Re: Checkpoint cost, looks like it is WAL/CRC

Simon Riggs <simon(at)2ndquadrant(dot)com> writes:
> Is there also a potential showstopper in the redo machinery? We work on
> the assumption that the post-checkpoint block is available in WAL as a
> before image. Redo for all actions merely replay the write action again
> onto the block. If we must reapply the write action onto the block, the
> redo machinery must check to see whether the write action has already
> been successfully applied before it decides to redo. I'm not sure that
> the current code does that.

The redo machinery relies on the page LSN to tell whether the update has
occurred. In the presence of torn pages, that's of course unreliable.

> Having raised that objection, ISTM that checking for torn pages can be
> accomplished reasonably well using a few rules...

I have zero confidence in this; the fact that you can think of
(incomplete, inaccurate) heuristics for heap-page operations doesn't
mean you can make it work for indexes.

regards, tom lane