I've started work on the first release candidate of the final 1.3 release, 1.3.3. We have been maintaining an internal fork of 1.3 where I work, due to lack of RM attention to the 1.3 code line. The first task will be committing the relevant changes to branch-1.3. So far I have 35 commits queued for testing (not pushed up yet). This is not the complete list as I have not finished yet. All of the changes are bug fixes or operational improvements we've found of use and conform to the expectations of our compatibility policy, and confirming that or getting consensus on worthy exceptions will be part of the RC process for 1.3.3.
HBASE-20554 "WALs outstanding" message from CleanerChore is noisy HBASE-20523 PE tool should support configuring client side buffering sizes (Ram) HBASE-20404 Fixes to CleanChore correctness and operability HBASE-20352 [Chore] Backport HBASE-18309 (Support multi threads in CleanerChore) to branch-1 HBASE-20141 Fix TooManyFiles exception when RefreshingChannels HBASE-20052 TestRegionOpen#testNonExistentRegionReplica fails due to NPE HBASE-19391 Calling HRegion#initializeRegionInternals from a region replica can still re-create a region directory HBASE-19905 ReplicationSyncUp tool will not exit if a peer replication is disabled HBASE-19553 Old replica regions should be cleared from AM memory after primary region split or mergeHBASE-19163 "Maximum lock count exceeded" from region server's batch processing HBASE-19358 Improve the stability of splitting log when do fail over HBASE-19364 Truncate_preserve fails with table when replica region > 1 HBASE-19435 Reopen Files for ClosedChannelException in BucketCache HBASE-18786 FileNotFoundException should not be silently handled for primary region replicas HBASE-18248 Warn if monitored RPC task has been tied up beyond a configurable threshold HBASE-15134 Add visibility into Flush and Compaction queues HBASE-17991 Add more details about compaction queue on /dump (Guangxu Cheng) HBASE-18058 Zookeeper retry sleep time should have an upper limit (Allan Yang) HBASE-17924 Consider sorting the row order when processing multi() ops before taking rowlocks (Allan Yang) HBASE-17118 StoreScanner leaked in KeyValueHeap (binlijin) HBASE-17731 Fractional latency reporting in MultiThreadedAction HBASE-19551 hbck -boundaries doesn't work correctly (Toshihiro Suzuki) Amend HBASE-18830 TestCanaryTool does not check Canary monitor's error code HBASE-18762 Canary sink type cast error HBASE 17959 Canary timeout should be configurable on a per-table basis HBASE-17965 Canary tool should print the regionserver name on failure HBASE-17930 Avoid using Canary.sniff in HBaseTestingUtility HBASE-16399 Provide an API to get list of failed regions and servername in Canary (Vishal Khandelwal) HBASE-15727 Canary Tool for Zookeeper (churro morales) HBASE-16091 Canary takes lot more time when there are delete markers in the table (Vishal Khandelwal) HBASE-17912 Avoid major compactions on region server startup HBASE-17816 HRegion#mutateRowWithLocks should update writeRequestCount metric HBASE-17798 RpcServer.Listener.Reader can abort due to CancelledKeyException (Guangxu Cheng) HBASE-17072 CPU usage starts to climb up to 90-100% when using G1GC HBASE-17205 Add a metric for the duration of region in transition HBASE-16947 Some improvements for DumpReplicationQueues tool HBASE-19816 Refresh repliation sinks on UnknownHostException Please hold off on committing to branch-1.3 until this work is completed. I don't think there is much risk of that, but you've been advised, thanks in advance. -- Best regards, Andrew Words like orphans lost among the crosstalk, meaning torn from truth's decrepit hands - A23, Crosstalk
