Re: Still a bug in the VACUUM ??? !!!

From: Andreas Schmitz <a(dot)schmitz(at)cityweb(dot)de>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: Still a bug in the VACUUM ??? !!!
Date: 2003-02-28 09:16:29
Message-ID: 200302281016.29799.a.schmitz@cityweb.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin pgsql-hackers


> Ultimately, we need a way to reproduce the problem for debugging.
> If it is a Postgres bug, it should be possible to reproduce it.
>
> regards, tom lane

Hello,

the vacuum seems to be ok. I discovered that the "cluster" kills the data for
some reason. I could reproduce it on to sun (SS20 sol9, e450 sol8, redhat 7.3
-> postgresql-7.3.2). I took the same dump to all of these machines und ran
the cluster command.

tables: multicom(pk)->multithumb(fk) (see attached schema)

I checked before the cluster run:

select id from multithumb where id not in (select id from multicom) order by
id desc;

and it returned 0 row(s).

after this I ran "cluster multicom_txt_t_idx on multicom;" and repeated the
statement above. 33 rows of data from the multicom table where gon and 32
references in multithumb were still there. the same 33 rows (id and data)
were killed on all three databases. I tried it a few times and I always got
the output. same id's same data.

I picked the id's and had a look at the data.

select * from multicom where id=18379;
id | zulieferer | prioritaet | rubrik | kategorie | datum
| kennung | dachzeile | ueberschrift | unterzeile |
autor | zwischentitel | vorspann | kurztext | gueltig_bis |
region | dateiname | volltext
| titleidx
-------+------------+------------+------------+-----------+------------------------+------------------------------+-----------+--------------+------------+-------+---------------+----------+----------+------------------------+----------+-------------------------------------------+--------------------------------+----------
18379 | otz | 10 | Wirtschaft | BIL | 2003-02-26
19:47:00+01 | onbildotzBILWirNational37676 | | |
| | | | | 2003-03-02 19:47:00+01 |
National | /var/data/multikom/otz/m/WIRT/bulaptop.as | <p
class="contentfliess"> </p> |
(1 row)

No txtidx was build on that datarows 'cause no long text was inserted. I
filled the empty fields with the "<NODATA>" string. But it didn't help.

can someone give me some advise on this.

thanks and regards

-andreas

--
Andreas Schmitz - Phone +49 201 8501 318
Cityweb-Technik-Service-Gesellschaft mbH
Friedrichstr. 12 - Fax +49 201 8501 104
45128 Essen - email a(dot)schmitz(at)cityweb(dot)de

In response to

Responses

Browse pgsql-admin by date

  From Date Subject
Next Message Andreas Schmitz 2003-02-28 09:17:40 Re: Still a bug in the VACUUM ??? !!! (SCHEMA)
Previous Message Michael A. Schulte 2003-02-28 08:57:51 Re: ExclusiveLock and Python

Browse pgsql-hackers by date

  From Date Subject
Next Message Andreas Schmitz 2003-02-28 09:17:40 Re: Still a bug in the VACUUM ??? !!! (SCHEMA)
Previous Message Christoph Haller 2003-02-28 09:04:55 Re: Can pessimistic locking be emulated?