Re: Corrupt index

exactly what are your indexes on this, as well as any triggers?

Amir Becher wrote:
> Here is an example of the misbehavior that I am
> seeing:
>
> A simple table with 3 colums (and approximately a
> million records):
> column_a INTEGER PRIMARY KEY
> column_b DATE NOT NULL
> column_c DOUBLE NOT NULL
>
> When the index (primary key) gets corrupted, I get the
> following behavior when I do a simple SELECT (using
> the psql program):
>
> SELECT * FROM example_table WHERE (column_a = 12345);
>
> The result looks like:
>
> column_a | column_b | column_c
> --------------------------------
> 67890 | 2001-01-01 | 100
>
> The main point is that "column_a" is the wrong number.
>
> When I use the EXPLAIN command, it tells me that it's
> doing an index scan using the primary key index.
> Issuing a "REINDEX TABLE example_table" command fixes
> the problem.
>
> I usually detect the problem when an INSERT fails
> (batch insert fails due to duplicate key). I also
> delete a few thousand records before the inserts. All
> of the inserts and deletes are done throught the JDBC3
> driver.
>
> Please let me know what additional information you
> would like me to provide.
>
> --- Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
>>Amir Becher <abecher(at)yahoo(dot)com> writes:
>>
>>>I am experiencing frequent index corruptions
>>
>>(almost
>>
>>>daily).
>>
>>Please define your problem: what misbehavior are you
>>actually seeing?
>>Give us facts, not interpretations.
>>
>> regards, tom lane
>
>
>
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - File online, calculators, forms, and more
> http://tax.yahoo.com
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo(at)postgresql(dot)org
>