KUDU-1074
KuduClient::ListTables() does not retryKUDU-930
Move normalization of slot refs to BinaryPredicate
Fix links to Cloudera 'beta' docs across all topicsKUDU-456
Implement AUTO_FLUSH_BACKGROUND flush modeKUDU-420
c++ client: implement HT timestamp propagation via scan tokens
Support for encoding REINSERT
contention on glog disk can block reactor threads
client does not fail over snapshot scans when querying lagging replicas
Queue should reset all_replicated_opid when becoming LEADER
[java client] check array length instead of reference equality to EMPTY_ARRAY
Set default ThreadPoolBuilder max_queue_size to a low number, say 50
API should allow introspection of current split points
Avoid holding RPC handler threads on replicas that are part of a degraded tablet
Primary Key Syntax
pushed predicates not displaying on scans web page
Dont trace ProbeStats on a per-mutation basis
Make Kudu's wire format be compatible with Impala's tuple/row layout
master should show versions of tservers, version summary
Clients should be able to apply row strings
Change "Cloudera Kudu" to "Kudu" in source files and web UIs
Don't force un-materialized slots that are referenced in predicates to be materialized
Perform partition pruning on the impala frontend
Support INSERT OVERWRITE / UPSERT
Add local_replica tool to delete a replica
Scanner timeouts aren't retried when waiting on a transaction
Update docs for ASF maven repository coordinates
ITClient test is flaky and fails on rerun
Java client scan failures are not easy to diagnose root cause
Apache Spark stucks while reading Kudu table.
Unshaded com.google.thirdparty.publicsuffix classes in kudu-client jar
File manager is broken on OS X
High lock contention on TableInfo objects
ToolTest.TestLocalReplicaTombstoneDelete fails sporadically
Propagate timestamps for scans
Handle snapshot reads that might block indefinitely
SIGSEGV if tablet copy attempt started while tablet is bootstrapping
LBM fd usage sometimes doesn't scale right
An empty delta block is orphaned when flushing or compacting with no edits
Exactly-once semantics on writes
Kudu CLI/shell for DDL and basic DML
On reads at a snapshot that touch multiple tablets, without the user setting a timestamp, use the timestamp from the first server for following scans
Proper safe time advancement by leaders and replicas
New leaders should not allow "up to date" reads until they've committed everything from previous terms
heap-use-after-free in mt-tablet-test
Handle log file deletion after a time/size
TSTabletManager startup very slow with lots of orphaned blocks
TSTabletManager: Run tablet copy client on background thread
Bootstrap can fail shortly after an alter-table
Validate table and column names
table_id is not set in tablet metadata