Re: "page is not marked all-visible" warning in regression tests

On Tuesday, June 05, 2012 03:32:08 PM Tom Lane wrote:
> I got this last night in a perfectly standard build of HEAD:
>
> *** /home/tgl/pgsql/src/test/regress/expected/sanity_check.out Thu Jan 12
> 14:06:14 2012 ---
> /home/tgl/pgsql/src/test/regress/results/sanity_check.out Mon Jun 4
> 20:28:39 2012 ***************
> *** 1,4 ****
> --- 1,5 ----
> VACUUM;
> + WARNING: page is not marked all-visible but visibility map bit is set in
> relation "pg_db_role_setting" page 0 --
> -- sanity check, if we don't have indices the test will take years to
> -- complete. But skip TOAST relations (since they will have varying
>
> ======================================================================
>
> I didn't manage to reproduce it in a few tries, and I don't recall
> having seen the like reported from the buildfarm either. So there's
> some low-probability bug in there ...
I have seen that twice just yesterday. Couldn't reproduce it so far.
Workload was (pretty exactly):

initdb
postgres -c fsync=off
pgbench -i -s 100
CREATE TABLE data(id serial primary key, data int);
ALTER SEQUENCE data_id_seq INCREMENT 2;
VACUUM FREEZE;
normal shutdown
postgres -c fsync=on
pgbench -c 20 -j 20 -T 100
WARNING: ... pg_depend ...
WARNING: ... can't remember ...

Greetings,

Andres