hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Íñigo Goiri (JIRA) <j...@apache.org>
Subject [jira] [Commented] (HDFS-12773) RBF: Improve State Store FS implementation
Date Tue, 13 Mar 2018 17:07:00 GMT

    [ https://issues.apache.org/jira/browse/HDFS-12773?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16397276#comment-16397276

Íñigo Goiri commented on HDFS-12773:

Thanks, [~linyiqun] for the comments. Uploaded [^HDFS-12773.007.patch]  with your comments.

A good way should update now() to monotonicNow() both in StateStoreFileBaseImpl#putAll and
StateStoreFileBaseImpl#isOldTempRecord, instead of reverting that one.
The problem is the synchronization across Routers.
Imagine Router 1, is writing the membership record and sets {{monotonicNow()}}, while this
is happening, Router 2, comes and checks that time, as {{monotonicNow()}} is not universal,
the values they will get are not comparable.

> RBF: Improve State Store FS implementation
> ------------------------------------------
>                 Key: HDFS-12773
>                 URL: https://issues.apache.org/jira/browse/HDFS-12773
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Íñigo Goiri
>            Assignee: Íñigo Goiri
>            Priority: Major
>         Attachments: HDFS-12773.000.patch, HDFS-12773.001.patch, HDFS-12773.002.patch,
HDFS-12773.003.patch, HDFS-12773.004.patch, HDFS-12773.005.patch, HDFS-12773.006.patch
> HDFS-10630 introduced a filesystem implementation of the State Store for unit tests.
However, this implementation doesn't handle multiple writers concurrently.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org

View raw message