hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Yi Liu (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (HDFS-7740) Test truncate with DataNodes restarting
Date Fri, 20 Feb 2015 08:48:12 GMT

    [ https://issues.apache.org/jira/browse/HDFS-7740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14328679#comment-14328679
] 

Yi Liu edited comment on HDFS-7740 at 2/20/15 8:47 AM:
-------------------------------------------------------

Thanks [~shv] for your comments. I will address them later, currently I'am on holiday for
Chinese traditional new year, so there is some delay for the update.

{quote}
I don't think your new tests will work with the ones existing in TestFileTruncate, because
you are restarting the whole cluster
{quote}
Oh, not noticed these before :) I thought if I use a different Mini DFS Cluster with different
ports, then they will not affect each other.

{quote}
I would propose to put your new cases into a new file
It would be good to start the cluster once in @BeforeClass, and make sure all DNs are up after
each test case.
I think this is possible as you do not need to reformat the cluster after each test
{quote}
Good idea, I will move them to a new class. Yes, we don't need to reformat the cluster.

{quote}
With DNs restarting one way to accelerate test running time is to shorten heartbeats, as we
did in TestFileTruncate.
{quote}
We already use the short hearbeats for the new test.



was (Author: hitliuyi):
Thanks [~shv] for your comments. I will address them later, currently I'am on holiday for
Chinese traditional new year, so there is some delay for the update.

{quote}
I don't think your new tests will work with the ones existing in TestFileTruncate, because
you are restarting the whole cluster
{quote}
Oh, not noticed these before :) I thought if I use a different Mini DFS Cluster with different
ports, then they will not affect each other.

{quote}
I would propose to put your new cases into a new file
I think this is possible as you do not need to reformat the cluster after each test
{quote}
Good idea, I will move them to a new class. Yes, we don't need to reformat the cluster.

{quote}
It would be good to start the cluster once in @BeforeClass, and make sure all DNs are up after
each test case.
{quote}
To make sure all DNs are up after each test case, I think we are referring to use @Before
instead of @BeforeClass
{quote}
With DNs restarting one way to accelerate test running time is to shorten heartbeats, as we
did in TestFileTruncate.
{quote}
We already use the short hearbeats for the new test.


> Test truncate with DataNodes restarting
> ---------------------------------------
>
>                 Key: HDFS-7740
>                 URL: https://issues.apache.org/jira/browse/HDFS-7740
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: test
>    Affects Versions: 2.7.0
>            Reporter: Konstantin Shvachko
>            Assignee: Yi Liu
>             Fix For: 2.7.0
>
>         Attachments: HDFS-7740.001.patch
>
>
> Add a test case, which ensures replica consistency when DNs are failing and restarting.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message