Re: Multicolun index creation never completes on 9.0.1/solaris

Josh Berkus <josh(at)agliodbs(dot)com> writes:
> OK, either tape sort is broken or trace_sort is. The larger my
> maint_work_mem, the MORE tapes it wants to use:

Yeah, that's right. More tapes = better, or at least that was the
conclusion we came to some years ago. It doesn't necessarily use all
those tapes --- it's hoping it will find that the data sorts into fewer
runs than that, and it can go directly to the final merge without any
intermediate merge passes.

It's conceivable that there's an upper bound to the number of tapes we
should try to use, but you really haven't provided any data that would
either confirm or refute that idea.

regards, tom lane