hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xiao Chen (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HDFS-12518) Re-encryption should handle task cancellation and progress better
Date Thu, 21 Sep 2017 03:12:00 GMT

     [ https://issues.apache.org/jira/browse/HDFS-12518?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Xiao Chen updated HDFS-12518:
-----------------------------
    Attachment: HDFS-12518.01.patch

Patch 1:
- Added the fix in {{ReencryptionUpdater}} to handle canceled future better
- Modified the way {{submissions}} in {{ReencryptionHandler}} is managed: on -cancel and valid
new -start, the zone submission tracker is reset. This means even if somehow an re-encryption
is ended up in an erroneous state, it can still be reset from crypto CLI.
- Fixed / improved a few related places: protect {{submissions}} with the handler's object
lock, logging fix, missed assert.
- Enhanced {{TestReencryption}} to ensure fault injector behave correctly when multi-threaded
(had a 1% failure rate on dist-test

> Re-encryption should handle task cancellation and progress better
> -----------------------------------------------------------------
>
>                 Key: HDFS-12518
>                 URL: https://issues.apache.org/jira/browse/HDFS-12518
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: encryption
>    Affects Versions: 3.0.0-beta1
>            Reporter: Xiao Chen
>            Assignee: Xiao Chen
>         Attachments: HDFS-12518.01.patch
>
>
> Re-encryption should handle task cancellation and progress tracking better in general.
> In a recent internal report, a canceled re-encryption could lead to the progress of the
zone being 'Processing' forever. Sending a new cancel command would make it complete, but
new re-encryptions for the same zone wouldn't work because the canceled future is not removed.
> This jira proposes to fix that, and enhance the currently handling so new command would
start from a clean state.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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


Mime
View raw message