Support Questions
Find answers, ask questions, and share your expertise

Hive metastore is unable to restart in cloudbreak new cluster

Explorer

I have a cluster provisioned via cloudbreak at hdp 2.6.5 , i am adding a new cluster at 3.1.0 hdp but my metastore, hiveserver2 are failing to start with following errors:

hivemetastore -

0: jdbc:postgresql://dpcb-cloudbreak-db-prod.> CREATE TABLE "BUCKETING_COLS" ( " SD_ID" bigint NOT NULL, "BUCKET_COL_NAME" character varying(256) DEFAULT NULL::c haracter varying, "INTEGER_IDX" bigint NOT NULL ) Error: ERROR: relation "BUCKETING_COLS" already exists (state=42P07,code=0) Closing: 0: jdbc:postgresql://dpcb-cloudbreak-db-prod.c71clklfxr8a.us-east-1.rds.amazonaws.com:3306/hive


hiveinteractive server - unable to connect to metastore which is i think due to above

is it that we cannot have two hdp versions running in cloudbreak and using the same external hive db

any pointers will be really helpful , stuck for long time


cloudbreak version: 2.9.0

2 REPLIES 2

Cloudera Employee

Hi @navdeep agarwal

Regarding the error you mentioned, it seems you are trying to use the same database for both clusters, I think Ambari need a different, empty database on the existing database instance (as I see it is an AWS RDS instance) for new cluster.

Have you tried to use a different database for your hdp 3.1 cluster?

Explorer

correct ! i guess hive of new version needs new database or you need to upgrade the database for the version , thanks for confirming.