I'd like to create a Configuration object in OSGi, but one that won't be persisted, so it won't be there when the framework is restarted. Similar to START_TRANSIENT for bundles.
Some background: I've got an OSGi (Felix) based client side application, deployed over OBR. The configuration object I'm talking about effectively boots the application. That works fine, but sometimes the content has changed while the context was stopped. In that case, it boots the application as OSGi revives all bundles and adds all configuration options. Then I inject the correct configuration, the application stops and then restarts again.
So it does actually work, but the app starts twice, and I can't get access to the framework before it reconstructs its old state.
Any ideas?
As BJ said there is no standard support for this in the Configuration Admin spec.
However the Felix implementation supports two features which may help you. First, you can set the felix.cm.dir property which configures where the configadmin saves its internal state (which by default will be somewhere under the Framework storage directory). You could set this to a location that you control and then simply wipe it every time you start OSGi (you could also wipe out the entire OSGi Framework storage directory on every start... some people do this but it has wider consequences that what you asked for).
Second, if you need a bit more control, Felix ConfigAdmin supports customising its persistence with a PersistenceManager service. You could probably implement this and return empty/doesn't-exist for the particular pids that you want to control.
The OSGi Config Admin spec does not support this. I also do not know of a non-standard means either for any of the CM impls I am familiar with.
Ok, what I did in the end was the following:
I created a special really small 'boot' bundle, which I do not provision from OBR, instead, I install it from the classpath.
That bundle controls the configuration, and I use START_TRANSIENT the moment I really want to load that configuration.
Not exactly pretty, it gets the job done. I do think transient configuration would make sense to have in OSGi.
Related
I'm creating a Liferay 7.1 OSGi bundle, which has some external dependencies in it. In consideration of time, we opted to embed the external JAR in our OSGi Bundle. I've managed to create a bnd file, which includes all of the ElasticSearch dependencies, and put them on the bundle classpath. I've used the source-code from github (https://github.com/liferay/liferay-portal/blob/master/modules/apps/portal-search-elasticsearch6/portal-search-elasticsearch6-impl/build.gradle) and the bnd.bnd file, to check what's imported.
When activating the bundle, an exception is thrown:
The activate method has thrown an exception
java.util.ServiceConfigurationError: org.elasticsearch.common.xcontent.XContentBuilderExtension: Provider org.elasticsearch.common.xcontent.XContentElasticsearchExtension not a subtype
at java.util.ServiceLoader.fail(ServiceLoader.java:239)
at java.util.ServiceLoader.access$300(ServiceLoader.java:185)
at java.util.ServiceLoader$LazyIterator.nextService(ServiceLoader.java:376)
at java.util.ServiceLoader$LazyIterator.next(ServiceLoader.java:404)
at java.util.ServiceLoader$1.next(ServiceLoader.java:480)
at org.elasticsearch.common.xcontent.XContentBuilder.<clinit>(XContentBuilder.java:118)
at org.elasticsearch.common.settings.Setting.arrayToParsableString(Setting.java:1257)
The XContentBuilderExtension is from the elasticsearch-x-content-6.5.0.jar,
the XContentElasticsearchExtension class, is included in the elasticsearch-6.5.0.jar. Both are Included Resources, and have been put on the classpath.
The Activate-method initializes a TransportClient in my other jar, hence it happens on activation ;).
Edit:
I've noticed that this error does NOT occur when installing this the first time, or when the portal restarts. So it only occurs when I uninstall and reinstall the bundle. (This is functionality I really prefer to have!). Maybe a stupid thought.. But could it be that there is some 'hanging thread'? That the bundle is not correctly installed, or that the TransportClient still is alive? I'm checking this out. Any hints are welcome!
Edit 2:
I'm fearing this is an incompatibility between SPI and OSGi? I've checked: The High Level Rest Client has the same issue. (But then with another Extension). I'm going to try the Low-Level Rest Client. This should work, as there are minimal dependencies, I'm guessing. I'm still very curious on why the incompatibility is there. I'm certainly no expert on OSGi, neither on SPI. (Time to learn new stuff!)
Seems like a case where OSGi uses your bundle to solve a dependency from another bundle, probably one that used your bundle to solve a package when the system started.
Looking at the symptoms: it does not occur when booting or restarts. Also it is not a subtype.
When OSGi uses that bundle to solve a dependency, it will keep a copy around, even when you remove it. When the bundle comes back a package that was previously used by another bundle may still be around and you can have the situation where a class used has two version of itself, from different classloaders, meaning they are not the same class and therefore, not a subtype.
Expose only the necessary to minimize the effects of this. Import only if needs importing. If you are using Liferay Gradle configuration to include the bundle inside, stop - it's a terrible way to include as it exposes a lot. If using the bnd file to include a resource and create an entry for the adicional classpath location, do not expose if not necessary. If you have several bundles using one as dependency, make sure about the version they use and if the exchange objects from the problematic class, if they do, than extra care is required.
PS: you can include attributes when exporting and/or importing in order to be more specific and avoid using packages from the wrong origin.
You can have 2 elastic search connections inside one Java app and Liferay is by default not exposing the connection that it holds.
A way around it is to rebuild the Liferay ES connector. It's not a big deal because you don't need to change the code only the OSGi descriptor to expose more services.
I did it in one POC project and worked fine. The tricky thing is to rebuild the Liferay jar but that was explained by Pettry by his google like search blog posts. https://community.liferay.com/blogs/-/blogs/creating-a-google-like-search (it is a series but it's kind of hard to navigate in the new Liferay blogs but Google will probably help) Either way it is all nicely documented here https://github.com/peerkar/liferay-gsearch
the only thing then what needs to be done is to add org.elasticsearch.* in the bnd.bnd file in the export section. You will then be able to work with the native elastic API.
Our application are built on Spring boot, the app will be packaged to a war file and ran with java -jar xx.war -Dspring.profile=xxx. Generally the latest war package will served by a static web server like nginx.
Now we want to know if we can add auto-update for the application.
I have googled, and people suggested to use the Application server which support hot deployment, however we use spring boot as shown above.
I have thought to start a new thread once my application started, then check update and download the latest package. But I have to terminate the current application to start the new one since they use the same port, and if close the current app, the update thread will be terminated too.
So how to you handle this problem?
In my opinion that should be managed by some higher order dev-ops level orchestration system not by either the app nor its container. The decision to replace an app should not be at the dev-ops level and not the app level
One major advantage of spring-boot is the inversion of the traditional application-web-container to web-app model. As such the web container is usually (and best practice with Spring boot) built within the app itself. Hence it is fully self contained and crucially immutable. It therefore should not be the role of the app-web-container/web-app to replace either part-of or all-of itself.
Of course you can do whatever you like but you might find that the solution is not easy because it is not convention to do it in this way.
Rather than manually making configuration changes to OSGi components in Felix, it's good practice to create sling:OsgiConfig nodes in the JCR to make sure that the settings are version controlled, applied the same across environments, etc.
Similarly, I want to disable an out-of-the-box component on each environment. Is there any way to achieve this via configuration? Rather than going to system/console/components and disabling it there?
Anything I've read about this has mentioned making changes internally within the bundle, but since it's one that I don't own, I'd need some external configuration to do this.
You can do this with the ScrService, which is published by the SCR runtime bundle. API documentation is here.
Note that ScrService is not "standard", i.e. it doesn't come from the OSGi specification. However it is supported by Felix and Equinox and Knopflerfish, so it's pretty much a de facto standard. In fact this service is used by the Web Console when you go to system/console/components.
Try to use attribute policy=ConfigurationPolicy.REQUIRE on #Component.
Then you could prepare a set of packages with configurations for each environment as part of the build. This is also a good practice for having different OSGI configurations for each environment.
So for some of enironments you could just simply not provide a configuration for a particular component. Such component would not run - it would have unsatisfied status.
I believe I've seen this approach in AEM itself.
You can also create a filter to remove current configurations but it still would require to disable the Component at least once. And this solution will work only if this component has mentioned policy.
The other way around is to prepare a Service that would be responsible for disabling other components - it could be configurable. But it doesn't sound like a good solution to me.
I have a dynamic application that uses OSGi to load modular functionality at runtime. OSGi bundles contain the modular functionality and the application loads the bundles when they are needed. This approach works okay, but I would like a more granular solution. The bundles contain components controlled through Declarative Services. I'd like to be able to load a bundle, and only enable the components that are needed within the bundle. I've done research in this area, but cannot find a solution that I'm satisfied with. One approach was to create a "gatekeeper" component that is always enabled in the bundle and through the ComponentContext let it call enable and disable component. It's a viable solution, but I could not figure out a way for the "gatekeeper" to "know about" the other components in the bundle without hard coding the component names as properties in the "gatekeepers" SCR xml descriptor.
What I prefer is a way to load bundles and "know about" all components within the loaded bundles. Be able to determine what bundle the components are located in and what state they are currently in (similar to the equinox console command 'ls' that lists out all components). I would like to enable and disable the components when needed.
How does the console do this and how could I do this in an application?
Update:
#Neil Bartlett: Sorry for the delay. I had to move on to something else. Now I'm back on this issue. Really would appreciate any further assistance. My application is role based. I need to enable components based on the functionality they provide. The goal is for all role based components to initially be disabled. Upon role change, a role manager polls each component for its provided functionality and determines whether to load it. Each component will broadcast what functionality it provides (through a common service interface). ScrService will not allow me to enable an initially disabled service component. Having the components initially enabled and let ScrService disable them as soon as possible during application startup does not fit my needs.
Have a look at ScrService. Bothe equinox and felix has it.
However, components can be made to load lazily, i.e. only when needed by other components/bundles; but that is perhaps not what you want.
In your service description, mark the components as enabled, but requiring configuration information provided by the Configuration Management service. you then can write a CM plugin service (can't remember the exact term) that can publish and modify the configuration of your components. Services by default are identified by their name, which by default is their implementation class name. Configuration data is passed as a map, and it can be empty. DS will make the service available as soon as the CM provides a configuration.
I have a similar issue, but for a different purpose:
- I have apache file install and configuration admin service to configure my components externally with property files.
- I needed to make sure that some components get the config from the outer file and the only way I've found so far is that I mark my components with ConfigurationPolicy.REQUIRED.
- But that way my plugin projects doesn't run in eclipse (where there are no config files).
- The component.xml also contains a default development configuration so I'm okay with that, just my component doesn't start until there is a config data avalilable with configadmin. My components ar unsatisfied this way, until someone creates a configadmin entry.
- I figured out that if I create a osgi command line extender that sends empty configurations to service pid's they would start up with default values in component.xml files.
- I just came here to find a way to list all bundles
But I think this solution I use can also work with your setup and that's why I write this.
Just mark all your components with the configurtationpolicy.require and you can selectively start and stop them with adding and removing configurations with configadmin. This could be hard if you already use the configadmin for other purposes too, but it may be managable as a last resort.
We are using WebSphere 6.1 application server with default classloader delegation mode i.e. PARENT-FIRST. We think about changing it to PARENT-LAST to be able to choose our jsf implementation or our webservices stack.
As PARENT-FIRST is the default I wonder how many people switched to PARENT-LAST, and what was the reason to switch, and if your life became better since you switched :)
We have a lot of applications in production so I cannot just switch to see what happens, if we do it we will have a lot of testing so I’d like to have to some feedback if you have switched to PARENT-LAST.
Thanks
On projects that I'm assigned to, we actually do switching to PARENT-LAST for most of our applications. The reason for that is usually an app-specific implementation of something, or a need for app-specific property bundle that Websphere uses too (overriding the Websphere setup of commons-logging, for example).
If something breaks after the switch, it is usually because of somewhat wrong setup of the application that suddently starts to be used (while before the switch it was overriden by Websphere's resources).
Portlet applications (deployed on WebSphere Portal Server) always switch their configuration to parent last. In my experience it is always better to switch to parent last, especially if you are using commons logging. This is because WebSphere includes a truck load of stuff in its own classloaders which are often a different versions/configurations to the one that you want to use.
If you are doing it, I would recommend that you script up the deployment of the application because it can be one of those things that are missed when you do a deployment.