Known issues in AM2CM
You must be aware of the known issues and limitations, the areas of impact, and workarounds in the AM2CM tool.
- CLR-78457: Unable to use Replication Manager in CDP.
- HDP Replication Manager is deprecated in Cloudera Runtime. You can use DistCp.
- BUG-123761: HDP 7.1.1.0 do not appear when selecting the HDP stack.
- HDP 7.1.1.0 is behind the paywall. Contact Cloudera support for more information on downloading.
- BUG-124410: When you are upgrading Ambari, delete Accumulo service.
- None. Accumulo is not supported on any of the Ambari 7.1.x versions.
- CDPQE-1655: After the cluster upgrade with knox sso and proxy enabled, the post upgrade Cloudera Manager tests are failing as it is going via knox sso.
- Disable Ambari SSO before migrating the cluster as Ambari specific configurations are not migrated to Cloudera Manager and you must explicitly enable SSO in Cloudera Manager after migration.
- BUG-123594: Oozie service check failure on HA cluster during EU
- If Ranger HA and/or Oozie Server HA is configured and a custom composite keytab file is being used, service checks for Ranger and Oozie will fail during the HDP 2.6 to HDP 7.1 Upgrade.
- BUG-124599: Knox stop fails intermittently.
- None. Retry the operation.
- OPSAPS-58925: Large Solr logs on a cluster can sometimes exceed the one-minute timeout for log estimation. However, after setting the JVM argument, the process continues on the agent. But the server times out after one minute of waiting for the response. When you increase the timeout, the agent is indefinitely processing the Ranger tag sync logs.
- Increase the log estimation timeout in the CMF
Java Argument:
-Dcom.cloudera.RoleLogEstimator.estimateTimeoutPerHostSeconds=180 -Dcom.cloudera.RoleLogEstimator.maxEstimateTimeoutSeconds=180 - BUG-124836: Oozie service check failure on HA cluster during EU.
- If Ranger HA and/or Oozie Server HA is configured and a custom composite keytab file is being used, service checks for Ranger and Oozie will fail during the HDP 2.6 to HDP 3.0 Upgrade.
- CDPD-20582: The HDP 2.6.5.x cluster by default has the yarn.scheduler.capacity.resource-calculator property set to org.apache.hadoop.yarn.util.resource.DefaultResourceCalculator.
- If CDP Private Cloud Base 7.1.4 is freshly installed, it has the yarn.scheduler.capacity.resource-calculator property set to org.apache.hadoop.yarn.util.resource.DominantResourceCalculator.
- CDPD-21182: Cannot create managed tables with Impala after upgrading to CDP Private Cloud Base. After setting Ranger policies and trying to create a managed table with Impala results in an AccessControlException error.
- If Impala was not part of the HDP cluster, you must manually add Impala to the cl1_hadoop group after upgrading to CDP Private Cloud Base cluster.
- CDPD-21234: After migrating to CDP Private Cloud Base, connecting to Phoenix through sqlline displays an error.
- None
- CDPD-21185: Ozone service does not start when Ozone with Ranger dependency is added to the CDP cluster after migration.
- Log in to Cloudera Manager UI
- Navigate to Clusters
- Select the Ozone service
- Go to the Configuration tab
- Search for Ozone Manager Environment Advanced Configuration Snippet (Safety Valve)
- Add Key - PARCEL_DIRNAMES and Value - CDH
- OPSAPS-59113: The CDP Private Cloud Base cluster preceives Livy2 as Livy. Hence, when interpreter.json is having the Livy2 Interpreter named configuration in it (which is cooped from the HDP cluster), it fails the zeppelin to start.
- You must not copy interpretor.json. And, you must manually configure the custom configurations as interpreter.json is not copied.
