[ https://issues.apache.org/jira/browse/HBASE-20095?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16389309#comment-16389309 ]
Reid Chan commented on HBASE-20095: ----------------------------------- I'm afraid v6 has the same as v5, my mistake..., please review v7 again. bq. Notifications will come in serially, and sometimes we'll set that we need to update the pool. There could be multiple tasks going on at the same time. However, we shutdown the pool after the first task finishes, so we could end up killing a bunch of the queued/executing tasks before creating the new pool. Is this ok because we'll simply clean up the files on the next pass? If so, why do we need to wait for one task to finish at all? Your concern is right, so v7 add a simple latch to avoid shutdown pool after the first task finishes, and delay the pool update to the last cleaner. WDYT. > Redesign single instance pool in CleanerChore > --------------------------------------------- > > Key: HBASE-20095 > URL: https://issues.apache.org/jira/browse/HBASE-20095 > Project: HBase > Issue Type: Improvement > Reporter: Reid Chan > Assignee: Reid Chan > Priority: Critical > Attachments: HBASE-20095.master.001.patch, > HBASE-20095.master.002.patch, HBASE-20095.master.003.patch, > HBASE-20095.master.004.patch, HBASE-20095.master.005.patch, > HBASE-20095.master.006.patch, HBASE-20095.master.007.patch > > -- This message was sent by Atlassian JIRA (v7.6.3#76005)