We are configuring DFS-R to act as a high availability service for our file shares. I realize there are some caveats but I can live with those. What I am looking to do though is reduce the timeout value for clients from the default two minutes to seconds
- even if this increases network load.
IE. A client connects to a dfs share that is replicated and works in the share through Explorer. The server the client is connected to, drops off line or is shutdown for maintenance, the users connected to that server then experience the hour glass for up to two minutes until the referral ttl expires and is processed and the client connects to the second server that also hosts the dfs share. Its this hour glass interruption and long failover time our users are complaining about because if they try and click another folder they get the Windows is Not Responding. If they are trying to access a mapped drive then obviously they can click and wait but we are looking for minimal interruption. Yes, clustering is an option but we are starting to span campuses and dfs-r reduces the spof at the share level in a cluster.
Any pointers would be great!
IE. A client connects to a dfs share that is replicated and works in the share through Explorer. The server the client is connected to, drops off line or is shutdown for maintenance, the users connected to that server then experience the hour glass for up to two minutes until the referral ttl expires and is processed and the client connects to the second server that also hosts the dfs share. Its this hour glass interruption and long failover time our users are complaining about because if they try and click another folder they get the Windows is Not Responding. If they are trying to access a mapped drive then obviously they can click and wait but we are looking for minimal interruption. Yes, clustering is an option but we are starting to span campuses and dfs-r reduces the spof at the share level in a cluster.
Any pointers would be great!