These Release Notes summarize release-specific changes:
Support For HDP 2.6 Update Which Includes LLAP Performance Improvements
HDP 2.6 clusters use newer HDP 2.6 and Ambari 2.5 releases. For more information refer to System Information.
Introduces support for enabling a Hortonworks Flex Support Subscription for HDCloud clusters and the cloud controller. For information on how to enable Flex Subscription in HDCloud, refer to Flex Subscriptions. If you are looking for more general information about Flex Support Subscriptions, visit the Hortonworks Support page at https://hortonworks.com/services/support/enterprise/.
Central Protected Gateway
The protected gateway (which was previously per-cluster) is now available centrally on the controller instance, providing a single-access point for all clusters and cluster resources. For more information, refer to Protected Gateway.
Introduces the ability to use your own SSL certificate in place of the self-signed certificate installed by default on the protected gateway. For more information, refer to Using Your Own SSL Certificate.
Encryption for Master, Worker, and Compute Node EBS Volumes
When creating a cluster, you can optionally select to have the cloud controller configure encryption for the EBS volumes attached to the master, worker, and compute nodes. If you choose to use this option, the following types of data are encrypted:
- Data at rest inside the volume
- All data moving between the volume and the instance
- All snapshots created from the volume
If the issue that you are experiencing is not listed below, check the Troubleshooting documentation.
Problem (BUG-81014): Cluster Template May Not Populate Some Parameters Correctly
When creating a cluster from a previously saved template, storage-related parameters may not be pre-populated correctly.
Workaround: Before creating a cluster, validate that the storage-related parameters were populated correctly. Alternatively, you can copy the CLI JSON output, paste it into a JSON file, and create clusters using HDC CLI.
Problem (BUG-77857): Clock on the Cluster Details Page Shows 0:00
In some circumstances, the clock on the cluster details page may show 0:00 as the running time, even if more time has elapsed.
Workaround: Navigate back to the cluster dashboard. This will refresh the clock.
Problem (BUG-82018): Cluster Notification Email Field Doesn't Accept a Dash
On the Confirm Create Cluster pop-up, the email field under
(Optional) Send notification when cluster is complete does not accept email addresses which include a dash. For example
email@example.com will not be accepted.
Workaround: This issue will be fixed in a future release. Meanwhile, use an email address which does not include a dash.
Problem (BUG-73946): Cluster History Doesn't Show HDP Version for Some Clusters
The History page may not show HDP Version for some clusters.
Problem (BUG-67740): CLI Option 'generate-cli-shared-skeleton' Doesn't Work
hdc create-cluster generate-cli-shared-skeleton doesn't work.
Workaround: Ignore this option. Use
Problem (BUG-69444): Connection Error After Password Reset
After reseting your password, you will get the following error when running
hdc configure to update the password in the CLI configuration file:
Error while connecting to https://ec2-35-156-72-90.eu-central-1.compute.amazonaws.com/identity/oauth/authorize as user: firstname.lastname@example.org, please check your username and password. (406 Not Acceptable)
- Edit the CLI configuration file directly with sudo. For example:
sudo vi /home/cloudbreak/.hdc/config
- Change the owner of the .hdc directory manually. For example:
sudo chown -R cloudbreak:cloudbreak /home/cloudbreak/.hdc/
- After performing these steps,
hdc configureand other CLI command will start working.
Hive 2 LLAP
Problem (BUG-90246): alter table...update statistics query failure due to JDODataStoreException
When you run
alter table...update statistics query and certain conditions are met (the table includes boolean columns, bitvectors are on, metastore is postgres), you may get the following error:
ERROR [HiveServer2-Background-Pool: Thread-199] metastore.RetryingHMSHandler: Retrying HMSHandler after 2000 ms (attempt 1 of 10) with error: javax.jdo.JDODataStoreException: Insert of object "org.apache.hadoop.hive.metastore.model.MTableColumnStatistics@10ac18ff" using statement "INSERT INTO "TAB_COL_STATS" ("CS_ID","AVG_COL_LEN","BIT_VECTOR","COLUMN_NAME","COLUMN_TYPE","DB_NAME","BIG_DECIMAL_HIGH_VALUE","BIG_DECIMAL_LOW_VALUE","DOUBLE_HIGH_VALUE","DOUBLE_LOW_VALUE","LAST_ANALYZED","LONG_HIGH_VALUE","LONG_LOW_VALUE","MAX_COL_LEN","NUM_DISTINCTS","NUM_FALSES","NUM_NULLS","NUM_TRUES","TBL_ID","TABLE_NAME") VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)" failed : ERROR: column "BIT_VECTOR" is of type bytea but expression is of type oid Hint: You will need to rewrite or cast the expression. Position: 346 at org.datanucleus.api.jdo.NucleusJDOHelper.getJDOExceptionForNucleusException(NucleusJDOHelper.java:54
|BUG-80709||When creating a cluster, the "Cluster Components" gateway access is checked by default.|
|BUG-81255||The protected gateway (which was previously per-cluster) is now available centrally on the controller instance. For more information, refer to [Protected Gateway](security-network.md#protected-gateway).|
|BUG-77211||Fixed the cluster cloning issues where some parameters, such as Cluster Type and parameters related to autoscaling, were not be pre-populated correctly.|
|BUG-76991||Fixed issues with Ambari Hive View 2.0.|
|BUG-78871||Updated image for HDP 2.6 (Cloud). See System Information for more information.|
HDP 2.6 clusters are built using HDP 2.6 and Ambari 2.5. Refer to:
HDP 2.5 clusters are built using HDP 2.5 and Ambari 2.4. Refer to: