On Sun, Aug 23, 2009 at 6:23 PM, Greg Stark<gsstark@mit.edu> wrote:
> On Sun, Aug 23, 2009 at 5:37 PM, Jeff Ross<jross@wykids.org> wrote:
>>
>> pg_clog is 32K. I've put it at http://www.openvistas.net/pg_clog
>
> Sorry, I'm getting a 404
For what it's worth this is what the heap dump shows. I don't
understand why you're seeing these tuples show up if they're all
properly marked as updated and XMAX_COMMITTED.
The last tuple is marked strangely I think. I don't think it's
supposed to have XMAX_INVALID if xmax is 0 but I could be wrong. Also,
I don't understand why it's marked as UPDATED if ctid and xmax aren't
set.
Could you also send
select (h).* from (select page_header(get_raw_page('pg_namespace',0))
as h) as x;
I'm wondering if the page allvisible flag is set. The visibility map
is one of the few 8.4 patches which impact transaction visibility.
lp lp_off flags lp_len t_xmin t_xmax field3 t_ctid t_info2 t_info
3 7616 1 141 2 897 0 (0,6) 16387 9482
HOT_UPDATED, UPDATED, XMAX_COMMITTED, XMIN_COMMITTED
6 7340 1 129 897 898 0 (0,7) -16381 9482
HOT_UPDATED, HEAP_ONLY_TUPLE, UPDATED, XMAX_COMMITTED, XMIN_COMMITTED
7 7220 1 117 898 899 0 (0,9) -16381 9482
HOT_UPDATED, HEAP_ONLY_TUPLE, UPDATED, XMAX_COMMITTED, XMIN_COMMITTED
9 7088 1 129 899 900 0 (0,10) -16381 9482
HOT_UPDATED, HEAP_ONLY_TUPLE, UPDATED, XMAX_COMMITTED, XMIN_COMMITTED
10 6944 1 141 900 0 0 (0,10) -32765 10506
UPDATED, XMAX_INVALID, XMIN_COMMITTED
--
greg
http://mit.edu/~gsstark/resume.pdf
