Best splittable compression for Hadoop input = bz2?
We've realized a bit too late that archiving our files in GZip format for Hadoop processing isn't such a great idea. GZip isn't splittable, and for reference, here are the problems which I won't repeat:
- Very basic question about Hadoop and compressed input files
- Hadoop gzip compressed files
- Hadoop gzip input file using only one mapper
- Why can't hadoop split up a large text file and then compress the splits using gzip?
My question is: is BZip2 the best archival compression that will allow a single archive file to be processed in parallel by Hadoop? Gzip is definitely not, and from my reading LZO has some problems.
Answer
BZIP2 is splittable in hadoop - it provides very good compression ratio but from CPU time and performances is not providing optimal results, as compression is very CPU consuming.
LZO is splittable in hadoop - leveraging hadoop-lzo you have splittable compressed LZO files. You need to have external .lzo.index files to be able to process in parallel. The library provides all means of generating these indexes in local or distributed manner.
LZ4 is splittable in hadoop - leveraging hadoop-4mc you have splittable compressed 4mc files. You don't need any external indexing, and you can generate archives with provided command line tool or by Java/C code, inside/outside hadoop. 4mc makes available on hadoop LZ4 at any level of speed/compression-ratio: from fast mode reaching 500 MB/s compression speed up to high/ultra modes providing increased compression ratio, almost comparable with GZIP one.