[jira] [Commented] (HBASE-18074) HBASE-12751 dropped optimization in doMiniBatch; we take lock per mutation rather than one per batch

[ https://issues.apache.org/jira/browse/HBASE-18074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16016831#comment-16016831 ] Allan Yang commented on HBASE-18074: HBASE-17924 is another improvement for lock effi

[jira] [Commented] (HBASE-18074) HBASE-12751 dropped optimization in doMiniBatch; we take lock per mutation rather than one per batch

[ https://issues.apache.org/jira/browse/HBASE-18074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16016722#comment-16016722 ] stack commented on HBASE-18074: --- After more study, the 4-5x row locks is not true either. T

[jira] [Commented] (HBASE-18074) HBASE-12751 dropped optimization in doMiniBatch; we take lock per mutation rather than one per batch

[ https://issues.apache.org/jira/browse/HBASE-18074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16016586#comment-16016586 ] stack commented on HBASE-18074: --- Counting, we take 4-5x the rowlocks that we do in older co

[jira] [Commented] (HBASE-18074) HBASE-12751 dropped optimization in doMiniBatch; we take lock per mutation rather than one per batch

[ https://issues.apache.org/jira/browse/HBASE-18074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16016470#comment-16016470 ] stack commented on HBASE-18074: --- [~enis] Thanks for taking a look. bq. We always acquired

[jira] [Commented] (HBASE-18074) HBASE-12751 dropped optimization in doMiniBatch; we take lock per mutation rather than one per batch

[ https://issues.apache.org/jira/browse/HBASE-18074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16016220#comment-16016220 ] Enis Soztutar commented on HBASE-18074: --- We always acquired locks for every mutatio