Re: pg_dump compress - Mailing list pgsql-general

From Roger Niederland
Subject Re: pg_dump compress
Date
Msg-id 4E7DE62B.7090708@niederland.com
Whole thread Raw
In response to Re: pg_dump compress  (Adrian Klaver <adrian.klaver@gmail.com>)
Responses Re: pg_dump compress
Re: pg_dump compress
List pgsql-general

On 9/23/2011 5:18 PM, Adrian Klaver wrote:
> On Friday, September 23, 2011 7:26:19 am Roger Niederland wrote:
>> On 9/23/2011 6:46 AM, hubert depesz lubaczewski wrote:
>>> On Thu, Sep 22, 2011 at 11:00:10PM -0700, Roger Niederland wrote:
>>>> Using pg_dump from the command line with the exe included in windows
>>>> 32 bit install in postgresql 9.1.
>>>> pg_dump  -U username -f somefile.sql.gz -Z 9 -i dbname
>>> please note that pg_dump --help shows:
>>>     -Z, --compress=0-9          compression level for compressed formats
>>>
>>> hint: the part "for compressed formats" is critical.
>>>
>>> plain dump is not compressed, you'd have to use -Fc to get compression,
>>> but it's good to use anyway.
>>>
>>> Best regards,
>>>
>>> depesz
>> On previous versions the -Z 9  option would compress the plain text sql
>> output to a gzipped file.
>> Now it now longer compresses.  Using the -Fc the outputs a compressed
>> file, which is only readable
>> by pg_restore.
> Seems there has been a code change in pg_dump.c that ignores compression level
> for plain text.
>
>  From line 537:
>
> /*
>      * Ignore compression level for plain format. XXX: This is a bit
>      * inconsistent, tar-format throws an error instead.
>   */
>
> The documentation has not caught up with this change.
>
>> Roger

Just did not expect that the compression would be removed for plain files.
In case anyone cares... 7-zip.org provides a 32 bit and 64 bit program
which can utilize windows pipes.
Probably others available, but this one works for me.

Thanks,
Roger


pgsql-general by date:

Previous
From: Marti Raudsepp
Date:
Subject: Re: get number and names of processes connected to postgresql
Next
From: Timothy Garnett
Date:
Subject: Documentation of what schema modifications cause what level of table locking