hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jing Zhao (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8127) NameNode Failover during HA upgrade can cause DataNode to finalize upgrade
Date Fri, 10 Apr 2015 17:39:13 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8127?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14489978#comment-14489978
] 

Jing Zhao commented on HDFS-8127:
---------------------------------

One possible fix is to extend the current {{-bootstrapStandby}} command:
# Admin can use a new {{-upgrade}} option for bootstrapStandby to tell SBN to create previous
directory (which may be empty if SBN is initially a brand new node).
# The SBN adds a new parameter in the http request while downloading the fsimage to indicate
that it understands the upgrade state
# While serving the downloading image request, the ANN makes sure the SBN knows the cluster
is in upgrade state.

> NameNode Failover during HA upgrade can cause DataNode to finalize upgrade
> --------------------------------------------------------------------------
>
>                 Key: HDFS-8127
>                 URL: https://issues.apache.org/jira/browse/HDFS-8127
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: ha
>    Affects Versions: 2.4.0
>            Reporter: Jing Zhao
>            Assignee: Jing Zhao
>            Priority: Blocker
>
> Currently for HA upgrade (enabled by HDFS-5138), we use {{-bootstrapStandby}} to initialize
the standby NameNode. The standby NameNode does not have the {{previous}} directory thus it
does not know that the cluster is in the upgrade state. If NN failover happens, as response
of block reports, the new ANN will tell DNs to finalize the upgrade thus make it impossible
to rollback again.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message