[ https://issues.apache.org/jira/browse/HBASE-21372?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16661213#comment-16661213 ]
Mike Drob commented on HBASE-21372: ----------------------------------- In TransitRegionStateProcedure we check retries >= max which will never be true now, so this is actually saying retry forever. Should document that behaviour, even if MAX_INT is already basically forever. > Set hbase.assignment.maximum.attempts to Long.MAX > ------------------------------------------------- > > Key: HBASE-21372 > URL: https://issues.apache.org/jira/browse/HBASE-21372 > Project: HBase > Issue Type: Sub-task > Components: amv2 > Reporter: stack > Assignee: stack > Priority: Major > Attachments: HBASE-21372.branch-2.1.001.patch > > > From parent issue, [~allan163] suggests that we not give up on assign unless > there a change -- an SCP triggers failure -- or at the extreme, an operator > intervenes. This jibes w/ how we're thinking about assign (or to put it > another way, we have no handling for the case where we exhaust retries). -- This message was sent by Atlassian JIRA (v7.6.3#76005)