Export - CSV (All fields)
Export - CSV (Current fields)
KUDU-2510
On-disk and provided master list error message is misleadingKUDU-2511
The SingleReplicasStayOrMove scenario of the kudu-admin-test is a bit flaky
Add ability to configure number of reactors in loadgen
Add MemorySanitizer (MSAN) build
Add a header checksum to our RPC protocolKUDU-16
Add server-side LIMIT for scanners
Clean up multi-master tech debt
Consensus/WAL/Transactions Optimizations and tests
With big rows we can easily get a scan response that is higher than our max allowed frame size
Consistent Operations
Re-visit C++ client scan retry logic
Consider switching to jemalloc
Review Java API for alter schema
Documentation that mentions file block manager should sound more ominous
Flakiness in TestAsyncKuduSession
Prevent follower from causing pre-elections when UpdateConsensus is slow
Selective predicates when selecting high number of columns burns CPU in SerializeRowBlock
Document and possibly track memory usage of maintenance threads
Add Java, Spark, and Python examples
DebugUtilTest.TestSnapshot flaky
TabletCopyClientTest.TestFailedDiskStopsClient fails with TSAN error
TSAN race in MaintenanceManager::UnregisterOp
tserver should allow starting with a small number of missing data dirs
Allow Kudu-Spark upsert API to ignore NULL column values
Add ability to configure the number of reactors in KuduClient
RpcStubTest.TestDontHandleTimedOutCalls flaky in ASAN
Spark integration should use the scanner keep-alive API
Initialize rowset iterators lazily
Cull rowsets for open-ended queries
READ_YOUR_WRITES scan with no prior operation fails
C++ client retries in a tight loop for SCANNER_EXPIRED
Replica.getRoleAsEnum exposes Protobuf generated classes in public API
Add dist-test support for the Java build
python client fails to compile in EL6 environments
Exceptions thrown by Java client have inappropriate stack traces
Local variables can overflow when serializing a 2GB message
org.apache.kudu.client.TestKuduClient.testCloseShortlyAfterOpen flaky
kudu-tool-test fails when environment has GLOG_colorlogtostderr=1
org.apache.kudu.client.TestKuduClient flaky
Redistribute tablet data when removing directories
Consider removing PB descriptors from PBC header
Backup file left by pbc edit tool will prevent the server from starting
In some rare senarios, tserver may crash with SIGSEGV while boostrapping tablets
Contribution guidelines don't mention how to add reviewers to gerrit
Add scanner keepAlive method to the java client
broken links in site/docs
Spark write to kudu, too many machines write to one tserver.
Add replica metric tracking time since there was a valid leader
Have ksck not use a single snapshot for all tablets
Tool to determine the leader master