Hi Roshan, The one way to fix this is to restart Flume, so the HDFS client does not renew the lease and the NN would close the file. If the format used is one that does not require any trailers, the file should still be recoverable. Flume can retry close() a limited number of times, but not indefinitely.
On Fri, Jan 31, 2014 at 1:20 PM, Roshan Naik <[email protected]> wrote: > What is the recommendation for this problem... > > When HDFS sink has trouble closing the tmp file on HDFS, the tmp file > lingers around without being closed or renamed. if HDFS sink is configured > to use compression, the data in this tmp file is not going to be > recoverable. > > Does it make sense for Flume to actually keep retrying the close() > operation itself in case of failure ? > > -Roshan > > -- > CONFIDENTIALITY NOTICE > NOTICE: This message is intended for the use of the individual or entity to > which it is addressed and may contain information that is confidential, > privileged and exempt from disclosure under applicable law. If the reader > of this message is not the intended recipient, you are hereby notified that > any printing, copying, dissemination, distribution, disclosure or > forwarding of this communication is strictly prohibited. If you have > received this communication in error, please contact the sender immediately > and delete it from your system. Thank You. >
