hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeffrey Zhong (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-10581) ACL znode are left without PBed during upgrading hbase0.94* to hbase0.96+
Date Fri, 21 Feb 2014 19:15:23 GMT

    [ https://issues.apache.org/jira/browse/HBASE-10581?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13908695#comment-13908695

Jeffrey Zhong commented on HBASE-10581:

Can we modify this patch to do Himanshu's suggestion? add a deleteRecursive call at the end
of all these conditional checks
I'm worried that if a customer has their own custom znodes.  Cleaning the data we don't know
may cause data loss for them if they're not aware of this side effect. Even sideline those
znodes, it will also cause pain for them to move them back.  So far all znodes we know are
covered after the fix. Therefore, unknown znodes if there is any mostly likely are users data.

IMHO, it's better to be conservative to work on data that we know.

> ACL znode are left without PBed during upgrading hbase0.94* to hbase0.96+
> -------------------------------------------------------------------------
>                 Key: HBASE-10581
>                 URL: https://issues.apache.org/jira/browse/HBASE-10581
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 0.98.0, 0.96.0, 0.96.1
>            Reporter: Jeffrey Zhong
>            Assignee: Jeffrey Zhong
>            Priority: Critical
>         Attachments: hbase-10581.patch
> ACL znodes are left in the upgrade process when upgrading 0.94 to 0.96+
> Those 0.94 znodes will choke HMaster because their data aren't PBed.  

This message was sent by Atlassian JIRA

View raw message