Re: PD_ALL_VISIBLE flag was incorrectly set happend during repeatable vacuum

On Sat, Dec 25, 2010 at 2:09 PM, Maxim Boguk <maxim(dot)boguk(at)gmail(dot)com> wrote:
> While I trying create reproducible test case for BUG #5798 I
> encountered very strange effect on two of my servers (both servers
> have same hardware platform/OS (freebsd 7.2) and PostgreSQL 8.4.4).
>
> Very simple test table created as:
> CREATE TABLE test (id integer);
> INSERT INTO test select generate_series(0,10000);
>
> And I trying repeateble vacuum of that table with script:
>  perl -e "foreach (1..100000) {system \"psql -d test -h -c 'vacuum test'\";}"
>
> And once per like an minute (really random intervals can be 5 minutes
> without problems can be 3 vacuum in row show same error)  I getting
> next errors:
> WARNING:  PD_ALL_VISIBLE flag was incorrectly set in relation "test" page 1
> ...
> WARNING:  PD_ALL_VISIBLE flag was incorrectly set in relation "test"
> page 30 for all pages of the relation.

The same problem happened on my customer's PostgreSQL 8.4.2
environment. Here are the error messages:

Dec 13 13:11:59 test postgres[28249]: [2-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 38
Dec 13 13:11:59 test postgres[28249]: [2-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:11:59 test postgres[28249]: [3-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 61
Dec 13 13:11:59 test postgres[28249]: [3-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:11:59 test postgres[28249]: [4-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 88
Dec 13 13:11:59 test postgres[28249]: [4-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:11:59 test postgres[28249]: [5-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 96
Dec 13 13:11:59 test postgres[28249]: [5-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:11:59 test postgres[28249]: [6-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 98
Dec 13 13:11:59 test postgres[28249]: [6-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:11:59 test postgres[28249]: [7-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 107
Dec 13 13:11:59 test postgres[28249]: [7-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
...
Dec 13 13:15:59 test postgres[9640]: [2-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 38
Dec 13 13:15:59 test postgres[9640]: [2-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:15:59 test postgres[9640]: [3-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 61
Dec 13 13:15:59 test postgres[9640]: [3-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:15:59 test postgres[9640]: [4-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 88
Dec 13 13:15:59 test postgres[9640]: [4-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:15:59 test postgres[9640]: [5-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 93
Dec 13 13:15:59 test postgres[9640]: [5-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:15:59 test postgres[9640]: [6-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 96
Dec 13 13:15:59 test postgres[9640]: [6-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
Dec 13 13:15:59 test postgres[9640]: [7-1] WARNING: 01000:
PD_ALL_VISIBLE flag was incorrectly set in relation "pg_statistic"
page 107
Dec 13 13:15:59 test postgres[9640]: [7-2] LOCATION: lazy_scan_heap,
vacuumlazy.c:676
...

This problem was reported some times, but has not been resolved yet.
http://archives.postgresql.org/message-id/4C23A3CF.4080506%40frolix.muddywaters.org
http://archives.postgresql.org/message-id/g2o4b46b5f01004010610ib8625426uae6ee90ac1435ba1%40mail.gmail.com

Though I investigated the source code around PD_ALL_VISIBLE flag,
I could not identify the cause.

Does anyone have any ideas what the cause is?

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center