Re: crash-safe visibility map, take three

On Tue, Nov 30, 2010 at 11:59 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> On Tue, Nov 30, 2010 at 11:40 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> Ouch.  That seems like it could shoot down all these proposals.  There
>>> definitely isn't any way to make VM crash-safe if there is no WAL-driven
>>> mechanism for setting the bits.
>
>> Heikki's intent method works fine, because the WAL record only clears
>> the visibility map bits on redo; it never sets them.
>
> Uh, no, because he also had that final WAL record that would set the
> bits.

Well, as already discussed upthread, that WAL record causes some other
problems, so make it Heikki's intent method, without the final WAL
record that breaks things.

>> We could actually allow the slave to set the visibility map bits based
>> on its own xmin horizon.
>
> Not in a crash-safe way, which is exactly the problem here.

Brilliant selective quoting.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company