Re: unable to create new table - Mailing list pgsql-admin

From Greg Stark
Subject Re: unable to create new table
Date
Msg-id 407d949e0906292057nd6b270cwd39e63d4e73f868@mail.gmail.com
Whole thread Raw
In response to Re: unable to create new table  (Scott Marlowe <scott.marlowe@gmail.com>)
Responses Re: unable to create new table  (Scott Marlowe <scott.marlowe@gmail.com>)
List pgsql-admin
On Mon, Jun 29, 2009 at 4:56 PM, Scott Marlowe<scott.marlowe@gmail.com> wrote:
>> ERROR: invalid page header in block 95 of relation "pg_attribute".
>
> Are you running aanti-virus software on this machine?  That's a common
> cause of these issus on windows.

Is it? I thought it mainly caused random errors when connecting or
opening files. Not data corruption on disk.

I would be more inclined to blame bad memory or a bad hard drive,
though that also seems a bit unlikely to bite so quickly.

Are the column names sensitive data? Could you send your pg_attributes
data file? The data file in question will be named
<datadir>/base/11510/1249. (Actually the 11510 might vary, but if
you've only created one database it'll be the one that isn't "1".) If
you can extract bytes 778240-786432 then great, otherwise the whole
file probably isn't very big.

Hm, actually if you've gotten as far as 800k already in pg_attributed
then you've done a whole lot more than create four tables or something
else has added a ton of extra data to this file. Incidentally my
regression database has a pg_attribute that's just about that size.
Have you run "make check"?

--
greg
http://mit.edu/~gsstark/resume.pdf

pgsql-admin by date:

Previous
From: Kevin Kempter
Date:
Subject: how much memory (work_mem) is a query using?
Next
From: Scott Marlowe
Date:
Subject: Re: unable to create new table