1

I am trying to use Kuzzle 2.13.3 using docker on a 8GB RAM 4vCPU self-managed cloud instance. Everything seems fine, except that the Kuzzle node shuts down on its own after random period of time. Here are the logs towards the end:

{"level":"error","message":"2021-07-22T06:25:13+00:00 [LOG:ERROR] [knode-icky-vampire-54285] [CLUSTER] Node too slow: ID card expired"}
{"level":"info","message":"2021-07-22T06:25:13+00:00 [LOG:INFO] [knode-icky-vampire-54285] Initiating shutdown..."}
{"level":"info","message":"2021-07-22T06:25:13+00:00 [LOG:INFO] [knode-icky-vampire-54285] Halted."}
[nodemon] clean exit - waiting for changes before restart

what I need to do to get past this issue?

James Z
  • 12,209
  • 10
  • 24
  • 44
Pradip Biswas
  • 21
  • 2

1 Answers1

0

When the node cannot refresh it's ID card inside Redis in time, it will shutdown to ensure always having an healthy cluster.

You can try to increase the heartbeat delay from the configuration with the cluster.heartbeat key. The default value is 2000 (ms).

Aschen
  • 1,691
  • 11
  • 15
  • Hi @Aschen, thanks a lot for your insightful comment. I will definitely try your suggestion. However, I am still curious: as I have not configured any cluster as such (i.e. I am running only one instance of the Kuzzle docker container), any clue what might be going wrong in the setup? Like, why would the Kuzzle node suddenly complain once every 3-4 days and shutdown? – Pradip Biswas Jul 23 '21 at 13:17
  • Even if you start only one node, Kuzzle is ready to be deployed as a cluster. When not part of a cluster, the overhead is minimal since no messages has to be exchanged between nodes. However the ID card mechanism is still present since a new node may join at any time – Aschen Jul 26 '21 at 11:23