Re: 16-bit page checksums for 9.2

On Jan 6, 2012, at 4:36 AM, Andres Freund wrote:
> On Friday, January 06, 2012 11:30:53 AM Simon Riggs wrote:
>> On Fri, Jan 6, 2012 at 1:10 AM, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
>>> * Simon Riggs (simon(at)2ndQuadrant(dot)com) wrote:
>>>> I discover that non-all-zeroes holes are fairly common, just not very
>>>> frequent.
>>>
>>> Curious, might be interesting to find out why.
>>>
>>>> That may or may not be a problem, but not something to be dealt with
>>>> here and now.
>>>
>>> But I agree that it's not the job of this patch/effort. It sounds like
>>> we have clear indication, however, that those areas, as they are not
>>> necessairly all zeros, should be included in the checksum.
>>
>> Disagree. Full page writes ignore the hole, so its appropriate to do
>> so here also.
> Well, ignoriging them in fpw has clear space benefits. Ignoring them while
> checksumming doesn't have that much of a benefit.

I agree with Andres... we should checksum zero bytes, because if they're screwed up then something is wrong with your system, even if you got lucky with what data got trashed.

As I mentioned before, 2 separate checksums would be nice, but if we can't have that I think we need to fail on any checksum error.
--
Jim C. Nasby, Database Architect jim(at)nasby(dot)net
512.569.9461 (cell) http://jim.nasby.net