Hi all:
    I use flume to collect logs with taildir source and HDFS sink .  I 
configured  the hdfs sink to scroll over time . But when I compare the hdfs 
filet to orginal log file , I always find  the former's  count is less than  
the latter's . So ,I checked the flume agent log printed in nohup.log , I found 
that no exception were thrown.  Any one can help me ?  


[email protected]

Reply via email to