I want to know from any Lotus Admin out there who runs a ‘clustered’ Domino/Notes environment:
In the event of a failover of a client’s mailbox from Domino-A to Domino-B, what has your experience been like? What challenges or roadblocks have you hit?
Our environment, although stable, sometimes needs to failover clients due to any reason (i.e. human error, network blip, etc). On many clients, it is not a seemless transition from one mailserver to another.
Are there any best practices that can make this as seemless as possible? What can clients be told about “expectations” of their experience during a failover?
Subject: Domino Client failover - The Real World Experience
just some thoughts …
An important point to remember is if a document is open on the server that has failed you must close the document in order to fail over to a replica on the clustered server. In my opinion, the high availability clustering provides is well worth some of its quirks. All of our mail clients work off local email replicas so for email fail over is seamless. Getting users comfortable with local email is the hard part. outlining the benefits for them will help in the transition from server based email. Replication tends to be a sticking point, but in my opinion it (replication) is actually one of the important reasons we use notes. Once our road warriors got use to local email they will now complain if for whatever reason they aren’t setup for local email. We utilize mobile lightweight directories too so email addressing is done locally thus a server being down isn’t even a consideration for address look-ups. Clustering applications is less seamless but again in my opinion still worth it.
Subject: Domino Client failover - The Real World Experience
One thing I did was to create a policy that I sent out to set the users secondary server in there location documents. So that if they failed over, that the client would know to search that server for address information.
For the most part, I don’t really get to many complaints when the server fails over. I have shut the one server down in the middle of buisness hours and not heard any complaints from it. I have worked in both enviroments, clustered and non. I prefer clustered. Saves allot of problem calls.