I see a lot of that on one of our clusters. There is pretty much no load on it. Clocks are quite well synchronized.
date +start:%H:%M:%S.%N ; for d in 2 3 4 5 6 8 9; do ssh 10.3.0.17$d date +%H:%M:%S.%N & done;date +end:%H:%M:%S.%N start:01:22:45.478726356 end:01:22:45.480993049 01:22:45.570313246 01:22:45.572781468 01:22:45.584917988 01:22:45.595550249 01:22:45.595120011 01:22:45.598196670 01:22:45.863486947 That is 3.0.9. I see the log message ~2000 times an hour On another cluster, 3.11 I see it happening ~5 times an hour. On 22 September 2017 at 01:23, Nate McCall <n...@thelastpickle.com> wrote: > >> The message in the debug log is >> >> DEBUG [GossipStage:1] 2017-09-21 09:19:52,627 FailureDetector.java:456 - >> Ignoring interval time of 2000275419 >> >> >> > Did you truncate the log message? There should be and "for [endpoint]" on > the end which should help you narrow things down to a set of problem nodes. > I agree with Jeff in that this is most likely NTP sync issue or network > flap, though. > -- Best Regards, *Alexander Kotelnikov* *Team Lead* DIGINETICA Retail Technology Company m: +7.921.915.06.28 *www.diginetica.com <http://www.diginetica.com/>*