Mark Stosberg wrote:
> This post suggested a performance improvement to the initial copy might
> appear in 1.2, by removing indexes and re-applying them.
>
> http://gborg.postgresql.org/pipermail/slony1-general/2005-June/002285.html
>
> >From reading RELEASE-1.2.0, I didn't see a mention to that effect. Is an
> optimization like this in place?
>
> If not, is the recommended practice to drop all the non-primary-key
> indexes, and then re-apply them later?
>
That change is in place:
prepareTableForCopy(tab_id) added the code:
-- Setting pg_class.relhasindex to false will cause copy not to
-- maintain any indexes. At the end of the copy we will reenable
-- them and reindex the table. This bulk creating of indexes is
-- faster.
-- ----
update pg_class set relhasindex = ''f'' where oid = v_tab_oid;
finishTableAfterCopy(tab_id) added:
-- ----
-- Reenable indexes and reindex the table.
-- ----
update pg_class set relhasindex = ''t'' where oid = v_tab_oid;
execute ''reindex table '' ||
@[EMAIL PROTECTED](v_tab_fqname);
This was added in 1.1.5, before 1.2 was released...
As a result, "recommended practice" is to leave the indexes alone.
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general
