Christopher Browne cbbrowne at ca.afilias.info
Wed Sep 24 09:25:37 PDT 2008
"Jérôme Jouanin" <jerome at jfg-networks.net> writes:
> Because Slony documentation warns : "Reasonable," in this context,
> is on the order of a dozen servers. If the number of servers grows
> beyond that, the cost of communications increases prohibitively, and
> the incremental benefits of having multiple servers will be falling
> off at that point.  Moreover, I think a remote site could not have
> the same level of network availability than a LAN, because of
> external constraints. And I think that isolation is preferable in
> this case.  Isn't it ?

If it's remote, then it's remote.

If you expect to have trouble getting data across the WAN to feed an
active Slony-I node, then I wouldn't expect log shipping to be
terribly much easier.  If either approach turns out to be problematic,
I'd expect the other to be, too.

Log shipping isn't particularly less fragile; it introduces additional
"moving parts" in that you need to move the log files across the WAN.

Actually, part of what I'd question is why it seems a good idea to
have 10 replicas in the first place, at the "main" site.
-- 
(format nil "~S@~S" "cbbrowne" "linuxfinances.info")
http://www3.sympatico.ca/cbbrowne/wp.html
"war is an inappropriate analogy; ``flame war'' is a misnomer.
 in any usenet exchange, the only casualty is time.
 there are better uses for regret."
    --thi <ttn at netcom.com>


More information about the Slony1-general mailing list