Hive Metastore Database Host must be configured when Hive metastore is configured to use a database

[#1]

Trying to install cloudera core on a single node (localhost)  After the host checker screen i get a failure:

Command Context Status Started at Ended at

Deploy Client Configuration	Cluster 1	Finished	Jul 23, 2015 10:08:31 PM UTC	Jul 23, 2015 10:08:31 PM UTC
Command failed to run because service Hive has invalid configuration. Review and correct its configuration. First error: Hive Metastore Database Host must be configured when Hive metastore is configured to use a database other than Derby.

[#2 · Accepted solution]

Did the page where you test your Hive database connection work? Did you make any changes to database configuration since running that test?

This error happens because you have invalid configuration for Hive, which is described in the error message you posted. Hive has no idea how to talk to its database because the configuration for the database host is missing.

[#3]

I am attempting to do a fresh install of Cloudera Core Services using the Cloudera Manager. Do I need to just start a fresh instance again and start all over again for some reason ?

[#4]

I'm also experiencing this issue. Its a fresh install with all defaults selected and no issues so far in the installation  outside of here.

[#5]

I couldnt get it work, so I had to remove and uninstall everything and start from scratch and it worked without any issues. Nothing different was done.

[#6]

I wanted to share my experience in case it helped a few others out there in the community.  In my case I didn't have a Spark gateway on the same node as HS2 (a mis-click on my part).  The issue was that the installer would not retry in a proper way (by my standards) even though I went back and fixed the role assignment.  All attempts to make the change result in the same error on the same step while trying to finish my cluster install.  I haven't disected the installtion but at somepoint prior to the 'final' steps it has already created the cluster layout and configuration in CM.  And this can not be changed through the install wizard.  I opened CM in another tab.  I could see my brand new cluster with all services and a big error icon.  I corrected my issue here, making sure the original assignment matched and with the Spark gateway on the HS2 node as well.  Then I went back to the other tab with the install wizard and hit 'Retry'. I proceeded and finished my install.  This is how you need to handle any errors that come up regarding the configuration and it has already been committed to the CM database.