[
https://issues.apache.org/jira/browse/HDFS-5037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13811253#comment-13811253
]
Aaron T. Myers commented on HDFS-5037:
--------------------------------------
I'm pretty confident that the eclipse failure is unrelated:
{noformat}
[ERROR] Failed to execute goal
org.codehaus.mojo:build-helper-maven-plugin:1.5:add-source
(add-jsp-generated-sources-directory) on project hadoop-hdfs: Execution
add-jsp-generated-sources-directory of goal
org.codehaus.mojo:build-helper-maven-plugin:1.5:add-source failed: Unable to
load the mojo 'add-source' in the plugin
'org.codehaus.mojo:build-helper-maven-plugin:1.5'. A required class is missing:
org.codehaus.mojo.buildhelper.AddSourceMojo
{noformat}
A few small comments:
# One too many 'a's: "+ // # edit autoroll threshold is a a multiple of
the checkpoint threshold "
# Is there some reason that the variables in the TestEditLogAutoroll class need
to be static? Seems like they should be instance variables to me.
# Doesn't look like you ever use the nn1 variable in the test.
# Similarly you only ever use the namesystem variable once, just to get a
reference to the edit log. Might want to just chain those calls.
# Are we positive that the test would fail without the auto roller in place?
i.e. is it not possible that the standby NN in the minicluster might trigger an
edit log roll and cause the test to pass without the auto roller having done
anything?
+1 once these are addressed.
> Active NN should trigger its own edit log rolls
> -----------------------------------------------
>
> Key: HDFS-5037
> URL: https://issues.apache.org/jira/browse/HDFS-5037
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: ha, namenode
> Affects Versions: 3.0.0, 2.1.0-beta
> Reporter: Todd Lipcon
> Assignee: Andrew Wang
> Priority: Critical
> Attachments: hdfs-5037-1.patch, hdfs-5037-2.patch, hdfs-5037-3.patch,
> hdfs-5037-4.patch
>
>
> We've seen cases where the SBN/2NN went down, and then users accumulated very
> very large edit log segments. This causes a slow startup time because the
> last edit log segment must be read fully to recover it before the NN can
> start up again. Additionally, in the case of QJM, it can trigger timeouts on
> recovery or edit log syncing because the very-large segment has to get
> processed within a certain time bound.
> We could easily improve this by having the NN trigger its own edit log rolls
> on a configurable size (eg every 256MB)
--
This message was sent by Atlassian JIRA
(v6.1#6144)
