[ https://issues.apache.org/jira/browse/HBASE-10583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13908732#comment-13908732 ]
Lars Hofhansl commented on HBASE-10583: --------------------------------------- Trying to grok the exact scenario. writeScanMetrics() is also called from ClientScanner.next(). So in this scenario we have: # less rows to scan than the cache size, *and* # we do not read all rows (if we did, we'd exhaust the scanner and writeScanMetrics() would be called in next(), right?) ? [~v.himanshu] In any way, patch looks good. > backport HBASE-8402 to 0.94 > --------------------------- > > Key: HBASE-10583 > URL: https://issues.apache.org/jira/browse/HBASE-10583 > Project: HBase > Issue Type: Bug > Reporter: Liu Shaohui > Assignee: Liu Shaohui > Attachments: HBASE-10583-v1.diff > > > see HBASE-8402 -- This message was sent by Atlassian JIRA (v6.1.5#6160)