<p><font size="2">Not just EDB :-)<br /><br /> - Luke<br /><br /> Msg is shrt cuz m on ma treo<br /><br />
-----OriginalMessage-----<br /> From: Chris Browne [<a
href="mailto:cbbrowne@acm.org">mailto:cbbrowne@acm.org</a>]<br/> Sent: Wednesday, April 04, 2007 10:07 PM Eastern
StandardTime<br /> To: pgsql-hackers@postgresql.org<br /> Subject: Re: [HACKERS] Modifying TOAST
thresholds<br/><br /> tgl@sss.pgh.pa.us (Tom Lane) writes:<br /> > Bruce Momjian <bruce@momjian.us> writes:<br
/>>> The big question is whether this is for 8.3 or 8.4.<br /> ><br /> > What I would definitely like to
seefor 8.3 is some performance testing<br /> > done to determine whether we ought to change the current defaults.<br
/>> (Both TOAST_TUPLES_PER_PAGE and EXTERN_TUPLES_PER_PAGE ought to be looked<br /> > at.)<br /> ><br /> >
Whetherit's possible to get the storage parameter in there depends on<br /> > how soon someone produces a patch.
Giventhat we understand this area<br /> > fairly well, I personally would be willing to give it a pass on the<br />
>"feature freeze" rule, as long as we have the patch by say mid-April.<br /><br /> If I had to end of April, I'd
volunteer. Alas, I will be unavailable<br /> between tomorrow and mid-April.<br /><br /> It seems to me that changing
thedenominator from 4 to something<br /> somewhat higher (16 comes to mind) would probably be "the patch," and<br />
wouldlikely be useful for a fair number of cases. I'd really like to<br /> see something like this (e.g. - something
thatwould start TOASTing<br /> for, say, tuples > 500 bytes) in 8.3; this seems to me to have the<br /> potential to
bea significant optimization that would be VERY cheap to<br /> implement. "Low hanging fruit," so to speak.<br /><br
/>EDB might like to use this as a route into a tunable "column store"<br /> some time later; I don't think that should
stopus from an improvement<br /> today...<br /> --<br /> "cbbrowne","@","linuxfinances.info"<br /><a
href="http://www3.sympatico.ca/cbbrowne/rdbms.html">http://www3.sympatico.ca/cbbrowne/rdbms.html</a><br/> "MSDOS didn't
getas bad as it is overnight -- it took over ten years<br /> of careful development." --
<dmeggins@aix1.uottawa.ca><br/><br /> ---------------------------(end of broadcast)---------------------------<br
/>TIP 4: Have you searched our list archives?<br /><br /> <a
href="http://archives.postgresql.org">http://archives.postgresql.org</a><br/><br /></font>
