[
https://issues.apache.org/jira/browse/YARN-7351?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16212937#comment-16212937
]
Eric Yang edited comment on YARN-7351 at 10/20/17 7:55 PM:
-----------------------------------------------------------
-1 after applying patch 003, query started failing when it is used in
combination with patch for YARN-7326.
{code}
[yarn@eyang-1 hadoop-3.1.0-SNAPSHOT]$ dig @localhost -p 5353 .
;; Warning: query response not set
; <<>> DiG 9.9.4-RedHat-9.9.4-51.el7 <<>> @localhost -p 5353 .
; (2 servers found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOTAUTH, id: 48353
;; flags: rd ad; QUERY: 0, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; WARNING: recursion requested but not available
;; Query time: 9 msec
;; SERVER: 127.0.0.1#5353(127.0.0.1)
;; WHEN: Fri Oct 20 19:49:49 UTC 2017
;; MSG SIZE rcvd: 12
{code}
This is because the response payload is bigger than UDP datagram. TCP channel
for response is working using the initialized NIOTCPChannel.
was (Author: eyang):
+1 for disabling TCP channel for now.
> High CPU usage issue in RegistryDNS
> -----------------------------------
>
> Key: YARN-7351
> URL: https://issues.apache.org/jira/browse/YARN-7351
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Jian He
> Assignee: Jian He
> Attachments: YARN-7351.yarn-native-services.01.patch,
> YARN-7351.yarn-native-services.02.patch,
> YARN-7351.yarn-native-services.03.patch,
> YARN-7351.yarn-native-services.03.patch
>
>
> Thanks [~aw] for finding this issue.
> The current RegistryDNS implementation is always running on high CPU and
> pretty much eats one core.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
