Heikki Linnakangas <hlinnaka@iki.fi> writes:
> On 02/12/2025 19:59, Peter Geoghegan wrote:
>> These tests will increase nbtree code coverage quite a bit, which is a
>> nice bonus.
> Committed, thanks for the review!
In the past couple of days, scorpion and skink have failed
the nbtree_half_dead_pages test with identical symptoms [1][2]:
@@ -41,8 +41,6 @@
(1 row)
vacuum nbtree_half_dead_pages;
-ERROR: error triggered for injection point nbtree-leave-page-half-dead
-CONTEXT: while vacuuming index "nbtree_half_dead_pages_id_idx" of relation "public.nbtree_half_dead_pages"
SELECT injection_points_detach('nbtree-leave-page-half-dead');
injection_points_detach
-------------------------
@@ -67,7 +65,6 @@
-- Finish the deletion and re-check
vacuum nbtree_half_dead_pages;
-NOTICE: notice triggered for injection point nbtree-finish-half-dead-page-vacuum
select * from nbtree_half_dead_pages where id > 99998 and id < 120002;
id
--------
No idea why it took a month for this to show up. Could some
recent change have affected it?
regards, tom lane
[1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=scorpion&dt=2026-01-02%2004%3A54%3A38
[2] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=skink&dt=2025-12-31%2003%3A34%3A51
