hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rohith Sharma K S <rohithsharm...@huawei.com>
Subject RE: NodeHealthReport local-dirs turned bad
Date Wed, 19 Mar 2014 12:08:13 GMT
Hi

There is no relation to NameNode format!!!!

Does NodeManger is started with default configuration? If no , any NodeManger health script
is configured?

Suspect can be  
    1. /hadoop does not have permission or 
    2. disk is full!!!!

Thanks & Regards
Rohith Sharma K S


-----Original Message-----
From: Margusja [mailto:margus@roo.ee] 
Sent: 19 March 2014 17:04
To: user@hadoop.apache.org
Subject: NodeHealthReport local-dirs turned bad

Hi

I have one node in unhealthy status:

	


Total Vmem allocated for Containers 	4.20 GB
Vmem enforcement enabled 	false
Total Pmem allocated for Container 	2 GB
Pmem enforcement enabled 	false
NodeHealthyStatus 	false
LastNodeHealthTime 	Wed Mar 19 13:31:24 EET 2014
NodeHealthReport 	1/1 local-dirs turned bad: /hadoop/yarn/local;1/1 
log-dirs turned bad: /hadoop/yarn/log
Node Manager Version: 	2.2.0.2.0.6.0-101 from 
b07b2906c36defd389c8b5bd22bebc1bead8115b by jenkins source checksum
82bd166aa0ada92b44f8a154836b92 on 2014-01-09T05:24Z
Hadoop Version: 	2.2.0.2.0.6.0-101 from 
b07b2906c36defd389c8b5bd22bebc1bead8115b by jenkins source checksum
704f1e463ebc4fb89353011407e965 on 2014-01-09T05:18Z



I tried:
Deleted /hadoop/* and did namenode -format again Restarted nodemanager but still in unhealthy
mode.

Is there any guideline what I should do?

--
Best regards, Margus (Margusja) Roo
+372 51 48 780
http://margus.roo.ee
http://ee.linkedin.com/in/margusroo
skype: margusja
ldapsearch -x -h ldap.sk.ee -b c=EE "(serialNumber=37303140314)"


Mime
View raw message