How do I convert a Hadoop system to HPC?

Question

I have multiple Limulus Hadoop 400 computing systems and would like to convert at least one to a SOLR search cluster for several large datasets. As I understand, the HPC provides IP node addressing and standard CentOs filesystem. This is all I need to support SOLR. Any assistance is appreciated in converting one system to the other. Thanks!

Answer 1

You mean you want to have a node (not the head node) that can reach out to the internet?

This can be done now with the firewall (NAT masquerading and port forwarding)

Or more specifically do you need a node to have it's own FQDN in addition to the login node?

Comments

In further research I've found SOLR supports HDFS so I believe the machine will work as configured. To clarify, the head node IP is all I need. Thanks.

---

Perfect. BTW if I recall your systems are CentOS 6 based. Our current systems are CentOS 7 with the latest (and last) Cloudera/Hortonworks HDP/Ambari distribution. We still have CentOS 6 systems here for support reasons and we'll probably keep these until mid-2022. Also if you need a CentOS 6 RPMS there is a repo that will provide them (the main repos have stopped) Down load it as follows:

curl https://www.getpagespeed.com/files/centos6-eol.repo --output /etc/yum.repos.d/CentOS-Base.repo

---

Can you please email me the default username/pw combo?