Support Questions

Find answers, ask questions, and share your expertise
Announcements
Celebrating as our community reaches 100,000 members! Thank you!

how to build ambari cluster on one node

avatar

until now we build every ambari cluster , as the following structure:

3 Journal nodes machines ( inside each journal node we build the kafka )

3 workers machines (data-node machines)

but in order to minimize the costs we want to build one linux machine that includes one Journal node machine ( master ) + worker + kafka

dose HORTON-WORKS have procedure for this structure - to build Journal node + worker + kafka on one single machine ?

Michael-Bronson
1 ACCEPTED SOLUTION

avatar
Master Mentor

@Michael Bronson

Do you mean that there is no NameNode on that single host? As you are talking about JournalNode so do you means that you want to have NameNode HA on a single node machine?
.

If NameNode HA is not a requirement then you can take a look at the HDP Sandbox and then export it's Blueprint then create a cluster based on that. Or Just take a HDP Sandbox then delete whatever unwanted services/components that you do not want in your single node cluster and then you can either use the same or export it's blueprint and then use it.

View solution in original post

9 REPLIES 9

avatar
Master Mentor

@Michael Bronson

Do you mean that there is no NameNode on that single host? As you are talking about JournalNode so do you means that you want to have NameNode HA on a single node machine?
.

If NameNode HA is not a requirement then you can take a look at the HDP Sandbox and then export it's Blueprint then create a cluster based on that. Or Just take a HDP Sandbox then delete whatever unwanted services/components that you do not want in your single node cluster and then you can either use the same or export it's blueprint and then use it.

avatar

namenode should be defined on this machine , all we want is to minimize all in one machine ( amster + worker + kafka )

Michael-Bronson

avatar

@Jay regarding the sendbox from where to download the BP file? , as you know we separate the worker from the master machine so how to integrate them on one machine )

Michael-Bronson

avatar
Master Mentor

@Michael Bronson

HDP sandbox is a Single Node Cluster. Which can be downloaded from https://hortonworks.com/products/sandbox/

Following is the link which shows how to use the HDP Sandbox (Single Node Cluster) https://hortonworks.com/tutorial/learning-the-ropes-of-the-hortonworks-sandbox/

avatar

@Jay just to be clear I want to install the sandbox on my existing Linux redhat machine ( I am already configured the host-name and IP on that local machine ) , so this is relevant for me ?

Michael-Bronson

avatar

@Jay , we have some problem when we deploy OVA template , can you get you assit regarding that - https://community.hortonworks.com/questions/186221/cant-deploy-sandbox-hdp-264-ova-template-from-vsp...

Michael-Bronson

avatar
Master Mentor

@Michael Bronson

I was re-reading this thread and looks like i have misunderstood your requirement.

Can you please confirm if my understanding regarding your query is correct or not?

Based on your issue description i understood that you want to configure

1. A Single Node Cluster.

2. That single node cluster should have "one Journal node machine ( master ) + worker + kafka". Means in thjis Single Node you want to install a JournalNode + DataNode + Kafka Broker


My Query: In this case where will the NameNode be. Also as you are using a term Journal Node which is available in NameNode High Availability mode . So are you planning to install both the name Nodes in the same Single Host (by making some port conflict resolution) ?

avatar

@Jay your steps 1 and 2 are correct , we want to install the "master" + "worker" + "kafka" on one single linux node ,

Michael-Bronson

avatar

@Jay do you have suggestion how to install Master + worker + kafka on single node ?

Michael-Bronson