Qubole Release Notes - 23-Apr-2015

Release Notes for release on Apr 23, 2015

Major Enhancements



QDS now supports Dumbo

QDS now supports Dumbo. Please read the documentation for more details.



QDS supports multipart move for large files

QDS now supports multipart move for large files. Please reach out to help@qubole.com to use this feature.



QDS supports private schemas in Redshift and Postgres

QDS now supports both private and public schemas for Redshift and Postgres.



QDS supports Presto Query in Workflow

Now workflows can use Presto Query as a sub command.



All other minor changes and bug fixes are listed below



HADOOP-1

  • HAD-400: TCP connection leaks in Namenode fixed

  • Improved Hadoop daemon log cleanup and compression of older logs to save local disk space

  • HAD-379: Multiple commands waiting for cluster bringup of non-us-east region cluster would cause failures in some of them. Caused by Race condition in creating SSH tunnels that has now been fixed.



AWS

  • HAD-353: Spot instances are no longer launched in the same AWS launch group. This was causing nodes with different bids (used in 100% spot-node clusters) to fail if any one of them had a low bid.



HADOOP-2

  • HADTWO-200: Don't schedule shell command AMs when they occupy certain percent of the queue resources already. This is done to avoid deadlocks.

  • HADTWO-177: Autoscaling on MapReduce side

  • HADTWO-143: Max nodes for hadoop2 cluster now refreshable via UI



PIG

  • QBOL-1851: Pig errors should show on UI (they were previously being logged to a file on local disk)

  • QPIG-12 Pig temp files would be cleaned up automatically (space leak was causing Out-of-space issues on local disk)

  • QPIG-12: Disabling Hadoop shared cache for pig (was causing slow space leak on HDFS)



PRESTO

  • Presto upgraded to 0.95

  • PRES-373: Data Export should work on tables created via presto using CTAS

  • PRES-359: Fix configuration for clusters that are left in pending state when the query which was bringing them up is cancelled to make them usable for Presto queries.

  • PRES-342: Fix incorrect assessment of cluster start time to ensure presto daemon checks work fine



Spark

  • SPAR-202. Notebook server should start faster

  • In Yarn-Client mode, Shell command AM memory should be dynamically set based on Spark driver memory (to prevent OOM issues)

  • Allow avro Hive Table access via SparkSql



QDS

  • UI-1627 - On creating new cluster the region and aws availability zones mismatch

  • UI-1635 - Analyze results display fix for multiple spaces

  • UI-1471: Overview: Selecting to and from as today does not show commands run today in charts

  • UI-1615: Fixed white space issues on cluster QCUH metrics

  • UI-1444: Fixed window resize issue with composer

  • UI-1477: Fix for Query tagging inconsistencies and History filter tag search issues.

  • UI-1048: Workflow delete subcommand bug

  • UI-1538: Add Qubole and App log separation for Shell Commands

  • UI-1520: Reset cluster labels list on edit mode

  • UI-1484: Prevent new line and empty inputs in comments. Show error message in UI for the same

  • UI-1402-switching cluster type does not filter the node types

  • UI-1583 Fix for 500 internal error when user is not associated with any account.

  • SCHED-55:Fixes cloning of a owner while cloning a periodic job

  • QBOL-3992: special handling of results for accounts which has result encryption enabled

  • UI-1173 Fix javascript error in cpl in cases where current_qbol_user is not available

  • UI-1173 Fix javascript error to load cluster info when the user doesn't have read permissions on Account but has read permissions on Cluster.

  • UI-1274 Escape values that are displayed in Templates UI

  • QBOL-3790 Fix for disappearing Modify action on Groups tab in control panel.

  • UI-1174 Check for the right perms while killing a job

  • QBOL-3891 Fixed the way deleted users and roles are handled.

  • UI-1238 UI-1272 UI-1273 UI-1274 XSS fixes - Escaping output properly

  • UI-1459 Use roles and groups for cluster management

  • QBOL-3918: fix in csv and making json output as default

  • UI-1502: Overview - show single data value when range from-to is same

  • UI-1475: Alphabetic ordering of Users in analyze query history search filter

  • Result not loading issue fix

  • QBOL-3878. If logs are more than 100K then we truncate it and show only first and last 10K. In such cases we fetch all the Job Tracker URLs in the truncated text and show them.

  • UI-1203 - Zendesk Integration

  • Persist interpreter confs on s3

  • Disable join_account on signup feature

  • UI-1495 Change cancel command color code

 

Have more questions? Submit a request

Comments

Powered by Zendesk