Cassandra nodetool status down and generation time very old

Question

Our cassandra cluster went down. It was not maintained properly. The nodetool status from each node is different. So, when I tried the gossip info, the generation timestamp if very old.

/192.168.0.35
  generation:1539742113
  heartbeat:8705953
  DC:datacenter1
  STATUS:NORMAL,-1117919639473599473
  SEVERITY:0.0
  HOST_ID:4bc33d49-6f71-4e33-9790-f9ae46180613
  LOAD:1.097982388E9
  NET_VERSION:8
  RPC_ADDRESS:192.168.0.35
  RACK:rack1
  RELEASE_VERSION:2.1.6
  SCHEMA:04c4cc4f-7598-3e69-9455-8d8896b0624d
/192.168.0.36
  generation:1504926520
  heartbeat:2147483647
  DC:datacenter1
  STATUS:shutdown,true
  SEVERITY:0.0
  HOST_ID:0708fc6f-0423-46fc-8f56-8e7512a83de4
  LOAD:1.2352913099E10
  NET_VERSION:8
  RPC_ADDRESS:192.168.0.36
  RACK:rack1
  RELEASE_VERSION:2.1.6
  SCHEMA:04c4cc4f-7598-3e69-9455-8d8896b0624d
/192.168.0.34
  generation:1541989374
  heartbeat:1890169
  DC:datacenter1
  STATUS:NORMAL,-100967197570522715
  SEVERITY:0.20000000298023224
  HOST_ID:f3e8f987-055d-42f9-b9aa-0516d6890235
  LOAD:4.243156213E9
  NET_VERSION:8
  RPC_ADDRESS:192.168.0.34
  RACK:rack1
  RELEASE_VERSION:2.1.6
  SCHEMA:04c4cc4f-7598-3e69-9455-8d8896b0624d
/192.168.0.33
  generation:1540772736
  heartbeat:5580091
  DC:datacenter1
  STATUS:NORMAL,-1055061873163068507
  SEVERITY:0.20080320537090302
  HOST_ID:c827ef32-f076-456f-a0ad-2f3289ef3e5b
  LOAD:3.123707903E9
  NET_VERSION:8
  RPC_ADDRESS:192.168.0.33
  RACK:rack1
  RELEASE_VERSION:2.1.6
  SCHEMA:04c4cc4f-7598-3e69-9455-8d8896b0624d

What should I do to make nodes gossip more? Thank you

Answer 1

In this case, you should restart the node check the status. because when the node get started it establish gossip again on all the nodes.