I have three servers; two are inside our firewall and are in a cluster, the third is not part of the cluster and is outside the firewall (actually inside the DMZ but allowed to face the web).
The customer facing server dbs are required to be read only and nothing from those dbs is allowed to replicate internally.
I have the replication for all dbs on the customer facing server set up to not replicate replicate ACLs from any server so I can maintain different ACLs inside and outside the firewall.
Last week, I modified one of the internal db's ACLs and the changes replicated to the external db cutting of access from the web.
I modified the external ACL back to what it should be, forced a two way replication, and everything remainded correct on both the internal and external dbs.
This happened once some time ago; no root cause was ever identified
I found this post but I really need to have separate ACLs internally and externally.
Any idea why the internal ACL overwrote the external one and, more importantly, how can I assure this doesn't happen again?
TIA for your thoughts.