[
https://issues.apache.org/jira/browse/HBASE-27793?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrew Kyle Purtell reopened HBASE-27793:
-----------------------------------------
This seems poorly executed for several reasons. HBCK1 is effectively read only,
this should have only been a change to HBCK2. Unknown servers are NOT
inconsistencies. Or otherwise the word becomes meaningless. An _inconsistency_
in a database system is a _data_ inconsistency. The fix versions were not set
correctly. The commit was committed even though it caused a spotless error, so
the precommit results were ignored.
Appreciate your help [~zhangduo] in bringing it up to our current practice, but
I'd rather revert this change given my reasoning above. Does that seem
reasonable?
My -1 can be cured by
1. Making this change in HBCK2.
2. Reporting unknown servers with informational messages only, not as an
"inconsistency"
3. Following our current committer guidelines, including paying attention to
precommit results, including use of spotless.
4. Proper attention to fix versions.
> Make HBCK be able to report unknown servers
> -------------------------------------------
>
> Key: HBASE-27793
> URL: https://issues.apache.org/jira/browse/HBASE-27793
> Project: HBase
> Issue Type: Bug
> Components: hbck
> Reporter: Rajeshbabu Chintaguntla
> Assignee: Rajeshbabu Chintaguntla
> Priority: Major
> Fix For: 2.6.0, 3.0.0-alpha-4, 2.5.5
>
>
> Currently hbck is not reporting unknown servers it would be helpful to report
> those as inconsistencies so that directly hbck2 schedulerecoveries option
> can be used to recover on unknown servers otherwise the taking the action for
> inconsistencies reported due to unknown servers may corrupt if not done
> properly.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
