Support Questions
Find answers, ask questions, and share your expertise

Hosts inspection Fails - Connection Refused


I created the cluster, the installation of packages/parcels is done. Able to ssh each host from the master host. Now the thing's "Inspect the Hosts" is getting failed showing the "connection Refused"


I had looked for /etc/hosts file in all the 3 hosts, it's like 


Kindly help

Thanks in advance




AWS uses intzernal IP's to communicate,I can see all the /etc/hosts entries are commented out except


All the 3 hosts including the Ambari host should be in every /etc/hosts on all the 3 nodes for easy identification use the format Internal IP :FQDN:ALIAS see example below.

10.0.7.x          ec2.xxxxx      Slave1
10.0.2.x          ec2.xxxxx       Slave2
10.0.4.x         ec2.xxxxx        Slave3

And before launching host inspector confirm you all host can ping each other hoping you followed the Cloudera pre-requisite recommendation for preparing the environments



Hi @Shelton , edited my hosts file as below... localhost localhost.localdomain localhost4 localhost4.localdomain4
::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 master slave1 slave2

still my problem persists in running inspection on hosts



I am not able to figure it out. If needed logs, which logs and from where shall I attach ?

Master Collaborator

As @Shelton mentioned, I do not recommend to use Public IP addresses at all.

If you need to understand what is happening in the background check the cloudera manager logs on the CM instance. Also check out if you are able to SSH from the CM host to the new nodes from command line.

Cloudera python scripts are using a lot of time python socket library to convert IP to hostname or hostname to IP, and my experience is that you should have these information on all hosts consistent and you should not mix with public IP addresses.


Now the inspector test is successful only on the Master node...and the other two nodes failed inspector run tests.cdh6.PNG


@Shelton @Tomas79  Pls help me here. It's asking for swappiness change to 10, Just changing in the master node helps?

Master one is successfully able to complete the inspection test, while the other two getting failed "Connection Timed Out"



The swappiness is a configuration that should be changed on all hosts in the cluster.

Talking about swappiness  those are the steps you should have done while preparing the environment  I  remenber asking you that about the firewall ensure on each host you have disabled the FW  disabling it in the VNET or PVC level  doesnt disable  it automatically at the host level




Any updates?

; ;