> On Sept. 13, 2016, 6:58 a.m., ramkrishna vasudevan wrote: > > hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HeapMemStoreLAB.java, > > line 115 > > <https://reviews.apache.org/r/51785/diff/2/?file=1496671#file1496671line115> > > > > Because in merge you don't close the segment you put it back here. So > > they are for sure from the chunkPool only?
The pooledChunkQueue is not empty only if there is chunk pool enabled, otherwise the queue is NULL and can not be used. This queue is used only to keep chunks from the pool. > On Sept. 13, 2016, 6:58 a.m., ramkrishna vasudevan wrote: > > hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/MemStoreCompactorIterator.java, > > line 89 > > <https://reviews.apache.org/r/51785/diff/2/?file=1496674#file1496674line89> > > > > Intentional change? So the other test case that was added due to a bug > > here is still passing? Then fine. Yes, this is intentional change. I promissed to Anoop to fix this later, when 14921 was commited. This is not related to this patch. All tests pass. > On Sept. 13, 2016, 6:58 a.m., ramkrishna vasudevan wrote: > > hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/SegmentFactory.java, > > line 98 > > <https://reviews.apache.org/r/51785/diff/2/?file=1496675#file1496675line98> > > > > You know for sure that they are from pool? Or just any segment from > > the MSLAB? The chunks that are not from the pool are not kept in this pooledChunkQueue. So this operations is relevant only for chunks from the pull. - Anastasia ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/51785/#review148665 ----------------------------------------------------------- On Sept. 15, 2016, 4:49 p.m., Anastasia Braginsky wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/51785/ > ----------------------------------------------------------- > > (Updated Sept. 15, 2016, 4:49 p.m.) > > > Review request for hbase. > > > Repository: hbase-git > > > Description > ------- > > This is a step toward final compacting memstore that allowes two modes of > work: index-compaction and data-compaction. > > The index-compaction means that when the new segment is pushed into the > pipeline, it is flattened and probably merged with old segments in the > pipeline. The new merge "feature" induces no data-copy-compaction and no > speculative SQM scan. > The compacting memstore of the data-compaction type means the usage of the > data-copy-compaction. > > > Diffs > ----- > > src/Makefile.am 2dd7913a1477f3c3560be4e2c1450b93fb3afc78 > src/launcher/default_executor.cpp PRE-CREATION > > Diff: https://reviews.apache.org/r/51785/diff/ > > > Testing > ------- > > > Thanks, > > Anastasia Braginsky > >
