[
https://issues.apache.org/jira/browse/HDFS-898?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12803024#action_12803024
]
Konstantin Shvachko commented on HDFS-898:
------------------------------------------
You are absolutely right if everything goes well generation stamps help
recognizing prehistoric replicas of newly created blocks.
When I say the problem is relevant I mean to emphasize two cases:
# In case of corrupt blocks, when all generation stamps of its replicas are
less than the block's gen. stamp on the NN, we do not have means to recognize
whether the replica is just a stale replica of this particular block or some
other (prehistoric) block. This happens when you don't have up to date
replicas of the block and a data-node with a prehistoric replica of previously
removed block with the same id starts up.
# The really old blocks, created before introduction of generation stamps, can
still cause the problem.
I agree this is not exactly the problem as you defined it, but they are still
related problems.
> Sequential generation of block ids
> ----------------------------------
>
> Key: HDFS-898
> URL: https://issues.apache.org/jira/browse/HDFS-898
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: name-node
> Affects Versions: 0.20.1
> Reporter: Konstantin Shvachko
> Assignee: Konstantin Shvachko
> Fix For: 0.22.0
>
>
> This is a proposal to replace random generation of block ids with a
> sequential generator in order to avoid block id reuse in the future.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
