[
https://issues.apache.org/jira/browse/HBASE-25007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17194021#comment-17194021
]
Michael Stack commented on HBASE-25007:
---------------------------------------
{quote}but the same trick can not be used to fix meta
{quote}
Perhaps have it scan ROOT or the meta location in the Registry too?
{quote}What is the value of HBCKSCP?
{quote}
It is for fixing a corrupted/broken cluster. All the SCPs that were scheduled
'naturally' are long gone or bypassed or dropped but meta still has references
to 'unknown servers'.
HBCPSCP is run by an operator via hbck2. Usually when it runs, it just does SCP
stuff and will cleanup 'unknown server' references and do reassigns but if an
operator runs it again and the SCP run can't find any references to the
deadserver in hbase:meta or elsewhere, then it just does the little extra trick
of scanning full hbase:meta because the operator insists there are 'unknown
servers' that the normal SCP is not finding.
"{color:#808080}It should never happen in normal running
{color}{color:#808080}* cluster but if we do drop accounting of servers, we
need a means of fix-up.
{color}{color:#808080}* Eventually, as part of normal CatalogJanitor task,
rather than just identify
{color}{color:#808080}* these 'Unknown Servers', it would make repair, queuing
something like this
{color}{color:#808080}* HBCKSCP to do cleanup, reassigning them so Master and
hbase:meta are aligned again.{color}"
It has come in handy over these last few months effecting repairs when running
upgrades on old clusters in varying states of consistency especially when they
were larger and the number of problems overwhelming.
You thinking fix on restart of Master? Looking for bad references? Sure. Could
CatalogJanitor look for them too?
> 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)
