<postgres at dac.e4ward.com> writes:
> On Nov 22, 2007 5:14 PM,  <slony1-general at lists.slony.info> wrote:
>> On 22/11/2007 19:08, postgres at dac.e4ward.com wrote:
>> > documentation, but it seems slonik is always suposed to be installed
>> > and run from the same server where postgres is installed.
>>
>> That's not my understanding....in particular, since connection strings
>> passed to slon always include host and port fields, I don't see why they
>> need to be on the same machine as the DB.
>
>
> I understand that and in fact, as I said, I'm successfully running the
> slave slon daemon from the master server, remotely. BUT I happen to
> have postgres installed on that server too.
>
> My problem is not where to run slon, because I'm running it always
> from servers where postgres is running too. My problem is to execute
> slonik remotely from a server where I don't have neither postgres or
> slon suning.
>
> Do I explain myself?

Yes, you're explaining yourself...

You will have a difficult time running slonik on a host that doesn't
have *any* of PostgreSQL running, because it will typically be
dynamically linked to the PostgreSQL client library, libpq, and you
will therefore need to have at least that part (libpq) of PostgreSQL
installed.

If your distribution has separate packages for:
a) Server components (e.g. - postmaster)
b) Libraries (e.g. - libpq)
c) Client components (e.g. - psql, pg_dump, ...)

then you could presumably avoid installing the server components and
still be able to get Slony-I components to work.  I'd certainly want
to have PostgreSQL "client components" like psql and pg_dump handy,
though...
-- 
(reverse (concatenate 'string "moc.enworbbc" "@" "enworbbc"))
http://linuxfinances.info/info/x.html
"How can you dream the impossible dream when you can't get any sleep?"
-- Sam Robb