ElasticSearch instance not reachable from outside the server - Azure Windows 2012

Question

I have set up Elastic Search 2.3.0 as a service in an Azure VM with Windows Server 2012 R2. I can access elastic search instance from within the server using

http://localhost:9200

but I cannot access from outside the server.

What have I tried?

• Just for testing I have made the VM available outside the virtual network.
• Opened up the port 9200 in Windows Firewall settings as an Inbound Traffic rule
• Added an endpoint in Azure portal settings opening up port 9200

Tried this with VM Classic as well.

Also for testing whether the VM is alright, I had set up IIS in that machine, opened up the port 80. The default IIS page is accessible from outside.

Only thing I have not tried is setting ES in a Linux VM.

I can't find anything in the elasticsearch logs. On the browser (Chrome), there are no logs in the Network. It just spins, waiting for the server, and snaps with "ERR_EMPTY_RESPONSE".

Can anybody that had done this shed some light please?

UPDATE: Here is what I can see in netstat:

TCP    127.0.0.1:9200         machine-name:0       LISTENING
TCP    127.0.0.1:9300         machine-name:0       LISTENING

Answer 1

Starting with ES 2.0, the elasticsearch process only binds to localhost, which explains why you can query ES from inside but not from the outside.

You need to change the following setting in your elasticsearch.yml configuration file:

network.host: 0.0.0.0

Answer 2

Resolved for me on Windows server 2012 with Elasticsearch 7.1.1 adding in C:\ProgramData\Elastic\Elasticsearch\config\elasticsearch.yml the following config:

http.host: 0.0.0.0

Answer 3

I had the same case. Looking in the log

tail -f /var/log/elasticsearch/NODE.log

I saw this:

 [NODE_NAME] publish_address {10.XXX.XXX.XXX:9300}, bound_addresses {10.XXX.XXX.XXX:9300}
 [NODE_NAME] bound or publishing to a non-loopback address, enforcing bootstrap checks
 [NODE_NAME] node validation exception
 [1] bootstrap checks failed
 [1]: the default discovery settings are unsuitable for production use; at least one of [discovery.seed_hosts, discovery.seed_providers, cluster.initial_master_nodes] must be configured
 [NODE_NAME] stopping ...
 [NODE_NAME] stopped

Solution:

Modify the archive /etc/elasticsearch/elasticsearch.yml in the line discorvery seed with this:

discovery.seed_hosts: ["127.0.0.1", "[::1]"]

Done.

Answer 4

I would suggest to set the network.host to special values

eg.

network.host: [_local_, _site_]

This works for me.

Answer 5

Please note that for me success was achieved only after setting

http.host: 0.0.0.0

instead of

http.host: "localhost"