Archiving HTTP log files: bzip2 versus gzip
The bzip2 compressed files are half the size of the gzip compressed files
The servers I work with generate fairly large http log files which we archive away for later analysis. Each month a typical log file (tens of millions of 'hits') gets compressed. I used to do this with gzip, but recently have moved over to bzip2 because I had heard it was more efficient. But beforehand, in the spirit of "measure before you optimize", I compressed five files from last year before 'trans-compressing' them all to bz2 files as a test. The results:
|
While the time to compress files with bzip2 is much, much longer than gzip, the files are half the size of the gzip files, and are a stunning 3% the size of the original files. That means I can file away twice as many of these logs on a single CD, which is very handy! (Yes, some of us geeks really do spend time and effort thinking about such weighty topics.) |
||||||||||||||||||||||||||||||||||||||||||
