Mailing-List: contact hdfs-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Date: Mon, 13 Feb 2017 11:32:41 +0000 (UTC)
From: "Yiqun Lin (JIRA)" <jira@apache.org>
To: hdfs-issues@hadoop.apache.org
Message-ID: <JIRA.13033982.1484171354000.62139.1486985561836@Atlassian.JIRA>
In-Reply-To: <JIRA.13033982.1484171354000@Atlassian.JIRA>
References: <JIRA.13033982.1484171354000@Atlassian.JIRA> <JIRA.13033982.1484171354520@jira-lw-us.apache.org>
Subject: [jira] [Commented] (HDFS-11333) Namenode unable to start if plugins
 can not be found
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Mon, 13 Feb 2017 12:19:45 -0000


    [ https://issues.apache.org/jira/browse/HDFS-11333?page=3Dcom.atlassian=
.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D1586=
3516#comment-15863516 ]=20

Yiqun Lin commented on HDFS-11333:
----------------------------------

The suggestion from [~aw] also makes sense to me. Based on v01 patch, I thi=
nk  it's suffient to just rethrow the exception here.=20
{code}
+    try {
+      plugins =3D conf.getInstances(DFS_NAMENODE_PLUGINS_KEY,
+          ServicePlugin.class);
+    } catch (RuntimeException e) {
+      LOG.warn("Unable to load plugins", e);
       throw e;          <=3D=3D=3D
+    }
{code}
And this logic can be reused for datanode plugins loading.

> Namenode unable to start if plugins can not be found
> ----------------------------------------------------
>
>                 Key: HDFS-11333
>                 URL: https://issues.apache.org/jira/browse/HDFS-11333
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: namenode
>    Affects Versions: 0.21.0
>            Reporter: Wei-Chiu Chuang
>            Assignee: Wei-Chiu Chuang
>              Labels: supportability
>         Attachments: HDFS-11333.001.patch
>
>
> If NameNode is unable to find plugins (specified in dfs.namenode.plugins)=
, it terminates abruptly with the following stack trace:
> {quote}
> Failed to start namenode.
> java.lang.RuntimeException: java.lang.ClassNotFoundException: Class XXX n=
ot found
> =E2=80=82=E2=80=82at org.apache.hadoop.conf.Configuration.getClasses(Conf=
iguration.java:2178)
> =E2=80=82=E2=80=82at org.apache.hadoop.conf.Configuration.getInstances(Co=
nfiguration.java:2250)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.star=
tCommonServices(NameNode.java:713)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.init=
ialize(NameNode.java:691)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.<ini=
t>(NameNode.java:843)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.<ini=
t>(NameNode.java:822)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.crea=
teNameNode(NameNode.java:1543)
> =E2=80=82=E2=80=82at org.apache.hadoop.hdfs.server.namenode.NameNode.main=
(NameNode.java:1611)
> {quote}
> We should catch this exception, log a warning message and let it proceed,=
 as missing the third party library does not affect the functionality of Na=
meNode. We caught this bug during a CDH upgrade where a third party plugin =
was not in the lib directory of the newer version of CDH.


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

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