I am still playing with it, I will upload if I hit a dead end. But here is 
strange behavior I noticed, may be you can shed some light.
We switched Unicast IPs to private IPs as opposed to Public IPs.

In Node1: if I do [IP1,IP2,IP3], then node1 & node2 seem to be in a cluster 
and node3 isn't
In Node1; if I do [IP1,IP3,IP2], then node1 & node3 seem to be in cluster 
and node2 isn't.

Does this mean anything?
Also When I query cluster state, for the 2 nodes that joined, cluster 
version says "3", for the non-joing node, cluster version would show "2".



On Wednesday, October 1, 2014 2:20:15 AM UTC-4, David Pilato wrote:
>
> Try setting your logging.yml file with discovery: trace and upload your 
> logs to Gist.
> Just replace key/secret from logs with XXXXXX as you probably don't want 
> to share them :)
>
> Envoyé de mon iPad
>
> Le 30 sept. 2014 à 22:57, sammy <[email protected] <javascript:>> a 
> écrit :
>
> This is driving me insance, I am trying to setup a three node AWS cluster. 
> Each work fine individually, but cluster isn't working. I am using latest 
> of everything, ES 1.3.4, AWS plugin 2.3.0
>
> Note in the logs below, node 2 shows that it was elected as a master, yet 
> node 1 doesn't know about it.
>
> yml: (identical for both nodes)
>
> discovery.zen.minimum_master_nodes:2
>
> discovery.zen.ping.multicast.enabled: false
>
> discovery.zen.ping.unicast.hosts: [node1,node2,node3]
> cloud:
>
> aws:
>
>  access_key:########
>  secret_key:#################
>
> discovery:
>  type:ec2
>
>
> cluster
>    "status": "green",
>    "timed_out": false,
>    "number_of_nodes": 1,
>    "number_of_data_nodes": 1,
>    "active_primary_shards": 0,
>    "active_shards": 0,
>    "relocating_shards": 0,
>    "initializing_shards": 0,
>    "unassigned_shards": 0
> }
>
>
>
>
> Here is what I get in log:
>
> [2014-09-30 20:38:06,279][INFO ][node                     ] [Rex 1] 
> version[1.3.4], pid[11632], build[a70f3cc/2014-09-30T09:07:17Z]
>
> [2014-09-30 20:38:06,279][INFO ][node                     ] [Rex 1] 
> initializing ...
>
> [2014-09-30 20:38:06,320][INFO ][plugins                  ] [Rex 1] 
> loaded [mapper-attachments, cloud-aws], sites []
>
> [2014-09-30 20:38:08,537][INFO ][node                     ] [Rex 1] 
> initialized
>
> [2014-09-30 20:38:08,537][INFO ][node                     ] [Rex 1] 
> starting ...
>
> [2014-09-30 20:38:08,628][INFO ][transport                ] [Rex 1] 
> bound_address {inet[/0:0:0:0:0:0:0:0:9300]}, publish_address {inet[/
> 10.164.184.135:9300]}
>
> [2014-09-30 20:38:08,637][INFO ][discovery                ] [Rex 1] 
> rexCluster/cV8zAk42Ql6RPrAov4LRhQ
>
> [2014-09-30 20:38:11,649][INFO ][cluster.service          ] [Rex 1] 
> new_master [Rex 
> 1][cV8zAk42Ql6RPrAov4LRhQ][ip-10-164-184-135][inet[/10.164.184.135:9300]], 
> reason: zen-disco-join (elect
>
> ed_as_master)
>
> [2014-09-30 20:38:11,665][INFO ][http                     ] [Rex 1] 
> bound_address {inet[/0:0:0:0:0:0:0:0:9200]}, publish_address {inet[/
> 10.164.184.135:9200]}
>
> [2014-09-30 20:38:11,666][INFO ][node                     ] [Rex 1] started
>
> [2014-09-30 20:38:11,682][INFO ][gateway                  ] [Rex 1] 
> recovered [0] indices into cluster_state
>
> [2014-09-30 20:42:44,368][INFO ][node                     ] [Rex 1] 
> stopping ...
>
> [2014-09-30 20:42:44,389][INFO ][node                     ] [Rex 1] stopped
>
> [2014-09-30 20:42:44,390][INFO ][node                     ] [Rex 1] 
> closing ...
>
> [2014-09-30 20:42:44,395][INFO ][node                     ] [Rex 1] closed
>
> [2014-09-30 20:42:45,833][WARN ][common.jna               ] Unable to lock 
> JVM memory (ENOMEM). This can result in part of the JVM being swapped out. 
> Increase RLIMIT_MEMLOCK (ulimit).
>
> [2014-09-30 20:42:45,915][INFO ][node                     ] [Rex 1] 
> version[1.3.4], pid[11744], build[a70f3cc/2014-09-30T09:07:17Z]
>
> [2014-09-30 20:42:45,915][INFO ][node                     ] [Rex 1] 
> initializing ...
>
> [2014-09-30 20:42:45,953][INFO ][plugins                  ] [Rex 1] loaded 
> [mapper-attachments, cloud-aws], sites []
>
> [2014-09-30 20:42:48,165][INFO ][node                     ] [Rex 1] 
> initialized
>
> [2014-09-30 20:42:48,166][INFO ][node                     ] [Rex 1] 
> starting ...
>
> [2014-09-30 20:42:48,254][INFO ][transport                ] [Rex 1] 
> bound_address {inet[/0:0:0:0:0:0:0:0:9300]}, publish_address {inet[/
> 10.164.184.135:9300]}
>
> [2014-09-30 20:42:48,257][INFO ][discovery                ] [Rex 1] 
> rexCluster/3A7q_M1AT4GvVJkhfwzobQ
>
> [2014-09-30 20:43:18,260][WARN ][discovery                ] [Rex 1] 
> waited for 30s and no initial state was set by the discovery
>
> [2014-09-30 20:43:18,266][INFO ][http                     ] [Rex 1] 
> bound_address {inet[/0:0:0:0:0:0:0:0:9200]}, publish_address {inet[/10.164
> .184.135:9200]}
>
> [2014-09-30 20:43:18,266][INFO ][node                     ] [Rex 1] 
> started
>
> [2014-09-30 20:44:03,464][DEBUG][action.admin.cluster.state] [Rex 1] no 
> known master node, scheduling a retry
>
> [2014-09-30 20:44:03,464][DEBUG][action.admin.indices.mapping.get] [Rex 1] 
> no known master node, scheduling a retry
>
>
> <span style="color: #660;" class="styled-by
>
> ...

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" 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/elasticsearch/d6513c42-917b-467c-bcf6-ba2dc8194231%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to