[
https://issues.apache.org/jira/browse/HADOOP-702?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12468818
]
Konstantin Shvachko commented on HADOOP-702:
--------------------------------------------
> Data-nodes automatically "catching up" on a missed previous upgrade/discard
> seems like a good thing, isn't it?
Yes, but!
We are trying to protect the system from human mistakes. Now suppose that
adminK started the overnight upgrade
of the system before going home, he plans to come back in the morning and check
whether the upgrade was successful
or not. But another adminY comes to work earlier and not knowing about adminK
actions last night starts the upgrade again.
The data-nodes will automatically discard "previous" fs state before upgrading
because they can store only one backup per node.
So it can automatically discard the last working state of the file system if
the upgraded software had bugs affecting the namespace.
I see it as the main problem with our new approach.
> DFS Upgrade Proposal
> --------------------
>
> Key: HADOOP-702
> URL: https://issues.apache.org/jira/browse/HADOOP-702
> Project: Hadoop
> Issue Type: New Feature
> Components: dfs
> Reporter: Konstantin Shvachko
> Assigned To: Konstantin Shvachko
> Attachments: DFSUpgradeProposal.html, DFSUpgradeProposal2.html,
> DFSUpgradeProposal3.html, FSStateTransition.html, TestPlan-HdfsUpgrade.html,
> TestPlan-HdfsUpgrade.html
>
>
> Currently the DFS cluster upgrade procedure is manual.
> http://wiki.apache.org/lucene-hadoop/Hadoop_Upgrade
> It is rather complicated and does not guarantee data recoverability in case
> of software errors or administrator mistakes.
> This is a description of utilities that make the upgrade process almost
> automatic and minimize chance of loosing or corrupting data.
> Please see the attached html file for details.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
