Unhealthy NodeManager

[#1]

Getting this alert. What could be wrong ? How to fix it

1/1 local-dirs are bad: /hadoop/yarn/local; 1/1 log-dirs are bad: /hadoop.

Thanks

[#2]

Check health of your storage. Did permissions change?

[#3]

@Artem Ervits

Storage doesnt seem to have any issue. It has plenty of left storage.

Thanks

[#4]

Do you have these directories created? If these directories do not exist, application will write to filesystem and fill up your OS space.

[#5]

@Pankaj - Those are your local yarn directories where the containers will run. What is your disk size for these folders.

You probably need to increase the disk space.

$df -hk

on the disks and see how much space you have got on these mount points.

It is possible that one of your applications is erroring and filling up the log space.

Check the long running jobs.