[
https://issues.apache.org/jira/browse/HDFS-5328?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13790162#comment-13790162
]
Steve Loughran commented on HDFS-5328:
--------------------------------------
not seen this -is the jenkins box under load/memory swap?
regardless, hard coded delays are always brittle, and could be made
configurable with a longer timeout. After all, if the minicluster comes up
normally the test is no slower, and if it takes >10s to start, the test fails
which is far more expensive in terms of human intervention.
Maybe extend the default spinup time to 30s?
> MiniDFSCluster times out at startup
> -----------------------------------
>
> Key: HDFS-5328
> URL: https://issues.apache.org/jira/browse/HDFS-5328
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: test
> Affects Versions: 2.1.1-beta
> Reporter: Alejandro Abdelnur
>
> It seems MiniDFSCluster startup got slower lately and we are observing the
> following intermittent failure on startup in our jenkins boxes:
> {code}
> java.io.IOException: Timed out waiting for Mini HDFS Cluster to start
> at
> org.apache.hadoop.hdfs.MiniDFSCluster.waitClusterUp(MiniDFSCluster.java:972)
> at
> org.apache.hadoop.hdfs.MiniDFSCluster.initMiniDFSCluster(MiniDFSCluster.java:665)
> at org.apache.hadoop.hdfs.MiniDFSCluster.<init>(MiniDFSCluster.java:585)
> at org.apache.hadoop.hdfs.MiniDFSCluster.<init>(MiniDFSCluster.java:511)
> {code}
> By looking at the code the {{waitClusterUp()}} has a hardcoded 10secs timeout
> (10 1 sec waits).
> There are two things we should look at it, make the timeout configurable, and
> more important, see what has changed to make the startup slower.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
