On Fri, Apr 20, 2001 at 04:53:43PM -0700, G. Anthony Reina wrote:
> Nathan Myers wrote:
>
> > Does the replication have to be reliable? Are you equipped to
> > reconcile databases that have got out of sync, when it's not?
> > Will the different labs ever try to update the same existing
> > record, or insert conflicting (unique-key) records?
>
> (1) Yes, of course. (2) Willing--yes; equipped--dunno. (3) Yes,
> probably.
Hmm, good luck. Replication, by itself, is not hard, but it's only
a tiny part of the job. Most of the job is in handling failures
and conflicts correctly, for some (usually enormous) definition of
"correctly".
> > Reliable WAN replication is harder. Most of the proprietary database
> > companies will tell you they can do it, but their customers will tell
> > you they can't.
>
> Joel Burton suggested the rserv utility. I don't know how well it would
> work over a wide network.
The point about WANs is that things which work nicely in the lab, on a
LAN, behave very differently when the communication medium is, like the
Internet, only fitfully reliable. You will tend to have events occurring
in unexpected order, and communications lost, and queues topping over,
and conflicting entries in different instances which you must somehow
reconcile after the fact. Reconciliation by shipping the whole database
across the WAN is often impractical, particularly when you're trying to
use it at the same time.
WAN replication is an important part of Zembu's business, and it's hard.
I would expect the rserv utility (about which I admit I know little) not
to have been designed for the job.
Nathan Myers
ncm@zembu.com
