Version 7.2.3 Vacuum abnormality

From: Paul L Daniels <pldaniels(at)pldaniels(dot)com>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Version 7.2.3 Vacuum abnormality
Date: 2003-02-13 05:27:01
Message-ID: 20030213152701.37b7a260.pldaniels@pldaniels.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Good evening, tonight while running my routine vacuum, the following came up on my screen:

---8<---------------
NOTICE: Rel xamefiles: Uninitialized page 708135 - fixing
NOTICE: Rel xamefiles: Uninitialized page 708136 - fixing
NOTICE: Rel xamefiles: Uninitialized page 708137 - fixing
NOTICE: Rel xamefiles: Uninitialized page 708138 - fixing
NOTICE: Rel xamefiles: Uninitialized page 708599 - fixing
---8<---------------

There were a lot more than this, several hundered.

Here's the general details of the DB:
- Approximately 30,000,000 rows
- No triggers
- Table consists strictly of text, integers and one key ( SERIAL )
- PostgreSQL 7.2.3 ( yes, i'll update to 7.2.4 in the next day or so )
- Pentium III 1.2Ghz with 1Gb RAM running RedHat 8.0 ( Not my machine! )

Here's the lead up to events:

- Two days ago I DELETE'd approximately 7 million rows
- I proceeded to vacuum, but it was 'terminated' by another admin approximately 12 hours later
- I restarted the vacuum, which resulted in the following stats:

NOTICE: Pages 701193: Changed 8459, Empty 0; Tup 21042082: Vac 0, Keep 0, UnUsed 32056923

- I ran the DELETE script again which purged another 200,000 rows ( approx )
- I proceeded to vacuum, the table in question returned the following stats:

NOTICE: Pages 704754: Changed 9599, Empty 0; Tup 14385034: Vac 0, Keep 207650, UnUsed 38884420.

- Today I dropped an index off the table, ran the DELETE again, removing 457,636 rows
- Vacuum dumped the above NOTICES. Final output for the table is:

NOTICE: Index xamefiles_k_key: Pages 187175; Tuples 14521716: Deleted 666334.
CPU 8.92s/14.11u sec elapsed 1306.20 sec.

- The vacuum is still going along fine.

Backups are proceeding without incident ( using pg_dump )
DELETE's are proceeding without incident

Hope this is enough information.

Kind Regards.

--
Paul L Daniels http://www.pldaniels.com
Linux/Unix systems Internet Development
ICQ#103642862,AOL:cinflex,IRC:inflex
A.B.N. 19 500 721 806

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2003-02-13 05:30:07 Re: [HACKERS] More benchmarking of wal_buffers
Previous Message Christopher Kings-Lynne 2003-02-13 05:16:16 More benchmarking of wal_buffers