Thread: Question on Replication

Hi,
I just went through the install notes on erserver1.2 and saw a mention
in there that every table for replication requires a integer based
column. It also states that a new column is added for every table that
is replicated.

The tables I intend to replicate already have such a column. Can the
server be configured to use these instead so that the tables on both
server remain unchanged ?

What is the overhead in using replication and what is the timing of the
replication. I have a database and we add roughly 100k records during an
update operation.

I did not find a lot of information on that except 2 small documents in
distribution. can anyone point me to the site where there is a detailed
description?

Thanks
Alex

On Sun, Nov 09, 2003 at 12:59:00AM +0900, Alex wrote:
>
> The tables I intend to replicate already have such a column. Can the
> server be configured to use these instead so that the tables on both
> server remain unchanged ?

Yes.  I recently posted a brief discussion of how to do this to the
erserver list.  The archives are available on the gborg site.

> What is the overhead in using replication and what is the timing of the
> replication. I have a database and we add roughly 100k records during an
> update operation.

You will need lots of memory.  The replication canot happen, of
course, before the transaction commits.  You may find that it will
not keep up to 100k records in a single snapshot, unless you do that
infrequently.

> I did not find a lot of information on that except 2 small documents in
> distribution. can anyone point me to the site where there is a detailed
> description?

I suggest checking the mailing list.

A

--
----
Andrew Sullivan                         204-4141 Yonge Street
Afilias Canada                        Toronto, Ontario Canada
<andrew@libertyrms.info>                              M2P 2A8
                                         +1 416 646 3304 x110