Created on 08-05-2016 04:10 AM - edited 09-16-2022 03:33 AM
Using Cloudera Manager for CDH 5.8.0, deploying on AWS machines with RH 7.2.
When installing YARN, the NodeManagers fail to start. Looking at the logs, there is an exception as follows:
Service org.apache.hadoop.yarn.server.nodemanager.recovery.NMLeveldbStateStoreService failed in state INITED; cause: org.fusesource.leveldbjni.internal.NativeDB$DBException: IO error: /var/lib/hadoop-yarn/yarn-nm-recovery/yarn-nm-state/LOCK: Permission denied org.fusesource.leveldbjni.internal.NativeDB$DBException: IO error: /var/lib/hadoop-yarn/yarn-nm-recovery/yarn-nm-state/LOCK: Permission denied at org.fusesource.leveldbjni.internal.NativeDB.checkStatus(NativeDB.java:200) at org.fusesource.leveldbjni.internal.NativeDB.open(NativeDB.java:218) at org.fusesource.leveldbjni.JniDBFactory.open(JniDBFactory.java:168) at ....
This error is actually identical to the one reported in this post
Although the version/diagnostics are not the same.
Anyway, I've connected to one of NodeManager machines to take a look, and eventually I've seen that the home directory of yarn belonged to root and did not have any permissions:
d---------. 2 root root 6 Aug 5 05:53 hadoop-yarn
As soon as I've given permissions to it, the NodeManager has started properly.
chown -R yarn:yarn /var/lib/hadoop-yarn chmod 770 /var/lib/hadoop-yarn drwxrwx---. 3 yarn yarn 49 Aug 5 06:37 hadoop-yarn
However, I don't want to do this for all the machines (and potentially for many other users that I see are wrong in the /var/lib directory), and I think it must be my mistake when installing...
I've tried removing the YARN service and installing it again, but I get the same result.
Thanks for your help.
Created 08-08-2016 07:37 AM
The problem got solved.
The solution of the problem was as "simple" as re-running everything using CentOS 6.4. (Red Hat 7.2 is not really in the explicit list of supported versions).
Although at the same time I could connect to the AWS instances with the user "root" (in RH 7.2 AWS I had to do it with "ec2-user"). Not sure about the difference between the 2 users in AWS, but it may or may not be the reason why the issue was solved.
Thanks everybody.
Created 08-08-2016 07:37 AM
The problem got solved.
The solution of the problem was as "simple" as re-running everything using CentOS 6.4. (Red Hat 7.2 is not really in the explicit list of supported versions).
Although at the same time I could connect to the AWS instances with the user "root" (in RH 7.2 AWS I had to do it with "ec2-user"). Not sure about the difference between the 2 users in AWS, but it may or may not be the reason why the issue was solved.
Thanks everybody.
Created 09-01-2016 10:06 AM
having the same issue but not sure about the solution here , how is Red Hat 7.2 not in the explicit list of supported versions ?
Created 09-01-2016 12:39 PM
Fair enough. I've tried to remember and looked again for the info and found this,
https://www.cloudera.com/documentation/enterprise/release-notes/topics/rn_consolidated_pcm.html#os58
which says that 5.8 supports RH7.2
Maybe the problem is that at that time I read this page
https://www.cloudera.com/documentation/enterprise/5-6-x/topics/cdh_ig_req_supported_versions.html
in which support for CDH 5 in general goes only until RH7.1
To be honest, at that moment I was just trying to train myself and I did not need to stick to RH7.2, but of course now I would not dare to say which one is supported and which one isn't.
Created 09-01-2016 02:05 PM