Install ZeroMQ 2.1.7

Step #A-1: Download zeromq-2.1.7.tar.gz from http://download.zeromq.org/.

Step #A-2: Extract the gzip file:

[root@ip-10-0-0-194 tool]# tar -zxvf zeromq-2.1.7.tar.gz

Step #A-3: Build ZeroMQ and update the library:

Note: If you are facing "cannot link with -luuid, install uuid-dev." error when you are executing "./configure", then you need to install it. You can install it by executing "yum install libuuid-devel".


B. Install JZMQ

Step #B-1: Get the project from the Git by executing:
Step #B-2: Install it:

C. Setup Storm

Step #C-1: Download the latest version (for this tutorial, I'm using 0.8.1 version) from https://github.com/nathanmarz/storm/downloads.

Step #C-2: Unzip the downloaded zip file:

[root@ip-10-0-0-194 tool]# unzip storm-0.8.1.zip

Step #C-3: Now change the configuration based on your environment. Default location of the main Storm configuration file is: "/storm/conf/storm.yaml". Any setting you write on this file will overwrite default configuration file. Here is what I changed in the storm.yaml file:
Note: I have created the "storm" folder manually inside "/var" directory.

At this point, you are ready to start your Storm cluster. Here, I installed and setup everything on a single instance first (supervisor1 - 10.0.0.194) and then I created AMI from that instance and later created rest of the two supervisors and one nimbus node from that AMI.

Launch daemons by using the storm script (bin/storm) on each nodes. I started nimbus and UI daemons on the nimbus host and supervisor daemon on each of the supervisor nodes.

• bin/storm nimbus on 10.0.0.182
• bin/storm ui on 10.0.0.182
• bin/storm supervisor on 10.0.0.194,10.0.0.195,10.0.0.196

You can see Storm UI by navigating to your nimbus host: http://{nimbus host}:8080. For my case, it was: http://54.208.24.209:8080 (here, 54.208.24.209 is the public IP address of my nimbus host).

Note: For privacy purpose, I had to modify several lines on this post from my original post. So if you find something is not working or facing any issues, please do not hesitate to contact me.

Install Opscenter in CentOS environment

In my previous post, I talked about how to install Cassandra in CentOS environment. This is the follow-up post of my previous post and here I am going to show you how to install OpsCenter in the same environment. "OpsCenter is a browser-based user interface for monitoring, administering, and configuring multiple Cassandra clusters in a single, centralized management console (ref)". 

Last time, I installed Cassandra cluster on these nodes:

• cassandra node1 -> 10.0.0.57
• cassandra node1 -> 10.0.0.58
• cassandra node1 -> 10.0.0.59

Now I am going to install OpsCenter agents on these nodes and will treat 10.0.0.57 as my client node (that means, OpsCenter console will be deployed on that host). Before OpsCenter installation, make sure your Cassandra cluster is up and running successfully.

Step #1: Create a new yum repository definition for DataStax OpsCenter in 10.0.0.57 node.

[root@ip-10-0-0-57 ~]# vim /etc/yum.repos.d/datastax.repo

Step #2: Write the edition you want to install in the datastax.repo file (I am installing OpsCenter community edition):
Step #3: Install the OpsCenter pacakge:

[root@ip-10-0-0-57 ~]# yum install opscenter-free

The above steps will install the most recent OpsCenter community edition in your system. But I want to install a specific version of OpsCenter today (appropriate for the Cassandra version which I installed earlier). So to do that, at first I need to check the list of versions for OpsCenter which are available now:
I wanted to install 3.0.2-1 version. So, I'm installing it by:

[root@ip-10-0-0-57 ~]# yum install opscenter-free-3.0.2-1

Step #4:  If you do not want to try with the repository and want to install manually any specific version of OpsCenter, in that case you can download the rpm file from http://rpm.datastax.com/community/noarch/ and can install it by:

[root@ip-10-0-0-57 ~]# yum install opscenter-free-3.0.2-1.noarch.rpm

Step #5: Now configure your opscenter configuration file (/etc/opscenter/opscenterd.conf) to mention your web server's IP address or hostname:
Step #6: Now start your OpsCenter by:
Step #7: Now you can see your OpsCenter console by navigating to http://<opscenter_host>:8888. For my case, it would be: http://54.208.29.59:8888 as it's the public IP address for the host 10.0.0.57.

Step #8: Wait, you are not done yet! You still need to install your OpsCenter agents. For the first time when you open your OpsCenter console, it will ask you whether you want to create a new cluster or want to use existing cluster. In previous post, I installed a cluster with the name "Simple Cluster". So I want to install that existing cluster for my OpsCenter. So, I'm selecting "Use Existing Cluster" option.

Step #9: Now, you need to pass a list of hostnames or IP address of my cluster in each line at a time (leave other fields as they are):

Step #10: At this point you should be able to see your OpsCenter console.

Note that on top of your console, there is a notification labeled as "0 of 3 agents connected" with a link called "fix". This is because, none of your OpsCenter agents are installed yet. Click on that link and install agents automatically.

Step #11: Enter appropriate credentials of your machine. For my case, I'm writing "root" as my username and pasting my private key (including the commented part, i.e. --BEGIN RSA PRIVATE KEY):

Click on "Done" and finally install it by clicking on the "Install on all nodes" button.

Step #12: Accept all the fingerprints by clicking on the "Accept Fingerprints" button. Once you click on that button, each of your host will be downloading agent package by connecting to the internet and then it will install and configure agent in the system.

At the end of the installation you should be able to see a successful message on your screen:

As you see that each of the host downloads the agent package from internet, it is required that each of your host can talk to internet. If you do not have that setup, you can also install OpsCenter agent manually.

At the end of your installation you shouldn't be seen "0 of 3 agents connected" notification anymore.

Finally you are done! There are so many things you can do in OpsCenter. I highly recommend you to play with it and do some experiments by changing/applying different settings and configurations. It also comes with a large set of very good performance metrics, so do not forget to check those metrics too.



Note: For privacy purpose, I had to modify several lines on this post from my original post. So if you find something is not working or facing any issues, please do not hesitate to contact me.

Install Cassandra Cluster within 30 minutes

Installing Cassandra cluster is pretty straight forward and you will find a lot of detailed documentations on DataStax site. But if you still find it overwhelming, you can follow the steps I mention here. It's just a sorted version of all of the steps you need to install a basic setup of Cassandra cluster. But I still recommend you to read DataStax documents for detailed information.

I am using a EBS backed CentOS AMI for this setup. You can choose different AMI based on your needs.

Please check the current Cassandra version if you want to setup the most latest one. Today, I am going to install Cassandra 1.2.3 version.

Step #1: Create an instance in Amazon AWS. I choose the following one for this setup:

EBS Backed CentOS 6.3

This is an EBS backed 64-bit 6.3 CentOS AMI. For me, the IP address which was assigned for this server: 10.0.0.57.

Step #2: Once you created this instance, by default it will not allow you to login as root user. So, login to that server as ec2-user.

Step #3: Now allow root login on that machine:

Using username "ec2-user".
Authenticating with public key "imported-openssh-key"
[ec2-user@ip-10-0-0-57 ~]$ sudo su
[root@ip-10-0-0-57 ec2-user]# cd /root/.ssh
[root@ip-10-0-0-57 .ssh]# cp /home/ec2-user/.ssh/authorized_keys .
cp: overwrite `./authorized_keys'? y
[root@ip-10-0-0-57 .ssh]# chmod 700 /root/.ssh
[root@ip-10-0-0-57 .ssh]# chmod 640 /root/.ssh/authorized_keys
[root@ip-10-0-0-57 .ssh]# service sshd restart

Step #4: Now you should be able to login as root user on that server. So, login again as root user and check the current Java version:

[root@ip-10-0-0-57 ~]# java -version

The AMI which I used for this server didn't have Java pre-installed. So, I'm going to install last version of Java 1.6 on this server. It is recommended not to use Java 1.7 for Cassandra.

Step #5: Download Java rpm (jre-6u43-linux-x64-rpm.bin) from the following location:

Oracle jre-6u43-linux-x64-rpm.bin

Step #6: Copy your downloaded rpm to that server by using WinSCP. (You can use any other tool if you want).

Step #7: Give required permissions to that rpm and install it:

[root@ip-10-0-0-57 ~]# chmod a+x jre-6u43-linux-x64-rpm.bin
[root@ip-10-0-0-57 ~]# ./jre-6u43-linux-x64-rpm.bin

Step #8: Java should be installed successfully and it should show you expected information:

[root@ip-10-0-0-57 ~]# java -version
java version "1.6.0_43"
Java(TM) SE Runtime Environment (build 1.6.0_43-b01)
Java HotSpot(TM) 64-Bit Server VM (build 20.14-b01, mixed mode)

Step #9: Now download Cassandra 1.2.3 rpm from the following location and copy it to your server by WinSCP:

DataStax Cassandra 1.2.3 rpm

You can also install it by using DataStax repository (check DataStax Cassandra manual for that), I'm just following this way as a preference.

Step #10: After you copy the Cassandra rpm to your server, install it:

[root@ip-10-0-0-57 ~]# yum install cassandra12-1.2.3-1.noarch.rpm

Step #11: Cassandra should be installed successfully and you can check the status by this (By default, Cassandra server is stopped right after its installed):

[root@ip-10-0-0-57 ~]# /etc/init.d/cassandra status
cassandra is stopped

Step #12: During Cassandra installation, it downloads OpenJDK version of Java as a dependency and it will overwrite your java setting. So follow these to set back your Java to use Oracle version:

Step #13: At this point, your Cassandra is ready to start. But before you start your Cassandra, you need to update its configuration file for your cluster. All configuration file are present in the location "/etc/cassandra/conf" of your server by default if its packaged install. For the cluster, I will change one of the file from there which is cassandra.yaml. The cassandra.yaml is the main configuration file for Cassandra. 

There are so many properties you can change in the main configuration file and you can find its details here: http://www.datastax.com/docs/1.2/configuration/node_configuration. But as I am just installing the most basic version of Cassandra cluster, I will change only the following property:

• cluster_name: Name of your cluster. It will be same for all hosts or instances.
• initial_token: Used in versions prior to 1.2. For this setup, I will manually set this value. But that is not required and you can setup your cluster by using both initial_token & num_tokens property. You can read more about it on the web and can try out different configs once you are familiar with Cassandra.
• partitioner: It determine which node to store the data on. Remember, paritioner cannot be changed without loading your all data. So configure your correct partitioner before initializing your cluster.
• seed_provider: A list of comma-delimited IP addresses to use as contact points when a node joins a cluster. This value should be unique for all your hosts.
• listen_address: Local IP address of each host.
• rpc_address: Listener address for client connections. Make it 0.0.0.0 to listens on all configured interfaces.
• endpoint_snitch: Sets which snitch Cassandra uses for locating nodes and routing requests. Since, I'm installing Cassandra cluster on AWS with a single region, I will be using EC2Snitch. Again, this value should be unique for all of your hosts.

So, after modifying, here is my updated configuration file (reflects only changes which I made):

Step #14:

Murmur3Partitioner: This is a new partitioner which is available from Cassandra 1.2 version. initial_token value is depends on the partitioner you are using. To generate initial_token value for Murmur3Partitioner, you can run the following commands:

Note: Here, 3 is the number of nodes which I will be using for my cluster setup. Change that value based on your needs.

RandomPartitioner: If you want to use RandomParitioner then in that case you can generate your initial_token value by using the tooken-generator tool which comes with Cassandra installation:
Step #15: Now, I am going to create a new AMI from my current instance so that when I create a new instance from that, I will have Cassandra installation ready with expected configuration.

Since, I'm installing a 3 node cluster, I will create two new instances from the AMI which I just created. So, I got IP addresses like this:

- cassandra node1 -> 10.0.0.57
- cassandra node2 -> 10.0.0.58
- cassandra node3 -> 10.0.0.59

If you are using Amazon Virtual Private Cloud (VPC), then you have the option to choose specific IP address based on your needs.

Step #16: Remember, even though you created instances from the AMI, you still need to change some values in cassandra.yaml file which varies based on hosts. Those are:

Node #2:

initial_token: -3074457345618258603
listen_address: 10.0.0.58

Node #3:

initial_token: 3074457345618258602
listen_address: 10.0.0.59

Step #17: You may need to update your "/etc/hosts" file in case your hostname is not configured. I have updated that file in each server like this:
Step #18: That's it! Now you are ready to start your Cassandra. Execute this in each node to start your cluster:

[root@ip-10-0-0-57 ~]# /etc/init.d/cassandra start
Starting cassandra: OK

Step #19: You can check the status of your Cassandra by executing this:

Two major things here you need to look on the status, those are "Owns" and "Status" columns. You see here all nodes are up and sharing same percentage of the total ownership. 

As I said at the beginning, this is the most basic (or minimum) setup of Cassandra cluster. There are a lot of things you can change, tune and modify based on your needs. Once you are familiar with basic Cassandra, I highly recommend you to do some experiments by using different configurations. The more you try the more you learn!

In another post, I will write about how to install DataStax OpsCenter to monitor this Cassandra Cluster.

Note: For privacy purpose, I had to modify several lines on this post from my original post. So if you find something is not working or facing any issues, please do not hesitate to contact me.