Issues

plugins should validate regex before publishing pipeline

s3 source should read all data in the folder and subfolders

Live-info does not work for workflows

S3 Sinks broken in standalone SDK

Clicking another node before the config for the one selected loads highlights two nodes

Dates in the UI should always be displayed as human readable

Tracker should interpret and display external dataset schemas

Tracker should be able to search just * and show all datasets

Cross namespace dataset access in programs should be presented in tracker

NaiveBayesTrainer has permgen memory leak

cdap application names need to support "."

Using readless increments in concurrent transactions causes conflicts

Improve CDAP Master logging of events

Lifecycle APIs are inconsistent

Fix Javadocs under Java 8

CDAP Master and CDAP Upgrade Tool hang on incorrect Zookeeper quorum

cdap.sh start does not work in a tmux session

Port fix for CDAP-7021 to release-3.4 branch

Port fix for CDAP-7066 to release-3.4 branch

Port CDAP-6364 to release-3.4 branch

All file handles not getting closed in a CDAP Master leader-follower-leader transition

Restart of a system service does not wait for the service to stop

DynamicPartitioner should have a better way for the user to handle null values

CDAP should warn users when a partitioning column has the same name as a table column

Incorrect log statement in run record corrector

MapReduce programs fail on clusters with Phoenix enabled

Run record drop down should show AM/PM

Handle container delegation token update failures

Remove unsetting of 'mapreduce.jobhistory.address' from MapReduceRuntimeService

UI should be able to handle slow response from REST APIs

Live-info does not work for workflows

Authorization for system services

Use YarnClient instead of making an http request to Yarn webapp, for cluster metrics

Support concurrent runs of a Spark program

Not all master services respect 'master.service.memory.mb'

Application Config Integer field is being returned as float

improve cube performance when querying for more than one measure

WorkerContext.execute() is inconsistent with the Transactional API

Deprecated methods are present in the docs for ProgramClient

Scheduler Quartz jobs are not deleted upon namespace (app) delete

Add Worker Examples in cdap docs

Excessive WARN and ERROR logging when zookeeper server is down

Configurable parameters should have defaults in cdap-default.xml

Failure while listing privileges causes DefaultAuthorizationEnforcementService to fail

Attempting to delete system artifact, by specifying user namespace gives 200

Listing programs of a particular type in a nonexistent namespace returns 200

Management of entity roles (dot roles) in the Sentry extension should be a convenience, not a requirement

DatasetFramework is missing a getAllModules method

Perform deletion of log files upon namespace delete