My requirement is to install plain vanilla Hadoop(It’s not that I dislike HDP but I want to get my hands on dirty on plain hadoop). Post Ambari installation(I have 1 ambari server, 1 master and 1 slave node) done using vagrant and VM's, how should I move ahead to install plain hadoop.
I checked Ambari wiki page for stacks/services but It's not very clear.
Can anyone please guide me.,
I have installed multi node Ambari cluster (1 ambari server, 1 master node and 1 slave node). Now I want to install plain hadoop (not HDP) using Ambari. It's not that I dislike HDP but my requirement is to use plain hadoop. Can anyone please tell me as to how to do it.
I tried Ambari custom stack/services wiki page too but it seems I would need to extend HDP version only to create my custom one. Also, the instructions on wiki page are not clear.
Please guide.It's urgent.
You can check the HDP manual install guide at below link:
This lets you perform the installation and configurations manually, so definitely provides a hands-on approach
The other reference would be the apache hadoop guide:
Thanks Vivek for responding but I want to install plain hadoop (not HDP customized version) and that too using Ambari.
Can you please help.
You may want to check how to install custom stacks using Ambari:
For your requirement there is work involved like - defining your custom stack (something like "Plain Hadoop Stack"), defining services that would be part of that stack (may want to start with Core services first like HDFS, YARN, Zookeeper), define service configurations, add repository details in service config so Ambari knows from where to install (could be HDP public repo or your own defined repo that would point to the appropriate location of packages).
Another reference for custom services would be: https://github.com/abajwa-hw/ambari-workshops#ambari-stack-services
Also as mentioned earlier, if your objective is to get hands-on with Hadoop installation - HDP Manual install is a good starting point. You can later try an Ambari automated HDP install and that would give a good sense of how deployment works both ways
Hope this helps!
While installing the Hadoop stack through Ambari it fetches install packages from hortonworks repo, even if you install apache Ambari distribution. Check Repo URLs at: https://cwiki.apache.org/confluence/display/AMBARI/Install+Ambari+2.2.2+from+Public+Repositories
However, if you want to experiment with installation below are a few things that could help you:
1. Ambari downloads the binaries to be installed at "/var/lib/ambari-server/resources/stacks/HDP/2.4/services/" directory.
2. You can copy the required services binaries at "/var/lib/ambari-server/resources/stacks/HDP/2.4/services/" on Ambari server machine and try running the installation.
NOTE: the approach is not verified but worth a try.
Thanks Puneet for your answer.
I think using the binaries downloaded at the location specified above will install HDP customized version of Hadoop, but I need to install plain hadoop. Please confirm.
The binaries downloaded by Ambari will install HDP for sure but you need to delete them and place your custom binaries before running the cluster setup.
Thanks for your response.
1. Can you please help me with the steps of creating own repo of plain hadoop?
2. Can I use the common services specified at below location to create plain hadoop repo?
Hi Shakir here,
Did anyone got the solution for this thread.
Please let me know.