Re: XID wraparound and busy databases

Tom Lane wrote:
> Bruce Momjian <bruce(at)momjian(dot)us> writes:
>> Is enlarging the xid field something we should consider for 8.4?
>
> No. We just got the tuple header down to 24 bytes, we are not going
> to give that back and then some.
>
> If you are processing 6K transactions per second, you can afford to
> vacuum every couple days... and probably need to vacuum much more often
> than that anyway, to avoid table bloat.
>
> Possibly your respondent should think about trying to do more than one
> thing per transaction?

I'm wondering how many of those 6k xacts/second are actually modifying
data. If a large percentage of those are readonly queries, than the need
for vacuuming could be reduced if postgres assigned an xid only if that
xid really hits the disk. Otherwise (for purely select-type queries) it
could use some special xid value.

This is what I'm doing in my Readonly-Queries-On-PITR-Slave patch.

greetings, Florian Pflug