Re: BBU Cache vs. spindles

Kevin Grittner wrote:
> Greg Smith <greg(at)2ndquadrant(dot)com> wrote:
> > Kevin Grittner wrote:
>
> >> So you're confident that an 8kB write to the controller will not
> >> be done as a series of smaller atomic writes by the OS file
> >> system?
> >
> > Sure, that happens. But if the BBU has gotten an fsync call after
> > the 8K write, it shouldn't return success until after all 8K are
> > in its cache.
>
> I'm not concerned about an fsync after the controller has it; I'm
> concerned about a system crash in the middle of writing an 8K page
> to the controller. Other than the expected *size* of the window of
> time during which you're vulnerable, what does a BBU caching
> controller buy you in this regard? Can't the OS rearrange the
> writes of disk sectors after the 8K page is written to the OS cache
> so that the window might occasionally be rather large?

If the write fails to the controller, the page is not flushed and PG
does not continue. If the write fails, the fsync never happens, and
hence PG stops.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ It's impossible for everything to be true. +