Thanks, JM. Those are some good tests. James
On Sep 13, 2013, at 9:04 AM, Jean-Marc Spaggiari <[email protected]> wrote: > Here is the performances comparison between 0.94.11 and 0.94.12 > (Lines/minutes or Lines/seconds): > > 0.94.11 0.94.12 > org.apache.hadoop.hbase.PerformanceEvaluation$FilteredScanTest 10.17 10.37 > org.apache.hadoop.hbase.PerformanceEvaluation$RandomReadTest 810.58 833.65 > org.apache.hadoop.hbase.PerformanceEvaluation$RandomSeekScanTest 255.71 > 251.57 org.apache.hadoop.hbase.PerformanceEvaluation$RandomWriteTest > 25682.11 25243.45 > org.apache.hadoop.hbase.PerformanceEvaluation$RandomScanWithRange10Test > 28715.29 29399.64 > org.apache.hadoop.hbase.PerformanceEvaluation$RandomScanWithRange100Test > 18022.39 17789.79 > org.apache.hadoop.hbase.PerformanceEvaluation$RandomScanWithRange1000Test > 2346.85 2359.45 > org.apache.hadoop.hbase.PerformanceEvaluation$SequentialReadTest 2875.83 > 2877.06 org.apache.hadoop.hbase.PerformanceEvaluation$SequentialWriteTest > 26693.23 26576.31 > So it's pretty stable between the 2 versions. > > > > > 2013/9/13 lars hofhansl <[email protected]> > >> - Checked the tarball, the documentation >> >> - Loaded 100m rows, flushes, compacted, scanned >> - Did my usual tight loop performance tests >> - Used client to connected to some older secure clusters (there was an >> issue with that at some point during the 0.94.12 cycle) >> - Double checked with Phoenix folks that no needed low level APIs were >> changed >> >> >> All good. +1 >> >> >> Other committers, please have a quick look if you can. >> >> Thanks. >> >> -- Lars >> >> >> >> ________________________________ >> From: lars hofhansl <[email protected]> >> To: hbase-dev <[email protected]> >> Sent: Wednesday, September 11, 2013 1:15 PM >> Subject: [VOTE] The 1st hbase 0.94.12 release candidate is available for >> download >> >> >> The 1st 0.94.12 RC is available for download at >> http://people.apache.org/~larsh/hbase-0.94.12-rc0/ >> Signed with my code signing key: C7CFE328 >> >> As before 0.94.12 is mostly a bug fix release and has 33 issues resolved >> against it: >> [HBASE-9277] - REST should use listTableNames to list tables >> [HBASE-9279] - Thrift should use listTableNames to list tables >> [HBASE-7709] - Infinite loop possible in Master/Master replication >> [HBASE-7954] - Fix the retrying logic of memstore flushes to avoid >> extra sleep >> [HBASE-8760] - possible loss of data in snapshot taken after region >> split >> [HBASE-8930] - Filter evaluates KVs outside requested columns >> [HBASE-9167] - ServerCallable retries just once if timeout is not >> integer.max >> [HBASE-9182] - Allow non-admin users to list all table names >> [HBASE-9195] - Fix TestFSHDFSUtils against java7 test re-ordering >> [HBASE-9207] - An Offline SplitParent Region can be assigned breaking >> split references >> [HBASE-9231] - Multipage book is generated to the wrong location >> [HBASE-9233] - isTableAvailable() may be stuck if an offline parent >> was never assigned >> [HBASE-9252] - HConnectionManager#getZooKeeperWatcher() should be >> deprecated in 0.94 >> [HBASE-9256] - HBaseClient#setupIOStreams should handle all exceptions >> [HBASE-9266] - Javadoc: Document that HBaseAdmin.flush(...) is >> synchronous >> [HBASE-9286] - [0.94] ageOfLastShippedOp replication metric doesn't >> update if the slave regionserver is stalled >> [HBASE-9301] - Default hbase.dynamic.jars.dir to hbase.rootdir/jars >> [HBASE-9303] - Snapshot restore of table which splits after snapshot >> was taken encounters 'Region is not online' >> [HBASE-9326] - ServerName is created using getLocalSocketAddress, >> breaks binding to the wildcard address. Revert HBASE-8640 >> [HBASE-9329] - SnapshotManager should check for directory existance >> before throwing a warning. >> [HBASE-9344] - RegionServer not shutting down upon KeeperException in >> open region >> [HBASE-9356] - [0.94] SecureServer.INSECURE_VERSIONS is declared >> incorrectly >> [HBASE-9397] - Snapshots with the same name are allowed to proceed >> concurrently >> [HBASE-9415] - In rpcServer, replicationQueue is initialized with the >> max queue size instead of the max queue lenght >> [HBASE-9428] - Regex filters are at least an order of magnitude slower >> since 0.94.3 >> [HBASE-9429] - Add back MetaScanner.allTableRegions(Configuration >> conf,byte[] tablename,boolean offlined) method >> [HBASE-9432] - Backport HBASE-8781 to 0.94 >> [HBASE-9448] - [0.94] Shell needs to fall back after HBASE-9182 if >> talking to older servers >> [HBASE-9455] - Port HBASE-7113 'TestGzipFilter is flaky with jdk1.7' >> to 0.94 >> [HBASE-9243] - Add more useful statistics in the HFile tool >> [HBASE-9314] - Dropping a table always prints a >> TableInfoMissingException in the master log >> [HBASE-9377] - Backport HBASE- 9208 "ReplicationLogCleaner slow at >> large scale" >> [HBASE-9287] - TestCatalogTracker depends on the execution order >> >> The list of changes is also available here: >> https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12310753&version=12324790 >> >> Here're the jenkins runs for this RC: >> https://builds.apache.org/job/HBase-0.94.12/4/ and >> https://builds.apache.org/job/HBase-0.94.12-security/5/ >> >> Please try out the RC, check out the doc, take it for a spin, etc, and >> vote +1/-1 by September 18th on whether we should release this as 0.94.12. >> >> Thanks. >> >> -- Lars >>
