We have a clustered environment, 3 clusters with 2 servers (1 production, 1 failover) in each cluster. Cluster #1 has the resreservations.nsf file and this is where we make our meeting room reservations. We are having an issue in which 1 room is double booking. My question is this - does the RNRMGR task have to be running on the other clusters if there is no resreservations.nsf db on the servers? Is Calconn what is needed for calendaring to work or does rnrmgr need to be running as well? Our environment is setup so that one user manages the resreservations.nsf, therefore no other users access. Any advice is greatly appreciated.
Subject: RNRMGR Task in Domino
HI,
see double booking section
Date
Topic
2008-08-25
Documents related to Reservation database (Set up, move, troubleshoot, etc)
From IBM Developer
Figure 4. Multiple replicas of an R&R database
In addition to allowing replica copies of the R&R database, we’ve designed RnRMgr to actively use Domino clusters to perform application failover of R&R request processing. This means that if one server where RnRMgr is running becomes unavailable for a period of time, RnRMgr on its cluster mates will attempt to take control and process pending R&R reservation requests for all R&R databases that are shared between those servers.
For Notes/Domino 7, we have added the concept of a “primary” server for an R&R database. The primary server is where request processing normally happens. It is the same as the home server for the room in the Domino Directory, and it is where C&S requests are normally routed to and delivered. All other servers that have a replica are “secondary” servers. While there is an RnRMgr running on the secondary server, it does not process any requests in the R&R database. Instead, it simply keeps an eye on the primary server. When the primary becomes unresponsive long enough to be considered “offline,” the secondary server will begin to process any unprocessed reservation requests for any R&R databases that were on the primary server. This means the secondary server has become an “acting primary” server. When the primary server restarts, it will check the failover status of its R&R databases before it tries to begin work. If there is an acting primary server, then the primary server will do nothing with that database. If there is no acting primary server for that database, then the primary server will begin processing reservation requests as it normally does.
JYR
Subject: Also
Thanks JYR for the links!
To answer specifically:
My question is this - does the RNRMGR task have to be running on the other clusters if there is no resreservations.nsf db on the servers? No - RNRMgr only needs to be running on a server if there is a resource reservation database on the server.
Is Calconn what is needed for calendaring to work or does rnrmgr need to be running as well? - CalConn is used to support only end users since Domino 7. RNRMgr picks up this responsibility for Resources, so if there are users, you need Sched and CalConn. If there is a resource database, you also need RNRMgr. If there are no user mail files, you do not need Sched or CalConn.
Our environment is setup so that one user manages the resreservations.nsf, therefore no other users access.
Hmmm. Since the system is set up using mail-in databases, users can try to book rooms directly from their calendars. RNRMgr will sort these conflicts out and if all rooms are controlled by an owner, users will get a decline if they try this.
I’d start with this technote - which is in JYR’s list. RnRMgr is very good about avoiding double-booking unless something is wrong.
http://www-01.ibm.com/support/docview.wss?rs=203&uid=swg21255213
Subject: RE: Also
Many thanks to you both! This will help a lot. I appreciate your responses.