On 9/21/07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
I'm also starting to come around to liking the page-header-xid field
a bit more. I suggest that it could replace the "page is prunable"
flag bit altogether --- to mark the page prunable, you must store
some appropriate xid into the header field. This would avoid a useless
prune attempt immediately after a page is marked prunable.
Agree. I was thinking of minimum of the following:
1 xmin of INSERT_IN_PROGRESS tuple
2 xmax of DELETE_IN_PROGRESS tuple
3 xmax of RECENTLY_DEAD_TUPLE
When we attempt a prune, we can set xid to the minimum of the above
three. If none of the above tuples exist in the page, xid can be set to FrozenXid
A page need not be pruned if its xid is set to FrozenXid or is greater
than OldtestXmin.
In addition, I guess we can even set the xid also when we insert,
delete or update from the page unless its already set to a lower value.
Thanks,
Pavan
--
Pavan Deolasee
EnterpriseDB
http://www.enterprisedb.com