Elasticsearch maintenance procedure

hi all,
I'm starting to think about the correct procedure to apply security patches on the ES nodes (OS: CentOS).
I have several clusters with different purpose but I focused on the most important; this cluster is composed in this way:

  • 3 Master nodes
  • 10 Data nodes
  • 2 Coordinating-only nodes

generally, OS's security patches do not require server reboot but:

  • should I set a specific parameter before to start the procedure?
  • should I consider something particular to avoid any overload?

I suppose I will patch servers using this schema:

day1: 1 master and 2 data
day2: 1 master and 2 data
day3: 1 master and 2 data
day4: 1 coord and 2 data
day5: 1 coord and 2 data

The process described here can be applied to this approach - https://www.elastic.co/guide/en/elasticsearch/reference/current/rolling-upgrades.html

1 Like