I'm testing the availible filesystems on Linux for a new storage array. The first test was simply copying a large directory containing many small files (svn tree from KDE) and measure the time for this. The copied data were ~4GB, tested on a 14GB partition. The test were done on a 2.6.18 vanilla kernel, patched with ext4 and reiser4. I measured the time needed for copying the directory, and after this the time for removing the new dir. For the different filesystems I got the following results:
ext2: cp 4.54user 32.91system 10:52.29elapsed 5%CPU rm 0.60user 4.75system 2:28.63elapsed 3%CPU ext3: cp 4.79user 41.20system 15:10.08elapsed 5%CPU rm 0.78user 8.39system 2:21.12elapsed 6%CPU ext4: cp 5.06user 40.53system 14:44.94elapsed 5%CPU rm 0.81user 8.81system 2:40.12elapsed 6%CPU jfs: cp 4.39user 37.65system 58:02.18elapsed 1%CPU rm 0.82user 19.25system 32:28.72elapsed 1%CPU reiser4: cp 5.76user 88.80system 6:08.21elapsed 25%CPU rm 0.92user 41.97system 1:39.27elapsed 43%CPU reiserfs: cp 5.22user 89.56system 11:44.85elapsed 13%CPU rm 0.82user 40.32system 1:57.11elapsed 35%CPU xfs: cp 6.73user 81.04system 45:21.30elapsed 3%CPU rm 1.30user 31.71system 21:01.96elapsed 2%CPU I'm really confused about these times. As I understand XFS and JFS are designed for high performance. These results show the absolute opposite. Did I missed something or is my understandig of (these) filesystems wrong? I'm asking this here because the JFS-results are the worst. Thanks for help, Johannes ------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 _______________________________________________ Jfs-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jfs-discussion
