[
https://issues.apache.org/jira/browse/HDFS-6005?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Suresh Srinivas updated HDFS-6005:
----------------------------------
Attachment: HDFS-6005.patch
Changes:
# As proposed in the description, datanode no longer requires startup option
-rollingUpgrade -rollback and corresponding code is removed.
# When upgrade status is reported by namenode, trash is enabled in the
datanode, as before. During datanode restart (either during rollback or
downgrade), during startup, datanode restores the block files from the trash.
During rollback, the restored blocks are retained. During downgrade the
restored blocks are deleted. When namenode is finalized, the trash is disabled
and the block files in trash are removed.
> Simplify Datanode rollback and downgrade
> ----------------------------------------
>
> Key: HDFS-6005
> URL: https://issues.apache.org/jira/browse/HDFS-6005
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode, namenode
> Reporter: Suresh Srinivas
> Assignee: Suresh Srinivas
> Attachments: HDFS-6005.patch
>
>
> Problem:
> When rolling upgrade fails, the cluster can either be downgraded or rolled
> back. With the current functionality in this feature branch, it is possible
> to downgrade namenode, while datanode is incorrectly rolled back. This does
> not affect the cluster state. The old blocks that appear back on the datanode
> due to rollback will be deleted. Similarly it is also possible to rollback
> namenode, while datanode is not rolled back. This can cause problem where old
> blocks do not appear back on the datanode and can result in missing blocks.
> Solution:
> I propose making the following changes:
> During rollback or downgrade, the entire cluster must be restarted. The
> datanodes always restore the deleted blocks on restart and go back to trash
> disabled mode. There is no need for datanodes to be started up
> -rollingUpgrade -rollback, anymore.
> # On namenode downgrade, the restored blocks are deleted.
> # On namenode rollback, the restored blocks will be retained and any newly
> created blocks (since the start of rolling upgrade) are deleted.
> This is much simpler operationally and solves the problem described above.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
