Orlando Andico wrote:
> On Fri, 7 May 1999 unknown@stripped wrote:
> > of the filesystem, which I believe is UFS. If the implementation is the
> > same as FreeBSD's UFS subsystem, it always makes sure that the metainfo
> > for the filesystem is updated first, making it easier to recover from
> > unexpected power outages, ec. It's more robust compared to ext2 though.
> It is. To quote, "metadata is synchronously updated on UFS." Strange, but
> the last time I tried FreeBSD it didn't seem that slow (but then, that was
> a 486 and EVERYTHING was slow back then..)
> > If you'd rather use Linux, there are big file patches available for the
> > 2.0.x series, and there are also (preferable?) patches available for the
> > 2.2.x series. They should let you get up to 4 gig/file.
> I am not certain if these patches are supported in the libc. I think
> libc6.1 has Large File Summit support, although I am not sure if it is
> enabled in Red Hat 6.0.
Or could could just keep your tables small. Having to
worry about the size of the tables will make you design
them better and your queries will be faster :)