re-enter your cluster information in NLS

[lukedevon]

hi

I have installed two NLS instances and configured SSL as well. I tried to add second instance to the cluster.

Added the IP address of the 1st NLS and its cluster ID.

But I had to face for following error;

"Could not establish connection, this could be due to a slow connection, or you may want to re-enter your cluster information."

I tried all the possible ways; but still I am getting same error.

Can somebody help me on this issue plz?

Thank you
Luke

[cdienger]

Is there anything logged in /var/log/elasticseach/<cluster_uuid>.log when you try to add the new node ?

As a quick test, can you revert the SSL settings and test? If you followed https://assets.nagios.com/downloads/nag ... Server.pdf to set it up then it should just be a few changes to disable it.

[lukedevon]

Hi

Removed ssl in the second node and I was able to add it to the cluster. Which means , first NLS is configured with SSL and second one without SSL.
Once it has been added to the cluster, I configured and enabled SSL in the second NLS too.

Now its working fine.

However I have another question;

Suppose each nodes having 2 NIC and having IPs assigned. How can we point NLS to a particular IP ? I need NLS to be in eth1 but its always getting eth0.

Thanks
Luke.

[cdienger]

Can you clarify what you mean by "its always getting eth0" ? By default the services on NLS typically listen on all interfaces - run "netstat -na | egrep ":::80|:::9200|:::9300|:::5544" to verify.

[lukedevon]

Hi,

eth0 - ip x

eth1 - ip y

but i want nagiosls to bind/connect/interface with eth1 - ip y. But NLS always get the ip x of eth0.

Thank you
Luke.

[cdienger]

Are you referring to the web ui? What happens if you point your browser to the IP of eth1(http://eth1_ip/nagioslogserver)? The web interface should be listening on both interfaces and you can confirm that with "netstat -na | grep :::80". You can force it to listen to a specific inteface by modifying the "Listen" option in /etc/httpd/conf/httpd.conf.

For example, the default is:

Listen 80

Try changing it to:

Listen eth1_ip:80

and restarting Apache:

service httpd restart

[lukedevon]

Hi ,

It was help me to fix the message and now the cluster is up and running. However I would like to discus with you one of nls implementation that might helps for others as well.

Scenario;
1. I have four nodes nls cluster with SSL enabled.
2. In each nls instances having 2 network interfaces, For example: eth1, and eth2
3. eth1 - 192.168.1.30 and eth2 - 172.30.50.200; likewise rest of 3 nodes also have been assigned the particular IPs in respective network subnet.
4. eth1 - allows "user access traffics" for port 80/443 , eth2 - allows only data traffic. (logs)

Observation;
When checking the instance status : admin --> instance status, eth1, ip has assigned with the port 9300.

That us not what I expected. I want only eth2, Ips assigned for ports 9300.

how can it be done? what needs to be done during nls installation ?

Thank you
Luke.

[cdienger]

There's nothing to do during the install, but you can control which IP elasticsearch listens on by editing /usr/local/nagioslogserver/elasticsearch/config/elasticseach.yml and setting network.host. By default this is commented out and you'll need to remove the # and space after it, and set the IP accordingly:

network.host: 172.30.50.200

Restart elasticsearch afterwards:

service elasticsearch restart