On 1 August 2011 20:53, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> writes:
>> OK, so I should split this into 2 patches?
>> Even without the compression, it's probably worth the 16 -> 10 byte
>> reduction that would result from removing the 2nd CTID in the UPDATE
>> case, and that part would be a pretty small patch.
> Yeah, my point exactly. The rest of it might or might not be worth the
> extra complication.
OK, here's a patch for the first bit - just removing the second CTID
in the UPDATE case, and including a sanity check of the new tuple's
xmin and cmin.
It passes all the regression tests. I also tested it by doing a 10M
row UPDATE x=x+1 on a deferrable PK, and it gave about the expected
reduction in memory usage, with no difference in run time.
I'll test out the additional compression separately.
In response to
pgsql-hackers by date
|Next:||From: Florian Pflug||Date: 2011-08-02 08:33:39|
|Subject: Re: WIP fix proposal for bug #6123|
|Previous:||From: Robert Haas||Date: 2011-08-02 03:10:24|
|Subject: Re: pgbench internal contention|