> > The way I really hoped would be implemented was to have the index > > files exist within the file subdirectory, i.e. all files with indexes > > would be similar in structure to Type30, but have a .index > > subdirectory within them. So, from then on, the indexes would be > > "tied" to the main data file. > > But then you wouldn't have the opportunity to put your data files and > their indexes on different spindles / channels / file systems to > maximise throughput by spreading I/O.
In unix, use "soft links". See man pages on "ln -s". It lets you make an entry in the appropriate unix dir that UV will look for, but the file actually lives somewhere else. Kinda like a Pick Q-pointer, or an F pointer with <2> a fully qualified pathname to another directory. You can do that today to spread out INDEX.000, INDEX.001 INDEX.002, etc. across many spindles. Likewise, with Unidata, I'm sure you can use ln -s to relocate the X_... files elsewhere for better throughput. cds ------- u2-users mailing list [EMAIL PROTECTED] To unsubscribe please visit http://listserver.u2ug.org/
