Can the fuzzer such as AFL and symcc run separately in independent docker and conduct hybrid fuzzing by sharing the output folder of fuzzing?

[komorogrov]

In practice, I deployed symcc and AFL respectively in two different Dockers, and at the same time, I achieved the collaboration between the two by setting the output folders for shared fuzz testing between the two Dockers when docker starts up.

In this mode, symcc can indeed run normally, but it only generates output at the beginning and no valid test cases are produced subsequently. And from the output information, it can be seen that the order of symcc_helper_fuzzing when facilitating the seed queue is very chaotic.

I don't know what causes this.

So I want to know if this architectural pattern conforms to the operating principle of symcc_helper_fuzz. How does it not conform and what insurmountable problems exist?

Thanks