proc hadoop did not copy entirely a csv file

[blast]

Hi everyone ! 

i've tried many times to copy a csv file to a hadoop cluster.

The datafile is 4.6 Go.

Randomly, the file is copied entirely or stops at 2.1 Go.. 1.5Go..1.6Go sometimes 

i dont know how to make sure that the file is correcly copied, there is no error on the log. 

Did you ever had that kind of problem ? 

proc hadoop username="&user." password="&mdp." verbose;
   hdfs delete="&chemin_hdfs./&fichierIn." nowarn;
run;

proc hadoop username="&user." password="&mdp." verbose;
   hdfs copyfromlocal="&chemin_sorties.\&fichierIn." out="&chemin_hdfs."   ;
run;