We have to develop and maintain many Java web based applications (for the same company) of different sizes, scopes and life-spans. Some of them are huge and other ones are just simple pages that may live only a few months (or days), some are already implemented and need refactoring.
There have one thing in common though, they need access to (almost) the same information.
Problem
Due to the complexity of the data the company handles, we have to deal with many different sources, some of them inherited from the ancient times. Our domain objects may be mapped across many of those sources. As an example, a Contract domain object is mapped to our main database but its related (physical) files are stored in a document server, and the activity related to it is stored in a NoSQL database. Therefore, adding, removing, searching any of these objects involves many internal operations.
Our data sources are (although it could be any):
AS400 (using DB2 as a database)
Documentum document manager
Mongo DB
External web services
Other legacy sources
We normally use Glassfish as the application server and maven as our build tool.
Goal
Our goal is to create a business layer or library that all of our applications can access and it is:
Compact
Consistant
Easy to use
Easy to maintain
Accessible from many different clients
What we have found so far
We have been struggling for weeks and still we cannot find anything fully satisfactory. Some solutions:
Pack all the business logic in one or more jars: Very easy to share, but all the applications will have to contain all the jar dependencies and configuration files and take care of security, caching and other stuff. Difficult to maintain (we have to update the jars for every project when there are changes).
Create an Ejb project containing all the logic and access it remotely: Easy to maintain, security, caching and configuration only implemented once. We are afraid of the penalty of the remote calls. As we have noticed in our research, it seems to be a bad practice (we don't have much experience with ejbs).
Create an Ear project with everything inside and use local access: Well, this is faster than the remote version but it is a hell to maintain.
Go for OSGI: We are a bit afraid of this one since it is not as popular as Ejb and we have never used it seriously.
Is there a common practice for this kind of problem?
Many thanks!
I would not recommend put all logic into 1 EAR project and use local access. If you have a lot of code in the one place, it will be harder to maintain, test, deploy etc.
I would create mutlti-module maven project with common dependencies. One of the dependency - service with business logic and DAO access, which will expose API. With Maven project you can easy control version of the POM files. Different projects may work with different version of common service. Maven will handle version control for you. However it's require some configuration and implementation efforts.
Another option mentioned by you - standalone EAR with remote EJBs should work fine as well. Do not worry about performance and number of remote calls, unless you have heavy load. Simply cache remote EJB stubs on client to avoid unnecessary JNDI lookup.
Personally I prefer first option with shared dependency managed by Maven. It's clear and easy to maintain, easy to manage versions, deploy, configure. With Maven you don't need to change jar file manually for every project, you can simply use tools like Nexus
Related
What is the best way to store parameters and data for an EE7 application. I have to provide the web applications with information like a member fee or similar data (which may/can be altered several times in a year). The owner of the application should also have a central place where these data are stored and an application to change them.
Thanks in advance for any input
Franz
This is one question we are currently struggling with as we re-architect some of our back-end systems here, and I do agree with the comment from #JB Nizet that it should be stored on the database, however I will try to add some additional considerations and options to help you make the decision that is right for you. The right option will depend on a few factors though.
If you are delivering source code and automation to build and deploy your software, the configuration can be stored in a source code repository (i.e. as YAML or XML) and bundled with your deployable during the build process. This is a bit archaic but certainly widely adopted practice and works well, for the most part.
If you are delivering deployable binaries, you have a couple of options.
First one is to have a predetermined place in the file system where your application will look for an "override" configuration file (i.e. home directory of the user used to run your application server). This way you can have your binary deployable file completely separate from your configuration, but you will still need to build some sort of automation and version control for that configuration file so that your customer can roll back versions if/when necessary. This can also be one or many configuration files (i.e. separate files for your app server, vs. the application itself).
The option we are contemplating currently is having a configuration database where all of our applications can query for their own configuration. This can either be a very simple or complex solution depending on your particular needs - for us these are internal applications and we manage the entire lifecycles ourselves, but we have a need to have a central repository since we have tens of services and applications running with a good number of common configuration keys, and updating these keys independently can be error prone.
We are looking at a few different solutions, but I would certainly not store the configuration in our main database as: 1) I don't think SQL is best repository for configuration, 2) I believe we can get better performance from NoSQL databases which can be critical if you need to load some of those configuration keys for every request.
MongoDB and CouchDB both come to mind as good candidates for storing the our configuration keys if you need clearly defined hierarchy for you options, whereas Redis or Memcached are great options if you just need a key-value storage for your configuration (faster than document based too). We will also likely build a small app to help up configure and version the configuration and push changes to existing/active servers, but we haven't spec'd out all the requirements for that.
There are also some OSS solutions that may work for you, although some of them add too much complexity for what we are trying to achieve at this point. If you are using springframework, take a look at the Spring Cloud Config Project, it is very interesting and worth looking into.
This is a very interesting discussion and I am very willing to continue it if you have more questions on how to achieve distributed configurations. Food for thought, here are some of my personal must haves and nice to haves for our new configuration architecture design:
Global configuration per environment (dev,staging,prod)
App specific configuration per environment (dev,staging,prod)
Auto-discovery (auto environment selection depending on requestor)
Access control and versioning
Ability to push updates live to different services
Roger,thanks a lot. Do you have an example for the version predetermined place in the file system"predetermined place in the file system"? Does it make sense to use a singleton which reads the configuration file (using Startup annotation) and provides then the configuration data? But this does not support a dynamic solution.kind regards Franz
I have a Java web application using Spring, Hibernate and Wicket connecting to a MySQL database that I'd like to refactor and separate into several applications. I started by using Maven's multi-module system but in reality each of the applications would have its own release cycle, so I've ditched that effort now and I'm looking at creating individual projects for each of them. They will all continue to connect to the same database so I was going to move the model classes into a project of their own which can be used as a dependency.
I have a few questions regarding this setup:
Is moving the model classes to their own project a typical solution to the multiple apps/single database problem, or is there another way?
Is there a nice way of ensuring all the applications are using the same version of the model dependency?
Should I also include any base daos and services in this core project that each application could use or extend, or should I just include my GenericHibernateDao and let each application create its own daos and services? Obviously I will want to avoid changing this project as much as possible as it will require a new release of all the applications depending on it.
Is there any Hibernate related config I would need to change, such as connection pooling? Does it matter if each app has its own pool or should they share one? I'm not using caching at the moment, but I understand if I wanted to I would need a distributed cache?
How would I share application config such as db params, email host, sms gateway etc. between applications? Is there any way of defining them once somewhere to ensure they are all pointed at the same db?
Are there any other gotchas I may encounter further down the road with this setup, either with Maven or during deployment? Any tips or best practises I should follow?
These have been usual scenario with me, what I have usually done is..
- DAOs, Conn. Pool Management, Fail over related code can be managed by writing separate module [jar]
- You can then use this module in components as you have mentioned.
With this you will have separate connection pool for each of your component.
While working on a modular system architecture for an enterprise application I run into some problems with database initialization. We have a core library that provides base entities and base configuration. On top of this core several modules are build. They are pluggable and can have their own entities and configuration. Some characteristics:
Configuration, like system properties, resourcebundles, etc, are all stored in the database.
JPA is used to make the system database independent.
System runs on Java SE
Every module can bring its own tables, but they can also require to populate the core property table, or the core resourcebundle table. So somehow we need some mechanisme to run a DDL and DML initialization for the database. Some options:
Create simple sql scripts. Disadvantage is that they must be database independent and perhaps this is not the most developer friendly. Unless we can generate them with some DB diff tool?
Use Java classes to initialize via JPQL?
Store configuration in files? This avoids a lot (but not all) of configuration DML.
Use some tool like liquibase?
What would be the best practice for this (or a similar) problem?
Use a database for store all configuration data is the best option. Many products, such as WebSphere Portal or Liferay use a database to store the configuration data for each portlet or even for theme. Don't forget to include those that are used as part of an SOA and Business Rules.
Therefore, the use of SQL scripts is also the best choice. However, if you require very specific features of SQL, you may need to create several versions of same script for each database management system.
I am currently in an project that has the same idea of modules that add functionality to a core system.
Generally we are using maven and multiple src folders as well as maven profiles and different builds to be able to generate a deployable with different modules. (we do not have the necessity to push out single modules and install them later on - this might be different in your project. We just build different versions with different modules.)
Anyway, for the DB we are using liquibase. Firstly to manage the DB and the changes done to it. But also (and this might be helpful to you) to include/generate another SQL script that adds tables for the modules.
Each module has its own changeset-file that includes everything that is necessary for that module (also in different versions as the modules evolve through time). These can then be applied or not.
So, I think liquibase could also be useful in your case (even though it's main purpose is to manage DB changes).
This may be a very rudimentary question, but please help me out if this is well-known and has been solved elsewhere.
I have a multi-war setup (all maven modules) say kilo-webapp1 and kilo-webapp2 as two WARs that I need to deploy on a Tomcat instance. These two webapps both use services from a common service jar, say kilo-common-services.jar. The kilo-common-services.jar has its own spring context that is loaded by the users of the jar viz. kilo-webapp1 and kilo-webapp2 in this case. It so happens that the initialization of the services in kilo-common-services takes a long time and hence I want it to happen only once (to ensure that the time it takes to bring up the instance is not very high) which also helps me to use it as a second level cache that it kept current in the JVM instance. To do this, we resorted to the following steps:
Modify the catalina.properties of CATALINA_BASE in tomcat to have shared.loader as ${catalina.base}/shared/lib
Copied the kilo-common-services.jar and all of its dependent jars to the CATALINA_BASE/shared/lib. [Manual step]
Copy spring related jars to the CATALINA_BASE/shared/lib location [Manual step]
Created a beanRefContext.xml file in kilo-common-services.jar. Define a new ClassPathXmlApplicationContext here, where the constructor was provided with the location to the spring context file for the common services.
Noted the dependency scope of kilo-common-services.jar and every other dependency (like Spring related jars) as provided in the kilo-webapp1 and kilo-webapp2 pom files. For Spring this is needed to ensure that the classpath scanning actions are not triggered twice. Also this causes different ClassCastExceptions (for log4j lets's say) if not excluded via the provided scope.
web.xml for kilo-webapp1 and kilo-webapp2 indicated that the parentContext for them is the servicesContext defined in kilo-common-services.jar.
I was able to verify that only one instance of the services of kilo-common-services exist, but the setup as you might have imagined is painful. If someone has best practices about such a setup in an IDE like Eclipse, would really appreciate it. My problems are as below:
#2 is becoming a challenge. I am currently running mvn dependency:copy-dependencies on kilo-common-services to copy dependent jars from target/dependency to the shared/lib which is a woefully manual step. Time and again, I forget to regenerate dependencies and have to do a redeploy again.
#3 is also not straight-forward as time and again there are newer common dependencies and we always have to remember to copy it to shared lib to avoid ClassCastExceptions
#5 is again a maintenance nightmare.
Also as time progresses, there will more such disparate common jars that need to be shared and it would involve pain for each of those jars. Feel free to critique the setup and propose a better one in its place that may be easy to use (from an IDE as well). Would be happy to provide any other details.
Thanks in advance!
The problem is that your architecture is broken (and that's why you're struggling with the solution). You have two solutions:
1) If you want to share a service that takes a long time (to initialise) between two war applications, make that a separate service completely and access it via rest or any kind of remoting.
2) Merge both webapps into one.
Having the common library is the shared lib folder is going to bring you lots of headaches, and you'll end up rolling it back.
My (personal) approach would be to merge both applications, but keep the packages separate enough and have separate spring configurations. In this way, at least you still keep the logic separation of both webapps.
Also since both run on the same container, there's little gain from having 2 separate wars (unless you're planning to move them to different containers very soon).
About the IDE, you can use the maven-cargo-plugin to start up a tomcat with several web applications with (almost) any configuration you want.
We are developing restful soa, with spring and tomcat and utilizing Domain Driven Design (well thats the plan anyway). There is migrationProject and a initial basic search service. Two separate WAR files, with two separate POMs. Both utilize the same Domain objects.
So I will have separate project that will be just the DomainObjects I will wrap them up into a jar, and then using maven and/or jenkins it will deploy automatically (whenever I configure (for example when pushed to a specific repository).
Having two copies of the same jar, sounds like a much worse idea to me. Its not your architecture that is broken, its your deployment and development process thats needs improvement, imho.
(my kind of related question).
Our long term plan is to have one project as the restful interface, with multiple Controllers that have service classes and repositories injected into them from their dependencies.
I have a large scale project I am working on at the moment using Eclipse. Normally, as a one man team, these problems would not be an issue, but as our team is not one person we need to be able to break up pieces of the project to be worked on by certain team members.
In simplicity, let's say I have two layers to be separated apart:
1. Each DAO is a separate Java project, to be worked upon individually
2. The web-tier service layer contains all of our service endpoints and must be able to reference all of the DAOs. This layer runs on Tomcat as a dynamic web project, and utilizes Adobe LiveCycle Data Services as the piece that handles creation and management of endpoints.
Now, the issue we are running into is that when we create a DAO and unit test it individually it runs great. But when we reference it into our service project and try to run it we begin to get all kinds of issues related to the fact that we have two different versions of certain jars referenced in and as such we begin to have errors when running the server.
As a result, we know we can solve the issue by pulling the problem jars and ensuring that this is not an issue again in the future, but as I said before this is a large scale project with multiple people working on it and we don't want to be spending our time weeding out dependency issues when under the gun.
We are looking for recommendations on where to proceed for alternative solutions? Our team is new to JavaEE and as such we don't have much of a bearing on what we can use to tie everything together in it, or if it is a viable solution. Should we be looking at turning our DAOs into EJBs and deploying them in an EAR library? If so, where would our service layer lie, and would the service layer be able to reference the DAO classes since the EJB maintains it's own classpath (from what we have read?) Are we looking down the wrong path, or are we completely wrong in our current understanding of JavaEE?
Any assistance is greatly appreciated. We are still in the framework stage of this project and we want to be sure that we will be able to maintain it in the long run.
I second the Maven recommendation. That can add all sorts of sanity to your project structure.
Maven can even generate Eclipse workspaces via mvn eclipse:eclipse
An important clarification on the EJBs note. As of ava EE 6 is you no longer need to separate EJBs from Servlets and can use them together in the very same jar in the war file.
So understand from that that using EJBs or not no longer has any impact on packaging or classloaders as it once did. These are now separate decisions. EARs and classloader separation should now be viewed as a feature you might want to use should you want classloader separation and the complexity it brings. Most applications simply do not need that and are more than fine with just a war file containing servlets, ejbs, jpa entities, cdi beans, jaxrs services and whatever else you need. You are free to decide how you want to separate them or if you want to bother separating them at all.
EJBs do make great DAOs due to transaction management, something you don't get from plain Tomcat but can be made available in Tomcat via TomEE and works fine in Eclipse. You should consider EJBs for that reason, not for dependency reasons.
Side note, as you're new to Java EE, you might find this helpful:
http://openejb.apache.org/examples-trunk/index.html
In order to have things organized when working with Java EE in teams of 1+ people I could suggest:
Use Maven to manage your build process and library dependencies.
Maven has a small learning curve, but once you grasp it you will be grateful. By using Maven you no longer depends on Eclipse to manage your classpath.
A thing about it that I think is really helpful when working in teams is the install feature. Suppose you are woking on the version 1.0 of an EJB module, say core-ejb-module-1.0, and you've got it to a stable state and want everyone working in the project to refer to it from now on.
You then run a maven command like this on it: mvn clean package install
Maven will clean this module, compile it, run tests, create the jar and then install it to a repository that you define. Could be any computer in your company.
Now you may tell the guys working on other projects to update this dependency version on their .pom file and in the next build they run, before compiling, maven will download this library and then use it. Really neat. No more classpath hell.
(There are other ways to always automatically refer to the latest library as stated in this post, but there are some caveats. Anyway it's just an example.)
Use JPA/EJB instead of DAO Pattern.
Some people say DAO meaning any sort of data access, others really mean that they use the DAO Pattern to access objects. If that is your case, you no longer need to use it when using JPA. (At least for most common scenarios).
In my case, I have a generic EntityService which is capable of doing CRUD operations on any Entity and has a centralized query management. Then every EJB's that should perform database related operations may inject this guy and do its job.
As a suggestion, with Maven, you project could be organized as such:
core project structure
core (The pom root)
core-ejb-module (Includes all generic EJB's, like the EntityService for instance.)
core-jpa-module (Includes all JPA generic definitions, like Interfaces, MappedSuperclasses and such.)
core-jsf-module (Includes all JSF generic definitions, like abstract controllers, generic converters and wrappers for FacesContext, etc..)
Now that you have a core generic module setup, you could create:
an application structure
app (The pom root)
app-ear-module (Includes all other modules in this application. Shared jars goes in the ear /lib folder, so all other modules could reference to them.)
app-ejb-module-a (Includes EJB's for the business layer of your application. It uses the core-ejb-module)
app-ejb-module-b (You may have lots of ejb modules. You may even have a project that contains only ejb modules. Other apps will declare their dependency on them via Maven.)
app-jpa-module (Contains definitions for JPA Entities that represents you database tables. Depends on the core-jpa-module)
app-web-module (Holds the pages, Controllers and Converters for this application.)
I think you got the idea. Things tend to be loosely coupled and you may organize your projects as you like.
This is just a simple example to illustrate. I didn't explained a lot about Maven but if you're interested I think it will help you indeed.
I hope it gives you some ideas and may help you in any way.
[]'s
If you can run all the sub-components using the same set of dependencies, you may find it helpful to migrate to a Maven build.
With Maven, you can define a top-level project that manages all the 3rd party dependency versions in one place, so all modules are built, tested and deployed against the same library versions. You are also likely to find Maven a good fit for the multi-module approach you have adopted, as it ensures that a project is rebuilt correctly if one of its dependencies changes.
You would still be able to use dynamic web projects as before; Eclipse will automatically deploy the DAOs as part of the service project (IIRC you need to characterise the DAOs as utility modules).
If you do go down the EJB root, you are correct that each EAR will get its own class-loader, and can therefore use a varying set of dependencies. However, in your position I would tend to look at improving your dependency management first - it'll probably be cheaper and easier.