Re: How to transfer from place to plase without backup/restore - Mailing list pgsql-general

From Hristo Filipov
Subject Re: How to transfer from place to plase without backup/restore
Date
Msg-id 000901c81eae$3cb225f0$b61671d0$@net
Whole thread Raw
In response to Re: How to transfer from place to plase without backup/restore  (Shane Ambler <pgsql@Sheeky.Biz>)
Responses Re: How to transfer from place to plase without backup/restore  (Alvaro Herrera <alvherre@commandprompt.com>)
List pgsql-general
Hi Shane,
Thanks for the answer.
Actually I want to distribute specific database to a lot of different
clients(different OS/Linux;Windows/, different CPU/AMD;Intel/), and I was
thinking about only copying the files instead of dump/restore(which is must
slower operation). From your answer I guess it is impossible for now :-(.
Is there a way to tell PostgreSQL not install itself with OS or/and CPU
compatibility(for the prize of loosing performance), but that way the one
can move files free from on Computer to another?

Hristo Filipov

-----Original Message-----
From: Shane Ambler [mailto:pgsql@Sheeky.Biz]
Sent: 23 ???????? 2007 11:58
To: Hristo Filipov
Cc: pgsql-general@postgresql.org
Subject: Re: [GENERAL] How to transfer from place to plase without
backup/restore

Hristo Filipov wrote:
> Is it possible to transfer database from one place to another without
making
> backup from the source and restore at the target?
>
> I mean I want to copy the files as they are at the source to the target?
> What else I must do, so I can attach them to the target PostgreSQL
database?

That is basically what you are doing when you transfer (unless you are
moving the drive to a new machine) - how you do the transfer will depend
on your needs and what hardware and software is at the source and
destination.

If you are looking for a backup server to be ready to use as soon as
your production server fails then you can look at PITR or
synchronization such as slony-1

If you want to copy your database to a development machine occasionally,
then if they have the same hardware and postgres version then you should
be able to copy the data files across while the production server is
stopped without problems. Some differences between CPU and postgres
version can stop this from working.

You could also do the dump/restore directly from one machine to another
if they are on the same network without having the dump file to transfer
across and import which will give similar results.

eg. instead of -
pg_dumpall > mybackupfile
psql -h my.other.machine.address < mybackupfile

you can -

pg_dumpall | psql -h my.other.machine.address

If you have different cpu's (one Intel and one AMD64) or different
postgres versions then you will have to do the dump/restore even if it
is sent straight from one to another.

For more help specify what hardware, system and postgres version you
have on each machine and what you wish to achieve in transferring.



--

Shane Ambler
pgSQL@Sheeky.Biz

Get Sheeky @ http://Sheeky.Biz


pgsql-general by date:

Previous
From: Kevin Hunter
Date:
Subject: What makes a Postgres DBA?
Next
From: Greg Smith
Date:
Subject: Re: young guy wanting (Postgres DBA) ammo