03-19-2019 11:42 AM
We are planning to setup Cloudera 6.1.x Hadoop cluster spanning across two data centers. We are following "Cloudera Enterprise Reference Architecture for Bare Metal Deployments" and planned services distribution as mentioned in "Appendix A: Spanning Multiple Data Centers" section.
We also plan to have a node configured as both Edge Node & Cloudera Management Node. We will be using Oracle 12c as Metastore Database for Cloudera Manager & CDH Services and plan to install it on this Cloudera Management Node. We will also have additional Edge node on other data center.
How can we achieve High Availability for Oracle Metastore Database and not being a single point of failure?
03-19-2019 03:33 PM
Hi @Anishsk10 ,
Please take a look at this doc:
As the doc points out, please install each instance of the metastore on a separate cluster host, to maximize high availability. Also, Hive metastore HA requires a database that is also highly available, so make sure that your Oracle database is installed with replication in active-active mode. Refer to Oracle documentation to configure it correctly.
Thanks and hope this helps,
03-20-2019 01:10 PM