I want to access `wlp.password.encryption.key' variable programmatically. Is it possible?
It is possible, but only via an SPI (meaning only in a user or product extension, not in an application).
http://www.ibm.com/support/knowledgecenter/SSHR6W_9.0.0/com.ibm.websphere.javadoc.liberty.doc/com.ibm.websphere.appserver.spi.kernel.service_1.5-javadoc/com/ibm/wsspi/kernel/service/location/VariableRegistry.html
The VariableRegistry will be a discoverable OSGi Service (choose your favorite discovery mechanism, I recommend DS).
Related
Is it possible to have both profiles on Liberty 17.0.0.2.
By both, I mean WAS traditional full profile and Liberty profile on WAS Liberty 17.0.0.2.
Technically, no. The only thing that uses that versioning is Liberty based products and the liberty runtime (profile) itself. These do not include anything with the full profile / traditional WAS.
In more practical terms, products like "WebSphere Application Server Base" or "WebSphere Application Server ND" 9.0.0.4 includes the 17.0.0.2 liberty profile as well as the 9.0.0.4 "full profile".
I want to make an application like JConsole. Is it possible? If yes, what are the changes need to done at JVM level? I am planning to use Spring-Boot. As per my knowledge, JMX is enabled by default. Do I need to configure anything extra in my Spring-Boot app in order to access the JMXBeans which are exposed by default?
Here I'm not trying to expose any MBean instead I'm trying to access those beans which are already exposed by JVM. How to achieve it?
JConsole is a JMX compliant monitoring and management application. The architecture is quite simple. It's a client-server architecture. Where the client is the Remote application (Example JConsole or the one that you want to build) and the server is the JMX Agent. In your case, you want to build your own client which is possible.
I want to make an application like JConsole. Is it possible?
Yes, it is possible.
If yes, what are the changes need to done at JVM level?
What do you mean by changes at JVM level? You are simply creating a client application that connects to the server (JMX Agent) using a certain protocol. Remote Method Invocation (RMI) is the protocol used by JConsole for the connection to the JMX Agent. If you want to use RMI for communication, you don't have to do anything on the server side. But if you want to use some other protocol for communication, you can define your own Protocol adapter.
As per my knowledge, JMX is enabled by default.
As of Java SE 6 it is. But you can only monitor it locally. For connection from a remote machine, you need to define an RMI port to start listening for incoming connections.
Here I'm not trying to expose any MBean instead I'm trying to access those beans which are already exposed by JVM. How to achieve it?
Please check out the example from this link - Mimicking Out-of-the-Box Management Using the JMX Remote API. It shows you how to create a simple client application that connects to a remote JMX agent and access the MBeans. This should guide you in the right direction.
In a full WAS it is possible to stop and restart a JMS activation specification, either by using the administrative console or by using the J2CMessageEndpoint MBean. My question is if the J2CMessageEndpoint MBean will be available in Liberty Profile, making it possible to pause an activation specification?
If not, is there any other way that I can activate/deactivate the message endpoint in Liberty Profile?
There isn't a direct equivalent, but you can change the server xml so the activation specification binding no longer matches. When the server picks up the new configuration, which is dynamic by default, the binding won't match and the mdb will stop.
My organization is using Websphere Application Server with RAD. My unit is developing Web Services that take data from a consumer, and often pass the data through other Web Services.
Currently, we have two ways of defining time-out's for data to be returned from called services:
Using the Spring framework
Websphere profile configuration
If we need to update the values using Spring we need to re-deploy our service. If we define them as JVM properties we need to do a JVM configuration change and 'restart/recycle/stop and start' the application.
My co-worker had it suggested to him that there may be some better solutions to this problem. I'm wondering if this is true, and if so what they are?
edit:
One option we might be considering is "Application Policy Sets". We'd like to know if this is a good alternative method:
can we define Spring to leverage policy sets
can we define multiple policy sets for an application when timeout values for different services called by our service are different?
Does this console change take affect at run-time, or do we need to recycle?
I suggest you use JAX-WS policy sets and bindings on WAS level, because it allows you to:
define policy set for each web service separately
in policy set define web service parameters, from HTTP transport parameters to WS-Security policies such as UsernameToken, digital signature and encryption
web services remain unchanged, the code remains the same
there is only one place where you change it: in WAS console, application is not changed
The only drawback here is that web services should be generated from RAD (you can use top-down or bottom-up approach).
Please see more details in this excellent 3-part DeveloperWorks tutorial
Can a Worklight Server be deployed to a WebSphere application server which also runs other non-Worklight .ear applications? Or does Worklight need its own separate instance of WAS?
Just like you can deploy multiple instances of Worklight (v6 and above) projects - multiple .war files to the same WAS application server, there should not be issues to deploy it to an application server running other services.
That said, possible issues to consider:
When deploying a Worklight project, you will want to enable "application security"
(in the WAS admin console, Security > Global Security). If there are some other web
applications for which application security is undesired, you need a different WAS server
instance.
Setting up, enabling and migrating security
The list of users that can use the web applications are configured through LDAP or
"federated repositories", or similar. If, for Worklight, you need to use a completely
different set of user logins than for the other web applications, then you need to use
multiple "security domains".
Configuring multiple security domains
The machine hosting the application server will probably need memory upgrades...
Deploying the Enterprise Archive (EAR) Using the WebSphere Admin Console
Probably also need to make clear seperation where required:
IBM WebSphere Developer Technical Journal: Co-hosting multiple versions of J2EE applications
Worklight is itself an application running inside a web container, whether that be Tomcat, WAS Liberty, or full WAS. It's essentially a layer running underneath the container to handle requests for Worklight applications, fielding their context root requests. If you create the WAR file for your Worklight app and extract out the deployment descriptor you'll find all the necessary filters and listeners that most other apps would have.
Things like adapters and wlapps are "installed" to this underlying layer, and are merely extracted and stored as whatever was packaged with them, such as the JS and CSS you used to make you app. In fact, with a standard Liberty install you can typically find your adapters in plain sight at (for the WL5.0.6 instance I have handy, it's different for WL6):
/opt/IBM/Worklight/server/wlp/usr/servers/worklightServer/worklight.home/worklight/data/export/adapters
So, in addition to what Idan has said, I also present you with the following docs (assuming WL6)
Overview of the Worklight Server installation process
Given my own experience, you should be perfectly able to install other EAR and WAR files to your existing WAS instance, just make sure your context roots are unique, as always ;)
I also second the memory considerations.