Re: BUG #6041: Unlogged table was created bad in slave node

From: Emanuel Calvo <postgres(dot)arg(at)gmail(dot)com>
To: Euler Taveira de Oliveira <euler(at)timbira(dot)com>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #6041: Unlogged table was created bad in slave node
Date: 2011-05-26 17:38:34
Message-ID: BANLkTikQBAFAEFRAu0jaPGP_UOaC30DeAQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs pgsql-hackers

2011/5/26 Euler Taveira de Oliveira <euler(at)timbira(dot)com>:
> Em 26-05-2011 08:37, Emanuel escreveu:
>
>> Description:        Unlogged table was created bad in slave node
>>
> Unlogged table contents are not available to slave nodes [1].
>

I know it. But the error raised isn't pretty nice for common users.
IMHO it could be an empty file with a message of WARNING level,
for notify administratros to watch up these tables.

>> postgres=# select * from voidy ;
>> ERROR:  could not open file "base/12071/17034": No existe el archivo o
>> directorio
>>
> I think we should emit the real cause in those cases, if possible (not too
> much overhead). The message would be "Unlogged table content is not
> available in standby server".
>

I think that detecting if table is unlogged && file not exists, rise
this message.
But the problem will comes when the standby is promoted to master, in that
case I didn't test what happens (file not exists, i think it will
couldn't insert data
directly or may force to the dba drop the table from catalog).

--
--
              Emanuel Calvo
              Helpame.com

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2011-05-26 22:27:29 Re: BUG #6021: There is no difference between default and empty access privileges with \dp
Previous Message Tom Lane 2011-05-26 16:34:34 Re: BUG #6041: Unlogged table was created bad in slave node

Browse pgsql-hackers by date

  From Date Subject
Next Message Alvaro Herrera 2011-05-26 17:41:40 Re: timezone GUC
Previous Message Kevin Grittner 2011-05-26 17:28:11 Re: [ADMIN] pg_class reltuples/relpages not updated by autovacuum/vacuum