Javanica fallback not found when contained in super class

[ctoestreich]

We are using the Annotated methods on classes where a service extends another service

class WishlistService extends CartService { ... }

And then inside CartService there is a some methods like

 @HystrixCommand(commandProperties = [
            @HystrixProperty(name = "execution.isolation.thread.timeoutInMilliseconds", value = "5000")
    ], threadPoolProperties = [
            @HystrixProperty(name = "coreSize", value = "20"),
            @HystrixProperty(name = "maxQueueSize", value = "150"),
            @HystrixProperty(name = "queueSizeRejectionThreshold", value = "150")
    ], threadPoolKey = 'CartService', fallbackMethod = 'transferItemsFallback')
    CartResponseDTO transferItems(CartKey fromKey, CartKey toKey, List itemIds){
     ...
}

CartResponseDTO transferItemsFallback(CartKey fromKey, CartKey toKey, List items, Throwable e) { ... }

When using the service wishListService in a class the following exception is thrown

com.netflix.hystrix.contrib.javanica.exception.FallbackDefinitionException: fallback method wasn't found: transferItemsFallback([class com.bsb.cart.domain.CartKey, class com.bsb.cart.domain.CartKey, interface java.util.List])
    at com.netflix.hystrix.contrib.javanica.utils.MethodProvider.getFallbackMethod(MethodProvider.java:82) ~[hystrix-javanica-1.5.2.jar:1.5.2]
    at com.netflix.hystrix.contrib.javanica.utils.MethodProvider.getFallbackMethod(MethodProvider.java:55) ~[hystrix-javanica-1.5.2.jar:1.5.2]
    at com.netflix.hystrix.contrib.javanica.aop.aspectj.HystrixCommandAspect$MetaHolderFactory.metaHolderBuilder(HystrixCommandAspect.java:120) ~[hystrix-javanica-1.5.2.jar:1.5.2]
    at com.netflix.hystrix.contrib.javanica.aop.aspectj.HystrixCommandAspect$CommandMetaHolderFactory.create(HystrixCommandAspect.java:208) ~[hystrix-javanica-1.5.2.jar:1.5.2]
    at com.netflix.hystrix.contrib.javanica.aop.aspectj.HystrixCommandAspect$MetaHolderFactory.create(HystrixCommandAspect.java:106) ~[hystrix-javanica-1.5.2.jar:1.5.2]

It would appear that when calling reflection for the class in com.netflix.hystrix.contrib.javinca.utils.MethodProvider line 92
return Optional.of(type.getDeclaredMethod(name, parameterTypes)); it does not resolve the super classes fallback method.

We are using groovy and grails and this may be a factor in the evaluation of the type methods via reflection. However calling the following DOES resolve the method.

type.getSuperclass().getDeclaredMethod(name, parameterTypes)

[mattrjacobs]

@dmgcodevil Any thoughts on this one?

[dmgcodevil]

Interesting, will take a look. Ok.
On 18 Apr 2016 7:04 p.m., "Matt Jacobs" notifications@github.com wrote:

@dmgcodevil https://github.com/dmgcodevil Any thoughts on this one?

—
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub
#1177 (comment)

[bischoje]

We have the exact same scenario and problem as ctoestreich describes in this issue. However, we are not using Groovy or Grails, just plain Java.

I'm going to attempt a workaround for this in my project by declaring the fallback methods as 'abstract' in the super class (which happens to be abstract in my project), thus forcing the subclasses to implement them. This workaround obviously won't work for everyone, so it would be good to make the fallback method search more robust so it can find inherited methods too!

[mattrjacobs]

ping @dmgcodevil

any thoughts?

[dmgcodevil]

Hi was busy with other activities, on the next weekend I will recall all
existing issues and try fix them at once. Sorry for inconvenience.
On 20 Jun 2016 11:26 p.m., "Matt Jacobs" notifications@github.com wrote:

ping @dmgcodevil https://github.com/dmgcodevil

any thoughts?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#1177 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/ABWS8SyhP5kQvQpZltDqiQ9ddemEg3fGks5qN1nMgaJpZM4IGg2z
.

[bischoje]

My workaround to implement in the subclasses is working for now. It would be great to resolve this though, thank you in advance!

[dmgcodevil]

@mattrjacobs can we close ?

[mattrjacobs]

Yes, thanks @dmgcodevil !