sql server replication: Greetings:

I have a situation where I need to replicate somewhere between 40 and 80
databases between two SQL Server 2000 clusters. I've already built a
40-database set of transactionally replicated items as a test (I should
mention these are fairly complex databases, with about 850 tables, 3000 views
and so on) and this works, after a fashion. System startup takes a good 15
minutes and I'm having a little trouble getting the logreader agents to all
come up correctly. I think I can solve most of this with tuning and by adding
a distributor machine. (Right now, the Publisher is also the Distributor and
I'm using all Pull subscriptions at the Subscriber.) What I don't have any
faith in is being able to go much beyond this, in case I really do end up
needing to replicate 80 active databases between the two clusters.

My question is this: Is there a programmatic or practical limit to the
number of databases you can replicate from a single SQL Server 2000 instance?
I'm running an active/passive OLTP cluster (4-CPU HP Opteron) and sending the
data to another active/passive OLAP cluster running the same hardware. I have
a gigabit connection between the clusters. (In case any of that helps scope
the response.) As I said, I've managed to get the 40-database transactional
replication system running, but it seems a little.... fragile.

Any help would be welcome in terms of suggestions for tuning or approaches I
might take to make the larger 80-database replication work. I'm very
concerned about the 'nominal' system load this many replication streams will
entail and also the I/O overhead associated with this activity. BTW: Both
clusters share a large SAN storage system.

Thank you in advance,

This article should help for topology/hardware tuning and optimization:
http://www.microsoft.com/technet/prodtechnol/sql/2000/maintain/tranrepl.mspx
Rgds,
Paul Ibison SQL Server MVP, www.replicationanswers.com
(recommended sql server 2000 replication book:
http://www.nwsu.com/0974973602p.html)

Yes. I've read that document and many others. I also have the SQL Server High
Availability and Performance Tuning books from MS Press. That's why I say I'm
confident I can sort out the issues with my 40-database replication model and
solve the fragility issues I'm seeing. My question was "Is there a
programmatic or practical limit to the number of databases you can replicate
from a single SQL Server 2000 instance?" My company may be about to make a
bad business decision in trying replicating 80 or more databases and I'm
wondering if that is even feasible or advisable. I'm hoping someone provide
me with information about the performance limits of the replication 'engine'.

[quoted text, click to view]