Elasticsearch nodes work wrong

[hasti_gh (Hasti Ghobakhloo)]

hi
i have a question about nodes in a cluster in ES. i have 4 nodes in my cluster for example :
host ip heap.percent ram.percent load node.role master name
Switch182 127.0.1.1 13 9 0.00 d * Cadaver
Porniko 192.168.1.32 4 43 d m Wiz Kid
ENAYATI-PC 192.168.1.215 7 51 d m Killraven
Ghobakhlo 192.168.0.231 7 48 d m Architect

ES work correctly but sometimes when node Wiz Kid become down or become start some CRUD dont work.for example Wiz Kid is start and DELETE in ES does not work and when Wiz Kid become down it work.why???
in a error that ES show it saied some shards are failed becouse of 192.168.1.32.
is it possible that 2 nodes become master in a time??? and this is result of this problem?? is it correct that nodes in ES are depend to each other or not???
thanks

[Christian_Dahlqvist]

What do you have ‘minimum_master_nodes’ set to in your config? As you seem to have 4 master eligible nodes it should be 3.

[hasti_gh (Hasti Ghobakhloo)]

discovery.zen.minimum_master_node=1
in elasticsearch.yml in config folder i have 1 master node.
only Switch182 127.0.1.1 13 9 0.00 d * Cadaver is master

[Christian_Dahlqvist]

That is a problem. minimum_master_nodes need, as described in the documentation, to be set based on the number of master eligible nodes you have in the cluster.