Thread: how to copy data between machines ?

how to copy data between machines ?

From
D.C.
Date:
Hi folks,

After searching for a while as to how to transfer data from machine to
machine, I came across the 'pg_dump' and 'pg_restore' commands.

On machine 'A', I did ..

    /usr/local/pgsql/bin/pg_dump test -f /tmp/blah

Then ...

    [socrates:/tmp] desmond% ls -l /tmp/blah
    -rw-r--r--  1 desmond  wheel  540 12 May 22:02 /tmp/blah

I then scp that file to machine B, and do ...

    pg_restore -d test /tmp/test

The result is ...

    pg_restore: [archiver] input file does not appear to be a valid
archive

Eh ???

What did I miss ?

I'm using the same version of postgreSQL on both machines, both
installed within minutes of each other.

Thanks in advance.

D.



Re: how to copy data between machines ?

From
D.C.
Date:
Le 12 mai 05, à 22:05, D.C. a écrit :

> pg_restore -d test /tmp/test

Sorry: /tmp/blah.  I didn't mix up the input files, but there are two
of them.  :-(  So the question still remains unresolved.  :-(

D.

--
Des Coughlan
coughlandesmond@yahoo.fr



Re: how to copy data between machines ?

From
Mark Greenbank
Date:
I think you need the '-F c' option in pg_dump to make pg_restore
understand the archive.
Mark

On 5/12/05, D.C. <coughlandesmond@yahoo.fr> wrote:
>
> Hi folks,
>
> After searching for a while as to how to transfer data from machine to
> machine, I came across the 'pg_dump' and 'pg_restore' commands.
>
> On machine 'A', I did ..
>
>     /usr/local/pgsql/bin/pg_dump test -f /tmp/blah
>
> Then ...
>
>     [socrates:/tmp] desmond% ls -l /tmp/blah
>     -rw-r--r--  1 desmond  wheel  540 12 May 22:02 /tmp/blah
>
> I then scp that file to machine B, and do ...
>
>     pg_restore -d test /tmp/test
>
> The result is ...
>
>     pg_restore: [archiver] input file does not appear to be a valid
> archive
>
> Eh ???
>
> What did I miss ?
>
> I'm using the same version of postgreSQL on both machines, both
> installed within minutes of each other.
>
> Thanks in advance.
>
> D.
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
>                http://archives.postgresql.org
>

Re: how to copy data between machines ?

From
D.C.
Date:
Le 12 mai 05, à 22:11, Mark Greenbank a écrit :

> I think you need the '-F c' option in pg_dump to make pg_restore
> understand the archive.

Thank you !!

D.



Re: how to copy data between machines ?

From
Tom Lane
Date:
"D.C." <coughlandesmond@yahoo.fr> writes:
> The result is ...
>     pg_restore: [archiver] input file does not appear to be a valid
> archive

> Eh ???

pg_restore is for restoring dumps that you make with the -Fc or -Ft
options to pg_dump.  The default output from pg_dump is a plain sql
script that you just feed to psql.

            regards, tom lane

Re: how to copy data between machines ?

From
Frank Bax
Date:
At 04:05 PM 5/12/05, D.C. wrote:
>After searching for a while as to how to transfer data from machine to
>machine, I came across the 'pg_dump' and 'pg_restore' commands.
>
>On machine 'A', I did ..
>
>    /usr/local/pgsql/bin/pg_dump test -f /tmp/blah
>
>Then ...
>
>    pg_restore -d test /tmp/test
>
>The result is ...
>
>    pg_restore: [archiver] input file does not appear to be a valid archive
>
>What did I miss ?


You forgot to read the manpage for pg_dump:

        -F format

        --format=format
               t      Output a tar archive suitable for input into
                      pg_restore. Using this archive format allows
                      reordering  and/or  exclusion of schema ele-
                      ments at the time the database is  restored.
                      It  is  also possible to limit which data is
                      reloaded at restore time.


Re: how to copy data between machines ?

From
D.C.
Date:
Le 12 mai 05, à 22:19, Tom Lane a écrit :

> "D.C." <coughlandesmond@yahoo.fr> writes:
>> The result is ...
>>     pg_restore: [archiver] input file does not appear to be a valid
>> archive
>
>> Eh ???
>
> pg_restore is for restoring dumps that you make with the -Fc or -Ft
> options to pg_dump.  The default output from pg_dump is a plain sql
> script that you just feed to psql.

Thanks again, and may I now make the most of your generosity by asking
another question ?  I got an error message when
./configuring postgreSQL, and now I find that I don't have 'history',
and I can't use the 'up' arrow to cycle through previous commands.  The
error was about 'readline', I think.

The installation I use is on an iBook, which might sound mad, but I
want a version to be able to take everywhere with me, and learn whilst
on the bus, in the train etc ..

Thanks again.

D.



Re: how to copy data between machines ?

From
Tom Lane
Date:
"D.C." <coughlandesmond@yahoo.fr> writes:
> Thanks again, and may I now make the most of your generosity by asking
> another question ?  I got an error message when
> ./configuring postgreSQL, and now I find that I don't have 'history',
> and I can't use the 'up' arrow to cycle through previous commands.  The
> error was about 'readline', I think.

Yup, you need libreadline to support that.

> The installation I use is on an iBook, which might sound mad, but I
> want a version to be able to take everywhere with me, and learn whilst
> on the bus, in the train etc ..

I use a Mac laptop too.  Personally, I get the readline sources
(check freshmeat.net for a pointer) and build/install them in the
default way, which drops 'em in /usr/local.  This is highly Not Mac
Approved I'm sure, but it gets the job done.  You could no doubt also
get readline from the fink project, if you'd rather have something
prebuilt (but then again why are you building PG by hand if that's
what you want ...)

            regards, tom lane

Re: how to copy data between machines ?

From
Peter Bierman
Date:
At 4:38 PM -0400 5/12/05, Tom Lane wrote:
>"D.C." <coughlandesmond@yahoo.fr> writes:
>>  Thanks again, and may I now make the most of your generosity by asking
>>  another question ?  I got an error message when
>>  ./configuring postgreSQL, and now I find that I don't have 'history',
>>  and I can't use the 'up' arrow to cycle through previous commands.  The
>>  error was about 'readline', I think.
>
>Yup, you need libreadline to support that.
>
>>  The installation I use is on an iBook, which might sound mad, but I
>>  want a version to be able to take everywhere with me, and learn whilst
>>  on the bus, in the train etc ..
>
>I use a Mac laptop too.  Personally, I get the readline sources
>(check freshmeat.net for a pointer) and build/install them in the
>default way, which drops 'em in /usr/local.  This is highly Not Mac
>Approved I'm sure, but it gets the job done.  You could no doubt also
>get readline from the fink project, if you'd rather have something
>prebuilt (but then again why are you building PG by hand if that's
>what you want ...)


There's nothing wrong with dropping stuff in /usr/local/ on a Mac.
That's how I do it too. Apple makes an effort not to stomp in
/usr/local for obvious reasons. :-)

-pmb

Re: how to copy data between machines ?

From
D.C.
Date:
Le 12 mai 05, à 22:38, Tom Lane a écrit :

> "D.C." <coughlandesmond@yahoo.fr> writes:
>> Thanks again, and may I now make the most of your generosity by asking
>> another question ?  I got an error message when
>> ./configuring postgreSQL, and now I find that I don't have 'history',
>> and I can't use the 'up' arrow to cycle through previous commands.
>> The
>> error was about 'readline', I think.
>
> Yup, you need libreadline to support that.
>
>> The installation I use is on an iBook, which might sound mad, but I
>> want a version to be able to take everywhere with me, and learn whilst
>> on the bus, in the train etc ..
>
> I use a Mac laptop too.  Personally, I get the readline sources
> (check freshmeat.net for a pointer) and build/install them in the
> default way, which drops 'em in /usr/local.  This is highly Not Mac
> Approved I'm sure, but it gets the job done.  You could no doubt also
> get readline from the fink project, if you'd rather have something
> prebuilt (but then again why are you building PG by hand if that's
> what you want ...)

Yep, that seems to work, thanks again !!

Running it on my iBook is brilliant fun: as I don't need it all the
time, a little 'wrapper' has gone into my $HOME/bin to 'su' start it up
when I want to run psql ...

I in fact was trained on oracle years ago when I worked in IT, but got
bored and walked out after two days.  Now a friend wants to start up a
small business, and I quite fancy setting up her DB with a
perl-generated webpage, to allow her to track customers, materials etc.
  She's not ready to take the plunge yet, and I'm just doing it for fun
-- and having loads of fun at the same time !

D.