After running 200 concurrent loggers repeated 10x, and letting the machine sit for a bit, the memory seems to have gone back down.
So it doesn't seem to be a reproducer.
Maybe we have a different bug where clients keep connecting to the controller and not closing the conn. (possible 2.4 vs 2.5 bug?)
After running 200 concurrent loggers repeated 10x, and letting the machine sit for a bit, the memory seems to have gone back down.
So it doesn't seem to be a reproducer.
Maybe we have a different bug where clients keep connecting to the controller and not closing the conn. (possible 2.4 vs 2.5 bug?)