HI,
Some mr tasks upon small files had run on our hadoop cluster these
days, with not such high
work load. While when i check it tonight, the cluster refused to
response. So i restart the hdfs
mapred.
BUT unexcepted exceptions were thrown when starting~ and even the
hadoop fs -ls command
could
I found it similiar to
HADDOP-3027(https://issues.apache.org/jira/browse/HADOOP-3027),
but its error msg is:
org.apache.hadoop.mapred.JobTracker: problem cleaning system
directory: /tmp/hadoop/mapred/system
while mine is:
org.apache.hadoop.mapred.JobTracker: problem cleaning system directory: