2009/11/24 RD黄永卫 <yongwei_huang@gtmc.com.cn>:
>
> 发件人: Scott Marlowe [mailto:scott.marlowe@gmail.com]
> 发送时间: 2009年11月25日 14:44
> 收件人: RD黄永卫
> 抄送: pgsql-bugs@postgresql.org; pgsql-general@postgresql.org
> 主题: Re: [GENERAL] How duplicate data produce when a UNIQUE index exite ?
>
> 2009/11/24 RD黄永卫 <yongwei_huang@gtmc.com.cn>:
>> Dear all,
>>
>>
>>
>> When "reindex" operation do,this error occurred:
>> Nov 22 10:22:27 SUC11 postgres[14145]: [2-1] ERROR: could not create unique
>> index
>> Nov 22 10:22:27 SUC11 postgres[14145]: [2-2] DETAIL: Table contains
>> duplicated values.
>> Nov 22 10:22:27 SUC11 postgres[14145]: [2-3] STATEMENT: REINDEX TABLE
>> "t_sfh_history"
>>
>> My question is:
>>
>> How the Duplicate datas produce when a UNIQUE index exite ?
>> Is this a bug of PostgreSQL ?
>
>
>
>>> My postgreSQL version is :
>>> PostgreSQL 8.1.4 on i686-pc-linux-gnu, compiled by GCC gcc (GCC) 3.2.3
>>> 20030502 (Red Hat Linux 3.2.3-56)
>
>> You are aware that 8.1.x is up to 8.1.18, right, that's 14 or so
>> updates you're missing. Could one of them have fixed a bug that
>> causes this? Sure. Look through the changelogs to be sure, but why
>> not just keep your pgsql version up to date? It's easier.
>>
>> Or are you running some redhat version that stays the same number
>> while getting bugs back ported to it or something?
>
> Either , or a pgsql bug can do this.
>
> What kind of " bad hardware " cause that ? May I reproduce that ?
Any kind of bad hardware. Bad memory, cpu, motherboard, RAID
controller, possibly the OS and on and on.
memtest86+ is a good starting place to see if your hardware has cpu /
ram /mobo issues.
But first, update pgsql. Then hunt down your duplicates and fix the
data in the database. I'm pretty sure there were a few problems like
this in the early 8.1 series releases.
