This is on a linux box, postgres 8.0.3 (yes, I know it's out of date but it was 
a test setup which seems to become live and I have not had time to upgrade the 
thing) which runs a daily process to crunch some logs and do some postGIS 
processing on them.

Tonight I got this message from my program:

Skipping the copy to disk
Done at Sat Jul  1 02:32:37 2006
Starting vacuum at Sat Jul  1 02:32:41 2006
Can't do the final vacuum and stats! ERROR:  could not access status of 
transaction 1081340514
SQL is <VACUUM ANALYZE;>

Running a vacuum verbose in the afflicted database got this complaint:
DETAIL:  0 dead row versions cannot be removed yet.
There were 19 unused item pointers.
0 pages are entirely empty.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
INFO:  analyzing "pg_catalog.pg_operator"
INFO:  "pg_operator": scanned 14 of 14 pages, containing 651 live rows and 0 
dead rows; 651 rows in sample, 651 estimated total rows
INFO:  vacuuming "public.weekly_log_data"
ERROR:  could not access status of transaction 1081340514
DETAIL:  could not open file "/data/postgres/psql8_test/pg_clog/0407": No such 
file or directory

Sure enough, there is no such beast:
[EMAIL PROTECTED] cmnd_logs]$ ls -lt /data/postgres/psql8_test/pg_clog/
total 1716
-rw-------    1 postgres postgres   155648 Jul  1 02:36 0007
-rw-------    1 postgres postgres   262144 Jun 15 02:09 0006
-rw-------    1 postgres postgres   262144 Apr 12 05:27 0005
-rw-------    1 postgres postgres   262144 Mar 12 02:41 0004
-rw-------    1 postgres postgres   262144 Feb 13 00:05 0003
-rw-------    1 postgres postgres   262144 Dec 28  2005 0002
-rw-------    1 postgres postgres   262144 Dec 18  2005 0001

=========================

I can reload the data with some tedium but not effort, but I am wondering if I 
might have run across a known bug in 8.0.3 ? Are there any things I should look 
at ? Or should I just start over on my week's logs on a newer, more up to date 
version ? (it's less than 5 hours of computer time to crunch one day of this 
stuff,and only a few minutes of operator time, so I can easily cut my losses 
and run...)

TIA,

Greg Williamson
DBA
GlobeXplorer LLC

---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?

               http://archives.postgresql.org

Reply via email to