[ https://issues.apache.org/jira/browse/HBASE-18294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16253139#comment-16253139 ]
Eshcar Hillel commented on HBASE-18294: --------------------------------------- Attaching a new patch with region level counters and thresholds. Per-region flush is based on heap/off-heap size and not based on data size. Still missing the way to set the off-heap threshold so currently it is equal to the on-heap threshold. Will change this once I figure out what is the best way. I tend towards adding a configuration property. I also think similar changes are required at the RS level. Meanwhile current patch is in RB. Thanks for the comments so far, please keep reviewing. > Reduce global heap pressure: flush based on heap occupancy > ---------------------------------------------------------- > > Key: HBASE-18294 > URL: https://issues.apache.org/jira/browse/HBASE-18294 > Project: HBase > Issue Type: Bug > Affects Versions: 3.0.0 > Reporter: Eshcar Hillel > Assignee: Eshcar Hillel > Attachments: HBASE-18294.01.patch, HBASE-18294.02.patch > > > A region is flushed if its memory component exceed a threshold (default size > is 128MB). > A flush policy decides whether to flush a store by comparing the size of the > store to another threshold (that can be configured with > hbase.hregion.percolumnfamilyflush.size.lower.bound). > Currently the implementation (in both cases) compares the data size > (key-value only) to the threshold where it should compare the heap size > (which includes index size, and metadata). -- This message was sent by Atlassian JIRA (v6.4.14#64029)