A question has already been posted here at Spring-boot jersey maven failed to run war file and its author has found a work around to get the Spring-boot app to run, but it's still an annoying bug.
Basically, when running a Spring-boot app with a Jersey configuration using ResourceConfig's public final ResourceConfig packages(final String... packages) with java -jar command or using Unix style ./app.jar start the following exception occurs:
java.io.FileNotFoundException:/path/app-1.0-SNAPSHOT.war!/WEB-INF/classes (No such file or directory)
Caused by: org.glassfish.jersey.server.internal.scanning.ResourceFinderException:
This does not occur when the app is ran with mvn spring-boor:run.
Not using packages("com.company.app.rest") is a work around, but it's a pain not to be able to have Jersey scan a base package.
Is this listed as a bug by the Spring-boot team?
I think it is but for some reason nobody complained hard enough. Can you please share your issue on #3413?
Related
When packaging our app with mvn package everything works fine. Then when we start our app with java -jar target\quarkus-app\quarkus-run.jar the app silently crashes. While debugging we found that it crashes while parsing an xml InputStream. It happens while initialising some classes.
This is the stacktrace that we had to dig out ourselves:
Exception occurred in target VM: Provider for javax.xml.parsers.DocumentBuilderFactory cannot be found
javax.xml.parsers.FactoryConfigurationError: Provider for javax.xml.parsers.DocumentBuilderFactory cannot be found
at javax.xml.parsers.DocumentBuilderFactory.newInstance(Unknown Source)
at org.optaplanner.core.impl.io.jaxb.GenericJaxbIO.parseXml(GenericJaxbIO.java:209)
at org.optaplanner.core.impl.io.jaxb.SolverConfigIO.read(SolverConfigIO.java:15)
at org.optaplanner.core.config.solver.SolverConfig.createFromXmlReader(SolverConfig.java:199)
at org.optaplanner.core.config.solver.SolverConfig.createFromXmlInputStream(SolverConfig.java:173)
at org.optaplanner.core.config.solver.SolverConfig.createFromXmlInputStream(SolverConfig.java:160)
When packaging the app in an uberjar this problem does not occur. Same when using dev.
We use graalvm-ce-java17-22.2.0, together with the 2.11.2.Final version of quarkus and the 8.29.0.Final version of optaplanner.
We tried to verify that there aren't any xml exclusion in the dependencies. Also we checked if quarkus and the quarkus maven-compiler-plugin are of the same version. Also we looked into the compiled jarfiles, if the xml we want to read is present. If it wouldn't be present, the code would crash even earlier. The class javax.xml.parsers.DocumentBuilderFactory is not listed in the quarkus-app-dependencies.txt
Adding the quarkus-optaplanner extension helped to identify the logger issue. So the problem with the silent crash is resolved. Adding quarkus-jaxp to the dependencies gets rid of the FactoryConfigurationError and everything works as expected.
I'm trying to get restart working with Spring Boot DevTools. I have been following the instructions provided here: https://docs.spring.io/spring-boot/docs/current/reference/html/using-boot-devtools.html
I am using gradle and included this in my build.gradle file:
bootJar {
excludeDevtools = false
}
I create the jar file and run the jar file:
java -jar app.jar
I am able to connect to the running application through Intellij. When I make a change I can see in the Intellij console that the updated classes are uploaded to the running process. And in the logs of the running process, I see the process attempts to restart. However, the process quits and spits out this log:
org.springframework.beans.factory.BeanDefinitionStoreException: Failed to parse configuration class [com.example.BootApplication]; nested exception is java.io.FileNotFoundException: class path resource [com/example/ExampleService.class] cannot be opened because it does not exist.
ExampleService is the class I modified.
I attempted to google the heck out of this, but could find nothing. I looked at many tutorials online, but could find nothing.
Has anyone encountered this or has anything I can try, would be much appreciated.
The issue is based on how compilation of a file works.
When you compile a file, it first deletes the already compiled file and then adds a new one. During this process, file system watcher consider it as two different updates (deletion of file and addition of a new file) if the poll time is too low. And deletion of file triggers deletion of file from the remote application and tries to restart the application without the file you changed and therefore, you get this error.
I was facing the similar problem but solved it by adding following in application.properties
spring.devtools.restart.pollInterval=10s
You can change the pollInterval that suits you
I'm working on a springboot 1.5.1 application that I'm trying to load a WSDL included in my resources directory in the wsdl directory. Depending on where my application executes I'm getting different results (command line, intellij, cloud foundry) and I can't seem to get all three to work at the same time.
I've tried several ways of locating the resource:
From prior to the migration to springboot we had this (worked in IntelliJ but not java -jar myboot.jar):
this.getClass().getResource("/wsdl/my.wsdl");
I switched it to the typically more correct version and got it to work in IntelliJ and java -jar but not Cloud Foundry:
this.getClass().getClassLoader().getResource("/wsdl/my.wsdl");
I switched it to use the Spring Resource Loader version and it worked in IntelliJ and CloudFoundry but not java -jar:
#Value("classpath:/wsdl/my.wsdl")
private Resource wsdlResource;
wsdlResource.getURL();
On the command line what I've noticed is that it seems to be thinking that BOOT-INF/classes is a JAR file (Note the ! after classes):
Caused by: javax.xml.ws.WebServiceException: Failed to access the WSDL at: jar:file:/C:/dev/redacted/build/libs/redacted.jar!/BOOT-INF/classes!/wsdl/my.wsdl. It failed with:
JAR entry BOOT-INF/classes!/wsdl/my.wsdl not found in C:\dev\redacted\build\libs\redacted.jar.
From looking at IntelliJ's URL, it's referring to the actual source folder which explains why it seems to always work.
What is causing this and how might I universally load these class path resources successfully with springboot?
I created a dummy Maven project that wraps CMU Sphinx4 in a servlet. The dependencies are sphinx4-core:1.0-SNAPSHOT and sphinx4-data:1.0-SNAPSHOT. When running locally (Eclipse) it works great, but then I deploy it to AWS Elastic Beanstalk and I get the following exception:
Property exception component:'null' property:'null' - java.io.FileNotFoundException: JAR entry edu/cmu/sphinx/models/en-us/en-us//means not found in /var/lib/tomcat8/webapps/ROOT/WEB-INF/lib/sphinx4-data-1.0-SNAPSHOT.jar
edu.cmu.sphinx.util.props.PropertyException: java.io.FileNotFoundException: JAR entry edu/cmu/sphinx/models/en-us/en-us//means not found in /var/lib/tomcat8/webapps/ROOT/WEB-INF/lib/sphinx4-data-1.0-SNAPSHOT.jar
Caused by: java.io.FileNotFoundException: JAR entry edu/cmu/sphinx/models/en-us/en-us//means not found in /var/lib/tomcat8/webapps/ROOT/WEB-INF/lib/sphinx4-data-1.0-SNAPSHOT.jar
I suspect it's that double slash in there but I cannot explain why it doesn't happen on my dev machine (OSX).
thanks!
Edit:
My dev machine: OSX running Eclipse Luna with AWS plugin. I am using a J2EE project with a single servlet, and getting Sphinx JARs (core & data) through Maven. When I run the server in my dev machine (port 8080) I feed it a WAV file through a POST message, the file gets loaded correctly and I get the analysis as a textual response.
My prod machine: AWS Elastic Beanstalk. I deploy using AWS plugin. When I comment out the Sphinx part of the code, the server works (i.e. gets the file, and returns it to me, etc.)
If the problem is not the double slash, then the JARs are not being cooked right. This would make sense since the file it's looking for is not a class so the JAR cooking part of Maven should be told how to include resource files in the JAR it makes when it deploys.
Any idea how to do this ?
thanks again!
Double slash seems to be a problem on some JVMs, on other JVMs it works.
I've just committed an update to avoid double slashes. Please update sphinx4 and try again.
Turns out the problem was I was developing (OSX) on Java 7 and deploying to AWS to Tomcat/Java 8. Once I deployed to Tomcat/Java 7, everything worked as expected.
I am getting weird errors on Heroku cedar using hibernate-memcached 1.3:
Caught CNFE decoding 438 bytes of data java.lang.ClassNotFoundException: org.hibernate.cache.entry.CacheEntry
Somehow hibernate-memcached 1.3 is failing to work with spymemcached 2.8.1 and up.
It does not matter if i explicitly list spymemcached 2.7.3 in my pom.xml or i set spymemcached scope as provided - i always get CNFE
I was able to replicate this issue locally when i upgraded to spymemcached 2.8.1
Thanks.
If you are using webapp-runner.jar to run your app, then yes, you're right: spymemcached class files are contained in the jar file to support caching of session data.
According to webapp-runner this is true for version 7.0.29.1 or newer.
The most recent version that does not contain spymemcached classes I was able to find, though, is 7.0.22.1.
I suspect that you can work around your problem by downgrading webapp-runner to version 7.0.22.1.
At least that worked for me when I ran into similar problems, getting exceptions like this:
org.springframework.beans.factory.CannotLoadBeanClassException:
Error loading class [net.spy.memcached.spring.MemcachedClientFactoryBean]
for bean with name 'memcachedClient' defined in class path resource [memcached-context.xml]:
problem with class file or dependent class;
nested exception is java.lang.NoClassDefFoundError: org/springframework/beans/factory/FactoryBean
These problems only occurred when I ran my war file with webapp-runner.jar. Running it from within Eclipse or with a local plain vanilla Tomcat 7 worked just fine.