Archives of Support Questions (Read Only)

This is an archived board for historical reference. Information and links may no longer be available or relevant
Announcements
This board is archived and read-only for historical reference. To ask a new question, please post a new topic on the appropriate active board.

hdp search installation

avatar

I have a smalll dev cluster with 5 nodes (2 masters and 3 slaves). I want to install solr (hdp search) in the cluster.

Which all nodes should i install solr (hdp search). Is there any document where i can refer the configuration inforamation.

1 ACCEPTED SOLUTION
7 REPLIES 7

avatar
Master Mentor

avatar
Master Mentor

HDP Search should be deployed on each node that runs HDFS. The HDP Search package is available as an RPM, located in the HDP-Utils repository.

avatar

Seems to be Link is not working , Can you please validate it.

avatar

Put the Solr instances on the slave/Datanodes.

In regards to configuration, see this https://community.hortonworks.com/articles/15159/securing-solr-collections-with-ranger-kerberos.html although its focused on securing Solr with Kerberos and Ranger, there is also a small section about the installation.

avatar

Thanks @Jonas Straub. the information you provided was extremely useful. As of now i am going to setup the solrcloud on a HDP cluster without kerberos. I followed your steps and it worked fine.

Is HDP search and solr on yarn both different.? Any clarity would be great.

avatar

Hi @Jonas Straub, i have a question regarding the HDP search installation. Since we are using the machines running datanodes to host the Solr instances. how do we control resources used by solr, so that the cluster resources are not dominated by solr alone. There are lot more other stuff running on the slave nodes. I guess we can't do the resource management with yarn for this case. I guess this would be a common usecase for most of them using solr with hadoop

avatar
Master Collaborator

If your index is of any significant size or has much of a load on it, I would recommend you install it on separate nodes.