Обсуждение: DB replicators comparison; (vs. DB upgrade via pg_dump)

Hello,

I am about to upgrade from 7.3.4 to 8.0.3, and I read that using a DB
replication tool is a good way to go about it.

I see there are multiple replicator choices:
  Slony-I
  Daffodil Replicator
  Mammoth Replicator

I was wondering if anyone has tried any 2 or 3 of them, and could share
the experiences.

My upgrade will be over WAN (old DB and new DB are in different hosting
facilities), and it's for a fairly small DB:

# du -sh ~postgres/data
4.6G    /var/lib/pgsql/data

Oh, and couldn't I just pg_dump my old DB, scp the file over to the new
DB host, and use psql -f dump-file-here ?  Seems simpler than messing
with replicators.

Thanks,
Otis

A long time ago, in a galaxy far, far away, ogjunk-pgjedan@yahoo.com wrote:
> I am about to upgrade from 7.3.4 to 8.0.3, and I read that using a DB
> replication tool is a good way to go about it.

Replication is useful for this purpose if it helps you cut down on the
time the transition takes.

If the database is relatively small, and it only takes a few minutes
to load the data in, then it's far simpler to do a dump/reload.

But if you discover that the dump/reload takes 4 hours, and you're
only permitted 20 minutes of downtime, the complexities of replication
may get forced on you...
--
output = ("cbbrowne" "@" "ntlug.org")
http://linuxdatabases.info/info/slony.html
Sheridan:"So how did you find out all of this?"
Bester:"I'm a telepath.  Work it out." <*>

Hello,

Thanks for the explanation :)  I thought there was some technical
limitation when dump/reload is used, and I just wasn't seeing it after
my dump/reload successfully got me from 7.3.4 to 8.0.3.

Still, DB replication sounds very useful, so I'd still be interested in
Slony-I, Daffodil, and Mammoth comparison, if anyone has used more than
one of them.

Thanks,
Otis

--- Christopher Browne <cbbrowne@acm.org> wrote:
> A long time ago, in a galaxy far, far away, ogjunk-pgjedan@yahoo.com
> wrote:
> > I am about to upgrade from 7.3.4 to 8.0.3, and I read that using a
> DB
> > replication tool is a good way to go about it.
>
> Replication is useful for this purpose if it helps you cut down on
> the
> time the transition takes.
>
> If the database is relatively small, and it only takes a few minutes
> to load the data in, then it's far simpler to do a dump/reload.
>
> But if you discover that the dump/reload takes 4 hours, and you're
> only permitted 20 minutes of downtime, the complexities of
> replication
> may get forced on you...
> --
> output = ("cbbrowne" "@" "ntlug.org")
> http://linuxdatabases.info/info/slony.html
> Sheridan:"So how did you find out all of this?"
> Bester:"I'm a telepath.  Work it out." <*>
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 7: don't forget to increase your free space map settings
>