[
https://issues.apache.org/jira/browse/YARN-4352?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15065685#comment-15065685
]
Sunil G commented on YARN-4352:
-------------------------------
Hi [~djp]/[~rohithsharma]
I analyzed this issue along with YARN-4306. {{MiniYARNCluster}} is used here
and YarnClient cannot connect to RM and throwing {{UnKnownHostException}}. This
causes timeout in all these test cases.
{{QualifiedHostResolver}} is used in SecurityUtils for this cases
({{useIpForTokenService}} is false). And I could see that {{/etc/hosts}} has 2
loop back entries and 2nd one is the hostname of the machine. Hence below code
will not return machine host name, instead it will return "localhost". {{
InetAddress.getByName(null)}}
{noformat}
// it's a simple host with no dots, ex. "host"
// try the search list, then fallback to exact host
InetAddress loopback = InetAddress.getByName(null);
if (host.equalsIgnoreCase(loopback.getHostName())) {
addr = InetAddress.getByAddress(host, loopback.getAddress());
} else {
addr = getByNameWithSearch(host);
if (addr == null) {
addr = getByExactName(host);
}
}
{noformat}
I have provided more detailed comment in
[YARN-4306-Analysis|https://issues.apache.org/jira/browse/YARN-4306?focusedCommentId=15030122&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15030122]
and possible solutions are
Solution:
1, loopback address configuration can be changed to use hostname first in the
list {{/etc/hosts}} in jenkins machine.
2, use DNS in jenkins machine
3, we can make changes in {{SecurityUtils#getByExactName}} where we check with
given hostname itself before doing with hostname + ".".
> Timeout for tests in TestYarnClient, TestAMRMClient and TestNMClient
> --------------------------------------------------------------------
>
> Key: YARN-4352
> URL: https://issues.apache.org/jira/browse/YARN-4352
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Junping Du
> Assignee: Sunil G
>
> From
> https://builds.apache.org/job/PreCommit-YARN-Build/9661/artifact/patchprocess/patch-unit-hadoop-yarn-project_hadoop-yarn_hadoop-yarn-client-jdk1.7.0_79.txt,
> we can see the tests in TestYarnClient, TestAMRMClient and TestNMClient get
> timeout which can be reproduced locally.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
