Closed aschaeffer closed 7 years ago
We are using 100% CPU bc we're constantly checking the non-blocking gRPC queue until a certain event arrives.
I just noticed that I disabled the timeout in the master branch for that loop, Ill reenable it.
I think we could just switch to the blocking gRPC API and its fixed
A general thought, though:
in how far should we consider 100% CPU usage a bug? Is it preventing you from doing anything else on the computer? plz describe.
I've added a micro sleep in pr #439 for the initialization routine. This fixes the 100% cpu load. The async method is not necessary but also not wrong. Therefore we should let it until we have a reason for either using sync or async in the initialization routine. Also I don't see a need for a timeout since Inexor Flex
can kill that process by itself.
When
Inexor Flex
startsInexor Core
but doesn't connect to it, theInexor Core
process consumes 100% CPU.