Nodes split

Hi,

Ia have a cluster with 3 nodes. How can it be possible that 2 of the nodes
are in green status showing number of nodes 3:
{

  • cluster_name: "production",
  • status: "green",
  • timed_out: false,
  • number_of_nodes: 3,
  • number_of_data_nodes: 3,
  • active_primary_shards: 60,
  • active_shards: 180,
  • relocating_shards: 0,
  • initializing_shards: 0,
  • unassigned_shards: 0

}

And the third one have status yellow showing number of nodes 2:
{

  • cluster_name: "production",
  • status: "yellow",
  • timed_out: false,
  • number_of_nodes: 2,
  • number_of_data_nodes: 2,
  • active_primary_shards: 60,
  • active_shards: 120,
  • relocating_shards: 0,
  • initializing_shards: 0,
  • unassigned_shards: 60

}

I assume that if 2 nodes can see all 3 nodes in the cluster then the third
node is in that cluster too.

This situation happend after few weeks cluster stable working, nothing was
restarted here.

What can I check to work this out ?

Thank you for help.
Best regards.
Marcin Dojwa

--

The logs from each node: https://gist.github.com/627e4616102437f87f1d
node1 and node2 shows 3 nodes in cluster, node3 shows 2 nodes in cluster.

Thanks.
Best regards.
Marcin Dojwa

2012/12/7 Marcin Dojwa m.dojwa@livechatinc.com

Hi,

Ia have a cluster with 3 nodes. How can it be possible that 2 of the nodes
are in green status showing number of nodes 3:
{

  • cluster_name: "production",
  • status: "green",
  • timed_out: false,
  • number_of_nodes: 3,
  • number_of_data_nodes: 3,
  • active_primary_shards: 60,
  • active_shards: 180,
  • relocating_shards: 0,
  • initializing_shards: 0,
  • unassigned_shards: 0

}

And the third one have status yellow showing number of nodes 2:
{

  • cluster_name: "production",
  • status: "yellow",
  • timed_out: false,
  • number_of_nodes: 2,
  • number_of_data_nodes: 2,
  • active_primary_shards: 60,
  • active_shards: 120,
  • relocating_shards: 0,
  • initializing_shards: 0,
  • unassigned_shards: 60

}

I assume that if 2 nodes can see all 3 nodes in the cluster then the third
node is in that cluster too.

This situation happend after few weeks cluster stable working, nothing was
restarted here.

What can I check to work this out ?

Thank you for help.
Best regards.
Marcin Dojwa

--