hi Peter,
There is no any operation log? Please using the state change log with
debug level.
I checked the code for creating 'symlink'. It looks like regular
operation with other interfaces at NameNode. GC, directory lock, shell
script may be doubtful point.
-Regards
Denny Ye
2012/2/24 Ti
Hi, Denny,
Thanks for the info, you are right about his, after some digging, I found
there were no data loss, the root cause of getting empty file is that
during the file rotation, the old file was closed, the new empty new file
was created/opened, then hdfs tried to create the small file(acting as
Hi,
My use case scenario involves data coming in periodically over https (post)
from thousands of clients. Apache is the web server.
About 50MB of data arrives every 15 minutes from each of these clients.
I need to move this data into HDFS for further Mapreduce processing.
Would appreciate if the
Thanks very much for the clarification.
So, we'd i guess ideally set the block size equal to the transfer rate for
optimum results.
If seek time has to be 0.5% of transfer time would i set my block size at
200MB (higher than transfer rate)?
Conversely if seek time has to be 2% of transfer time wo