On Thu, 2009-01-22 at 18:13 +0200, Heikki Linnakangas wrote:
> Simon Riggs wrote:
> > On Thu, 2009-01-22 at 16:15 +0200, Heikki Linnakangas wrote:
> >> That might be useful. But again, could just as well be implemented as an
> >> external tool like pglesslog.
> >
> > There is no WAL record for "no-op", at least not one of variable length.
>
> Hmm, maybe there should be? That seems like a useful thing to have for
> external tools.
>
> > The WAL files can't just have chunks of zeroes in the middle of them,
> > they must be CRC valid and chained together in the exact byte position.
> > There isn't any way to do this, even if there were, that's a seriously
> > complex way of doing that.
>
> Hmm, I think you could remove the records in the middle, rechain the
> remaining ones, recalculate the crc, and put an xlog switch record at
> the end. I agree that's seriously complicated, a no-op record would be
> much simpler.
If someone else suggested that mechanism you'd laugh and rip it to
shreds in an instant.
You are brilliant at seeing simple, practical ways of doing things and
that just ain't one of them. That's why for me this looks less and less
like a debate to determine the best way forwards.
I'm happy that you've chosen to spend your time on HS and I think we
should both return to that, for a rest. I'll be posting a new version
shortly.
-- Simon Riggs www.2ndQuadrant.comPostgreSQL Training, Services and Support