Upgrading Oracle Big Data Appliance from 2.5.0 to 3.* Release Fails during Step 7 on Addyarn (Doc ID 1900971.1)

Last updated on AUGUST 21, 2014

Applies to:

Big Data Appliance Integrated Software - Version 2.5.0 to 3.0.1 [Release 2.5 to 3.0]
Linux x86-64

Symptoms

During upgrade from v2.5.0 to v3.* on Oracle Big Data Appliance for secure clusters (Kerberos Enabled) which have HBase service with secure authentication and authorization off/disabled it is possible to get the following errors during upgrade on step 7:

************************************
Error [10519]: (//<BDANode01>//Stage[main]/Hadoop::Addyarn/Exec[add_yarn]/returns) change from notrun to 0 failed: /opt/oracle/BDAMammoth/bdaconfig/tmp/addyarn.sh &> /opt/oracle/BDAMammoth/bdaconfig/tmp/addyarn.out returned 1 instead of one of [0] at /opt/oracle/BDAMammoth/puppet/modules/hadoop/manifests/addyarn.pp:46
************************************

 

As UpgradeCluster call failed since HBaseUpgrade failed all subsequent upgrade calls for other CDH services have been suspended. Rerunning 'mammoth -p' will not re-execute UpgradeCluster call and will just try to restart the services which result in below failures since metadata hasn't been updated for Hive, Oozie, etc.

 

Note: At this point, mammoth -p should not be run again until the solution documented in the Solution section of the note is followed.

 

If mammoth -p is run again the next error encountered will be:

************************************
Error [10477]: (//<BDANode01>//Stage[main]/Hadoop::Upgradeactions/Exec[confighive]/returns) change from notrun to 0 failed: /opt/oracle/BDAMammoth/bdaconfig/tmp/confighive.sh &> /opt/oracle/BDAMammoth/bdaconfig/tmp/confighive.out returned 1 instead of one of [0] at /opt/oracle/BDAMammoth/puppet/modules/hadoop/manifests/upgradeactions.pp:17
************************************

 

If Hive Metastore Database Schema is then upgraded in Cloudera Manager the next error received on mammoth -p is:

************************************
Error [10477]: (//<BDANode01>//Stage[main]/Hadoop::Startcloudera/Exec[start_cloudera_services]/returns) change from notrun to 0 failed: /opt/oracle/BDAMammoth/bdaconfig/tmp/startcloudera.sh &> /opt/oracle/BDAMammoth/bdaconfig/tmp/startcloudera_1402949245.out returned 1 instead of one of [0] at /opt/oracle/BDAMammoth/puppet/modules/hadoop/manifests/startcloudera.pp:24
************************************

Checking in CM will show Oozie service in BAD health, as Oozie Database hasn't been upgraded.

 

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms