Occasionally, deadlocks occur when the main thread and asynchronous thread initialize the same Feign Api call

[dongfangding]

When I initiate a remote call for initialization using a Feign API in the main thread, and perform similar operations using the same Feign API in an asynchronous thread in the initialization of another class. I am following org.springframework.cloud.context.named.NamedContextFactory#getContext obtains the lock, there will be a call to the context.refresh(), which in turn will contain another lock. At certain times, due to uncontrollable sequence, the main thread may experience a deadlock and fail to start.

Sample

@Component
Class A {
   @PostConstruct
   public void init() {
       // Feign Api remote call
   }
}

@Component
Class B {
   @PostConstruct
   public void init() {
       Executors.newSingleThreadExecutor().execute(() -> {
             // Same feign API calls
       });
   }
}