Currently multilingual, CirrusGrid PaaS was initially created as pure Java Cloud and still maintains a primary focus on this programming language. The maturity of the platform technology was recognized by the Java community with Duke’s Choice Award already twice (in 2012 and 2018). This guide acquaints you with the distinctive features of Java hosting within the CirrusGrid platform and lets you easily get started with a broad functionality list.
Java Versions #
There are following Java versions and distributions available at CirrusGrid:
- AdoptOpenJDK: 8.0.242; 11.0.6; 13.0.2
- Amazon Corretto: 8.242.08.1; 11.0.6.10.1
- Eclipse OpenJ9: 0.11.0 (8u192-b12; 11.0.1); 0.15.1 (8u222-b10; 11.0.4); 0.17.0 (8u232-b09; 11.0.5; 13.0.1); 0.18.1 (8u242-b08; 11.0.6; 13.0.2)
- GraalVM CE (currently, for Java Engine and Maven only): 19.0.0; 19.1.1; 19.2.1; 19.3.1
- Liberica JDK: 8.0.242; 11.0.6; 13.0.2
- Oracle JDK Dev: 7.0_79; 8.0_202; 11.0.2
- Oracle OpenJDK: 7.0.231; 8.0.222; 11.0.4; 13.0.2; 14
- Zulu Community: 7.0.252; 8.0.242; 11.35.15; 13.28.11
Note: The 6th, 9th, 10th, and 12th Java versions cannot be created within the new environments anymore due to end-of-life of the respective releases. Herewith, the already existing instances remain fully operable (including redeploy, cloning, horizontal scaling).
You can choose the preferred version while creating an environment, or change it later. These procedures are described in the Java Versions document.
Java Application Servers #
Within a wide variety of software stacks, CirrusGrid supports a pure Java Engine and the following Java application servers:
Read the documentation below for additional information on Java servers specifics within the CirrusGrid PaaS:
Java Environment Creation #
To host a Java application, you need to create an appropriate environment. Just open the topology wizard in your CirrusGrid dashboard, navigate to the Java language tab, pick the desired Java application server, databases, and other stacks. If needed, customize settings (such as cloudlets, disk space, region, etc.) and click Create.
All added servers represent fully isolated containers, located on different hosts for more availability, while isolation eliminates the risks of interfering with one another. You can attach the Public IP address to any of these servers for being accessed directly. Otherwise (i.e. if using default settings), the incoming requests sent to your application will be proxied by Shared Load Balancer.
Java Application Deployment #
After environment creation, you can deploy your Java application. Herewith, Jelastic PaaS fully automates the deployment process allowing you to get the project up and running effortlessly.
The following deployment methods are supported:
- via the application archive – .war, .zip, .jar, and .ear archives
- from the GIT/SVN remote repository, using the Maven build node
- using plugins – Ant Task, Eclipse, IntelliJ IDEA, Maven and NetBeans
You can read the appropriate documents to learn more about the deployment of the Java applications:
There are also separate instructions for working with Bitbucket GIT repositories and managing your projects via Gitblit and WebDAV.
Domains Management #
You can bind a custom domain name to your application’s URL and use it instead of the default environment domain:
- CNAME redirect if using Shared Load Balancer; is recommended for dev and test environments
- DNS A Record if using public IP; can handle high traffic load and is suitable for production environments
Also, with the help of the swapping domains feature or the SwapExtIps API/CLI method, you can upgrade and modify your application with zero downtime (i.e. your users won’t notice any interruption).
Note: CirrusGrid PaaS allows using multiple domains within a single environment to increase its usability, efficiency, and scalability, while simultaneously saving your costs by avoiding a need to set up separate instances for different apps. Check the appropriate examples below:
Automatic Vertical Scaling #
CirrusGrid PaaS dynamically provides the number of cloudlets (i.e. RAM and CPU resources), which are required by your application to handle the current load. Just specify the maximum limit, and everything else will be performed by the platform automatically – no manual intervention required. This feature is called automatic vertical scaling and ensures that you never overpay for unused capacities without experiencing a resource shortage.
To set or change the vertical scaling limit, just use the appropriate slider within the topology wizard:
Horizontal Scaling: Manual and Automatic #
In case your application becomes highly popular and a single node is not enough, feel free to scale it horizontally. Herewith, to ensure even more reliability and high-availability, all the newly added nodes are created on the different hardware nodes.
Click the +/- buttons within the Horizontal Scaling section of the topology wizard to set the required number of nodes (load balancer will be added automatically).
Also, you can adjust the preferred scaling mode:
- Stateless – simultaneously creates all new nodes from the base image template
- Stateful – sequentially copies file system of the master container into the new nodes
Horizontal scaling can be performed not only manually but also automatically based on the current load on the node, which is monitored through the tunable triggers.
Within the Settings > Monitoring > Auto Horizontal Scaling section, you can adjust the scaling conditions due to your needs, i.e. the lower/upper limit (percentage) for the specified resource type during a certain period. If the load stays out of the set limits, the nodes removal/addition process will be called automatically.
There are five different types of resources, which are monitored by triggers:
- CPU
- Memory (RAM)
- Network
- Disk I/O
- Disk IOPS
The trigger starts monitoring the resource consumption immediately after addition, and when the usage level exceeds the stated limit, a trigger becomes invoked. Subsequently, if the load persists for a defined period, it will be executed, adjusting the number of nodes.
Java Clusterization #
The embedded application server clustering feature provides replication of sessions between pairs of nodes and eliminates the necessity of additional software or Memcached usage for increasing application availability.
CirrusGrid provides automated session replication between Tomcat and TomEE servers with the help of multicast to gain web application high availability within the Java cluster. This High-Availability feature can be enabled in the wizard during environment creation or topology tuning.
Herewith, the newer modern Auto-Clustering solution is provided for the GlassFish, Payara, and WildFly stacks:
Database Connection to Java Application #
A set of scalable and fully manageable database servers can be easily installed within the Java environment. To establish the connection, adjust your application following the instruction based on the database you require:
CirrusGrid PaaS provides high-quality Java hosting with improved availability, redundancy, and scalability strengths for your application.
Explore the benefits of Java cloud platform with a free trial at https://www.qloud.my/
