[
https://issues.apache.org/jira/browse/HDFS-11121?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15743244#comment-15743244
]
Wei-Chiu Chuang commented on HDFS-11121:
----------------------------------------
[~tasanuma0829] thanks for the patch. I reviewed the patch and I think it's
mostly Okay barring a few questions.
# I wonder if we can add a similar isStripedBlockId assertion in
BlockInfoStriped constructor.
# This is unrelated to your patch, but do you know if it is true that a
stripped block _always_ has a negative block id? Looking up
{{BlockIdManager#isStripedBlockID}} it seems true. I wasn't involved in the
design stage of EC, but this assumption is risky -- i have seen hadoop-2
clusters (with no stripped files of course) showing negative block ids.
> Add assertions to BlockInfo#addStorage to protect from breaking
> reportedBlock-blockGroup mapping
> ------------------------------------------------------------------------------------------------
>
> Key: HDFS-11121
> URL: https://issues.apache.org/jira/browse/HDFS-11121
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: erasure-coding
> Affects Versions: 3.0.0-alpha1
> Reporter: Takanobu Asanuma
> Assignee: Takanobu Asanuma
> Priority: Critical
> Fix For: 3.0.0-alpha2
>
> Attachments: HDFS-11121.1.patch
>
>
> There are not any assertions in {{BlockInfo.addStorage}}. This may cause that
> {{BlockInfo}} instances accept strange block reports and result in serious
> bugs, like HDFS-10858.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
