Another way to purge gossip info from each node is to:
1. Gracefully stop cassandra i.e. nodetool drain; kill Casandra PID 2. Move/delete files from $DATADIR/system/peers/ 3. Add JVM_OPTS="$JVM_OPTS -Dcassandra.load_ring_state=false" in jvm.options file 4. Restart Cassandra service. 5. Repeat above steps on each nodes in dc/cluster. 6. Once gossip info is purged, remove jvm option added in step 3 and restart instance again. Depending on cluster,load size you may get this done swiftly. ~Asad From: yuping wang [mailto:yupingwyp1...@gmail.com] Sent: Monday, July 29, 2019 10:56 AM To: user@cassandra.apache.org Subject: Re: Jmx metrics shows node down Is there workaround to shorten 72 hours to something shorter?(you said by default, wondering if one can set a non-default value?) Thanks, Yuping On Jul 29, 2019, at 7:28 AM, Oleksandr Shulgin <oleksandr.shul...@zalando.de<mailto:oleksandr.shul...@zalando.de>> wrote: On Mon, Jul 29, 2019 at 1:21 PM Rahul Reddy <rahulreddy1...@gmail.com<mailto:rahulreddy1...@gmail.com>> wrote: Decommissioned 2 nodes from cluster nodetool status doesn't list the nodes as expected but jmx metrics shows still those 2 nodes has down. Nodetool gossip shows the 2 nodes in Left state. Why does my jmx still shows those nodes down even after 24 hours. Cassandra version 3.11.3 ? AFAIK, the nodes are not removed from gossip for 72 hours by default. Anything else need to be done? Wait another 48 hours? ;-) -- Alex
