Install and Upgrade Notes

Beginning with Cloudera Manager and CDH 6.3.3, downloading new versions of these products will require a valid Cloudera Enterprise license file, and/or a username and password obtained from Cloudera. All Cloudera Manager package, CDH parcel and CDH package repositories now require authentication with valid credentials to access any version numbered 6.3.3 or later. For more information on using these credentials, see the documentation below.

Beginning with CDH 6.3.3 (and CDP Data Center 7.0), Cloudera Express is no longer available. Upgrades to Cloudera Manager or CDH 6.3.3 and higher are not supported when running Cloudera Express. A valid Cloudera Enterprise or CDP Data Center license must be in place before upgrading to Cloudera Manager 6.3.3 or 7.x or the upgrade will not be completed.

Downgrading from Cloudera Enterprise license to Cloudera Express license is also no longer supported in Cloudera Manager 6.3.3 and higher.

```
2019-07-18 14:54:52,483 ERROR [main] org.apache.hadoop.mapreduce.v2.app.MRAppMaster: Error starting MRAppMaster
java.lang.NoSuchMethodError: sun.nio.ch.DirectBuffer.cleaner()Lsun/misc/Cleaner;
    at org.apache.hadoop.crypto.CryptoStreamUtils.freeDB(CryptoStreamUtils.java:41)
    at org.apache.hadoop.crypto.CryptoInputStream.freeBuffers(CryptoInputStream.java:687)
    at org.apache.hadoop.crypto.CryptoInputStream.close(CryptoInputStream.java:320)
    at java.base/java.io.FilterInputStream.close(FilterInputStream.java:180)
```

Cloudera Bug: CDH-81350

After upgrading to Cloudera Manager 6.1 or 6.2, if the /tmp directory is mounted using the noexec flag, then starting or restarting YARN NodeManagers will fail for all CDH versions. Note that this change may cause configuration staleness on Cloudera Manager when upgrading to 6.3.0 for clusters running YARN.

Cloudera Bug: OPSAPS-50253

Cloudera Bug: OPSAPS-47436

Backup and Disaster Recovery (BDR) HDFS and Hive Replications will fail when replicating from secured (Kerberized) source clusters to destination clusters that have been upgraded to Cloudera Manager 6.1.0.

This also affects new installations of Cloudera Manager 6.1.0 on the destination cluster if an admin restarts the Cloudera Manager service.

Products affected: Cloudera Manager Backup and Disaster Recovery in a secure (Kerberized) environment

Releases affected: Cloudera Manager 6.1.0 (when used as the destination cluster of HDFS and/or Hive replication)

Users affected: Customers using HDFS or Hive Replication

Severity (Low/Medium/High): High

Root Cause and Impact:

In HDFS and Hive Replication, Cloudera Manager first runs a process on the destination cluster to verify if the replication is possible. Due to a bug, the source cluster is treated as an insecure (non-kerberized) cluster. As a result, replication fails.

You will see the exception javax.security.sasl.SaslException: GSS initiate failed [Caused by GSSException: No valid credentials provided (Mechanism level: Fail to create credential. (63) - No service creds)] in the process stderr logs.

Immediate action required: If you use BDR, do not upgrade a destination cluster to Cloudera Manager 6.1.0. Upgrade to Cloudera Manager 6.1.1 or higher when it becomes available.

If you have already upgraded your destination cluster to Cloudera Manager to 6.1.0, use the following workaround:

1. For an existing HDFS or Hive replication schedule, select Actions > Edit Configuration.
2. Save the schedule.

Please note that you will need to edit only one schedule even if you have multiple schedules.

Note: This workaround is not persistent. That is, if you restart the Cloudera Manager service, you must repeat the above workaround.

Cloudera Issue: OPSAPS-48865

Fixed in Cloudera Manager 6.1.1

You cannot upgrade to Cloudera Manager or CDH 6.0.0 from Cloudera Manager or CDH 5.15 or 5.16.

For details about the Cloudera Manager Enable Direct SQL option, refer to Hive Metastore Database.

You cannot upgrade to any Cloudera Manager or CDH 6.x general release from the Cloudera Manager or CDH 6.0 Beta release.

Use of Cloudera Express is limited to a total of 100 hosts running CDH6.0 or later across all environments used by an organization..

Affected Versions: CM 6.1 and higher

Cloudera Issue: OPSAPS-46868

Cloudera Data Science Workbench is not supported with Cloudera Enterprise 6.0.x. Cloudera Data Science Workbench 1.5.0 (and higher) is supported with Cloudera Manager 6.1.x (and higher) and CDH 6.1.x (and higher).

Due to the Sentry and Impala fine grained privileges feature in 5.16.0, if a role has the SELECT or INSERT privilege on an object in Impala before upgrading to CDH 5.16.0, that role will automatically get the REFRESH privilege during the upgrade.

If you are installing or upgrading to CDH 6.0.0 and using the PostgreSQL database for the Hue database, you must install psycopg2 2.5.4 or higher on all Hue hosts. SeeInstalling the psycopg2 Python package.

Cloudera Issue: OPSAPS-47080

Failed to execute command Bootstrap Solr Collections on service Solr

Workaround: If there are no collections added to it, remove the Solr service from your cluster before you start the upgrade.

Affected Versions: CDH 6.0.0, 6.0.1, 6.1.0, 6.1.1, 6.2.0, 6.2.1, 6.3.0, 6.3.1, 6.3.2

Fixed Version: CDH 6.3.3

Cloudera Issue: CDH-82042

The CDH upgrade process fails to delete Solr data from HDFS and the recreated collections fail to be initialized due to the existing indexes.

Affected Versions: CDH 6.0.0

Fixed Versions: Cloudera Manager 6.0.1

Cloudera Issue: OPSAPS-47502

When you install CDH with packages from a custom repository, ensure that the version of CDH you select for Select the version of CDH matches the version of CDH for the custom repository. Selecting the CDH version and specifying a custom repository are done during the Select Repository stage of installation.

If the versions do not match, installation fails.

Affected Versions: Cloudera Manager 6.x

Fixed Versions: N/A

Apache Issue: N/A

Cloudera Issue: OPSAPS-45703

Sqoop includes two connectors, one for Teradata and one for Netezza. The connectors are released in separate parcels and tarballs and can be installed in Cloudera Manager or manually. The versioning of the connectors takes the form <connector_version>c<major_cdh_version>. For example, 1.6c5 refers to the connector 1.6 for CDH 5. The manifest files do not prohibit installing the CDH 5 connectors on CDH 6, but they are not compatible with CDH 6.

If you have CDH 5 connectors installed, they will not be automatically upgraded during the CDH upgrade, and they are not compatible with CDH 6, so they should be uninstalled before the upgrade. Keeping the CDH 5 connectors will not cause the upgrade to fail, but instead will cause a failure to occur during Sqoop runtime. Cloudera will release the connectors for CDH 6 at a later time.

For more information about the Teradata and Netezza connectors, go to Cloudera Enterprise Connector Documentation and choose the connector and version to see the documentation for your connector.

New fail-fast checks for unsupported options were introduced in CDH 6. Users should check the jobs stored in their Sqoop metastore and remove all unsupported options. Some unsupported options were silently ignored in earlier CDH versions during upgrades, but in CDH 6, the same options fail instantly. See the following JIRAs in Apache Sqoop Incompatible Changes:

Changes in logical types cause code generated in Avro with CDH 6 to differ from code generated in Avro with CDH 5. This means that old generated code will not necessarily work in CDH 6. Cloudera recommends that users regenerate their generated Avro code when upgrading.

Parquet packages and the project’s group ID were renamed, and some of the class methods were removed.

If you directly consumes the Parquet API instead of using Parquet through a CDH component, your need to update and recompile your code. See Parquet API Change for details of the changes.

When doing a rolling upgrade from a CDH 6 version to a higher version, do not do any replication peer configuration changes. This includes removing a peer, adding a peer, and changing the configuration on a peer.

Due to the change of offset storage from ZooKeeper to Kafka in the CDH 5.8 Flume Kafka client, data might not be consumed by the Flume agents, or might be duplicated (if kafka.auto.offset.reset=smallest) during an upgrade to CDH 5.8.

Cloudera Issue: TSB-173

If your cluster has Spark 2.0 or Spark 2.1 installed and you want to upgrade to CDH 5.13 or higher, you must first upgrade to Spark 2.1 release 2 or later before upgrading CDH. To install these versions of Spark, do the following before running the CDH Upgrade Wizard:

Affected versions: CDH 5.13.0 and higher

Cloudera Issue: CDH-56775

Before upgrading to CDH 5.13 or higher, you may need to increase the size of the Java heap for Sentry. A warning will be displayed during upgrade, but it its the user's responsibility to ensure this setting is adjusted properly before proceeding. See Performance Guidelines.

Affected versions: CDH 5.13 or higher

Cloudera Issue: OPSAPS-42541

2017-06-08 17:43:37,173 WARN [Socket Reader #1 for port 41187]
org.apache.hadoop.ipc.Server: Unable to read call parameters for client 10.17.242.22on
connection protocol org.apache.hadoop.mapred.TaskUmbilicalProtocol for rpcKind
RPC_WRITABLE
java.lang.ArrayIndexOutOfBoundsException: 23
   at
...

This error could cause the task and the job to fail.

Workaround:

Avoid performing a rolling upgrade to CDH 5.11.0 or CDH 5.11.1 from CDH 5.10.x or lower. Instead, skip CDH 5.11.0 and CDH 5.11.1 if you are performing a rolling upgrade, and upgrade to CDH 5.12 or higher, or CDH 5.11.2 or higher when the release becomes available.

Cloudera Issue: DOCS-2384, TSB-241

After upgrading to 5.7 or higher, you might see a reduced Java heap maximum on Impala Catalog Server due to a change in its default value. Upgrading from Cloudera Manager lower than 5.7 to Cloudera Manager 5.8.2 no longer causes any effective change in the Impala Catalog Server Java Heap size.

When upgrading from Cloudera Manager 5.7 or later to Cloudera Manager 5.8.2, if the Impala Catalog Server Java Heap Size is set at the default (4GB), it is automatically changed to either 1/4 of the physical RAM on that host, or 32GB, whichever is lower. This can result in a higher or a lower heap, which could cause additional resource contention or out of memory errors, respectively.

Cloudera Issue: OPSAPS-34039