Re: Experimental patch for inter-page delay in VACUUM

Tom Lane wrote:
> "Zeugswetter Andreas SB SD" <ZeugswetterA(at)spardat(dot)at> writes:
> > One problem with O_SYNC would be, that the OS does not group writes any
> > more. So the code would need to eighter do it's own sorting and grouping
> > (256k) or use aio, or you won't be able to get the maximum out of the disks.
>
> Or just run multiple writer processes, which I believe is Oracle's
> solution.

Yes, that might need to be the final solution because the O_SYNC will be
slow. However, that is a lot of "big wrench" solution to removing
sync() --- it would be nice if we could find a more eligant way.

In fact, one goffy idea would be if the OS does sync every 30 seconds to
just write() the buffers and wait 30 seconds for the OS to issue the
sync, then recycle the WAL buffers --- again, just a crazy thought.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073