Hi,
Not just the file size, the space consumed also depend upon the replication factor and some other factors for example- if you copy or move file from one location to another. The system requires 2 copy of data, one is original and one is copy of that data. It takes twice the data size. Some space are reserved by system for internal process.
Either decrease the replication factor and give the cluster sometime to resync all the data.
Then try enter the file in hdfs or expand the partition size of datanode & namenode.

Thanks, hope this helps.
I