Hi

We have a infrastructure with 2 Graylog nodes behind a LB and 3 
Elasticsearch nodes (2 Data&Master, 1 Master --> because of the split brain 
problem). I've attached a painting of our infrastructure (graylog-web 
excluded).

<https://lh3.googleusercontent.com/-beICyjvnCkQ/Vzw83Q7tK1I/AAAAAAAAAAg/4fqTPTC7xqY37p24cNyMx2WeUax6r3VWgCLcB/s1600/IMG_1127.jpg>

The problem I have is, that if one of the data nodes goes down, graylog 
cannot write any data into the ES database. There is no error about failure 
to elect the master as it was with just two ES master nodes (fixed that by 
adding a master only node on one graylog server). 

If I shutdown the elasticsearch on esnode1 all other systems acknowledge 
this action with
2016-05-18 12:15:06,207][INFO ][cluster.service          ] [elasticsearch-
node2] removed {[elasticsearch-node1][w6iXmbtTRkq3571YA39vcQ][idlog-es-node1
.campus.unibe.ch][inet[/esnode1:9300]],}, reason: zen-disco-node_left([
elasticsearch-node1][w6iXmbtTRkq3571YA39vcQ][idlog-es-node1.campus.unibe.ch
][inet[/esnode1:9300]])

That's it. The cluster can restore it's stae only if I'm going to start the 
service on esnode1 again.

I have no idea why the failover doesn't work and would be more than happy 
if someone has an idea.

Thank you very much in advance. Kind regards, Stefan



Following are the ES related configurations of each server:

## Graylog 1 - /etc/graylog/server/server.conf
elasticsearch_shards = 1
elasticsearch_replicas = 1
elasticsearch_index_prefix = graylog2
allow_leading_wildcard_searches = true
allow_highlighting = true
elasticsearch_cluster_name = graylog-production
elasticsearch_node_name = idgraylog-srv1
elasticsearch_discovery_zen_ping_multicast_enabled = false
elasticsearch_discovery_zen_ping_unicast_hosts = esnode1:9300,esnode2:9300,
graylog2:9300
elasticsearch_analyzer = standard


## Graylog 2 - /etc/graylog/server/server.conf
elasticsearch_shards = 1
elasticsearch_replicas = 1
elasticsearch_index_prefix = graylog2
allow_leading_wildcard_searches = true
allow_highlighting = true
elasticsearch_cluster_name = graylog-production
elasticsearch_node_name = idgraylog-srv2
elasticsearch_discovery_zen_ping_multicast_enabled = false
elasticsearch_discovery_zen_ping_unicast_hosts = esnode1:9300,esnode2:9300,
graylog2:9300
elasticsearch_analyzer = standard

/etc/elasticsearch/elasticsearch.yml
cluster.name: graylog-production
node.master: true
node.data: false
path.conf: /etc/elasticsearch
path.data: /opt/es-data
network.bind_host: graylog2
discovery.zen.minimum_master_nodes: 2
discovery.zen.ping.multicast.enabled: false
discovery.zen.ping.unicast.hosts: ["esnode1:9300", "esnode2:9300", 
"graylog2:9300"]
script.disable_dynamic: true
indices.fielddata.cache.size: 40%


## ESNode1 - /etc/elasticsearch/elasticsearch.yml
cluster.name: graylog-production
node.name: "elasticsearch-node2"
path.conf: /etc/elasticsearch
path.data: /data
node.master: true
node.data: true
network.bind_host: esnode1
discovery.zen.minimum_master_nodes: 2
discovery.zen.ping.multicast.enabled: false
discovery.zen.ping.unicast.hosts: ["esnode1:9300", "esnode2:9300", 
"graylog2:9300"]
script.disable_dynamic: true
indices.fielddata.cache.size: 40%


## ESNode2 - /etc/elasticsearch/elasticsearch.yml
cluster.name: graylog-production
node.name: "elasticsearch-node1"
path.conf: /etc/elasticsearch
path.data: /data
node.master: true
node.data: true
network.bind_host: esnode2
discovery.zen.minimum_master_nodes: 2
discovery.zen.ping.multicast.enabled: false
discovery.zen.ping.unicast.hosts: ["esnode1:9300", "esnode2:9300", 
"graylog2:9300"]
script.disable_dynamic: true
indices.fielddata.cache.size: 40%

-- 
You received this message because you are subscribed to the Google Groups 
"Graylog Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/graylog2/9a3fd585-3a15-4f4d-8ac5-a715d6c76260%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to