On Linux, ext2 is the fastest filesystem available. You overcome the long fsck problem by limiting the number of inodes on the filesystem. With very large filesystems this can number into the tens of millions. If you use the -N option when you create the filesystem (using mke2fs) and only use the number of inodes you will need, the fsck runs amazingly fast. I've created some 800GB filesystems with under 1000 inodes and the fsck runs in seconds (assuming you're using this for large TSM data files). A journaling filesystem would not be a good choice anyway for large TSM data that's getting a huge number of writes to it.
Mitch West Suhanic wrote: > Hello: > > > What type of OS on Intel? > > > > If it's Linux then the problem is easy to overcome (and I don't mean by > using a > > journaling filesystem, either). > > > > Mitch > > > > If a journalling file system is not the solution. What is? Could you please > state the > solution you have in mind? > > thank you, > > west suhanic
