@Antoni Ivanov The cluster is new one, its size is small^_^ 2017-04-11 0:02 GMT+08:00 Silvius Rus <[email protected]>:
> On Sun, Apr 9, 2017 at 10:46 PM, Antoni Ivanov <[email protected]> wrote: > >> @Silvus How would this issue explain High cpu usage. It states the >> threads are in state “waiting” in the bug. >> > > The message I responded to stated 100% to 200% Catalog CPU usage, which > means 1 to 2 cores fully used, which translates to 1 to 2 threads actually > loading (and the other 15 or 14 possibly waiting on the same tables). > That's behavior consistent with IMPALA-4765. > > There could be other reasons for this behavior. I'm not saying > IMPALA-4765 is the root cause in this case, just that the described > behavior matches the known issue. >
