Simon

> Should be finished by #15348 this issue you mentioned had been solved locally before. memory leak still exists   

> executor yes,I have here is the code 

> @simmonn Could you please provide the heap dump file or stack info? Maybe this is caused by thread leak. I simulated it in the test environment. Executing the SSH...

> Yes, I changed the heartbeat to IGNORE then the thread leak can be resolved. Please test. @simmonn @peak-xu Thank you, I'll try this approach.

> Could you make sure the html selectors exist on that page? > > Also, could you make sure that the base url of those links are specified in `start_urls`...