[ https://issues.apache.org/jira/browse/HBASE-25007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17203613#comment-17203613 ]
Michael Stack commented on HBASE-25007: --------------------------------------- {quote}And the problem is that, we will not schedule SCP for them if something goes wrong. {quote} Correct. Maybe we could if an 'unknown server' has been in hbase:meta for more than two or ten heartbeats... The catalog janitor could schedule an SCP for any 'unknown server' found. {quote}So I think a normal SCP is enough? {quote} In most cases, yes. {quote}Or at least, we do not need to scan meta to find out the regions on a 'unknown server'? {quote} HBCKSCP only does this if the Operator runs an HBCKSCP and the SCP super-call returns that there are no matching crashed Servers. In this case, the Operator is insisting that SCP has 'missed' some references to the named server; in this latter case, HBCKSCP goes the extra running a full scan looking for any references to the passed server – even looking for references from Region Replicas. Fold this latter checking bit into SCP and then remove HBCKSCP? I've not run into case where meta was on the 'unknown server'. Meta has to be up to make any progress. On restarting Master to get the latest, in my 'cluster fixing experience', the Master has usually been started recently. > Make HBCK2 work for 'root table' > -------------------------------- > > Key: HBASE-25007 > URL: https://issues.apache.org/jira/browse/HBASE-25007 > Project: HBase > Issue Type: Sub-task > Components: hbck2 > Reporter: Duo Zhang > Assignee: Duo Zhang > Priority: Major > > We will also scan catalog table and fix them in HBCK2, we should add support > for root too. -- This message was sent by Atlassian Jira (v8.3.4#803005)