The SMB redirector does not detect a disconnect and instead operates on a Bad_Network_Path response. Then, the client reaches the VNetRoot and transitions to offline.
When a user connects to a CA share, the redirector is instructed to aggressively try reconnection, as the share will probably be available again soon. This attempt is made through persistent server-based file handles on the cluster side. These handles are shared throughout all cluster nodes that host the specific file server resource. The increased time-out period should guarantee that the client's session survives the failover to a different node within the cluster.
- Do not store scale-out file server data, such as Hyper-V VHDX and SQL databases, on the same shared folders as user data, such as Microsoft Office files. Create separate shares for each scenario. If necessary, create two different shares that point to the same root folder, with each share configured to use different settings. Then, point your users to the appropriate share.
- If you store user data on a shared folder, and users are pinning that share as offline available to their clients, do not enable the Continuously available option on that share.
- If you want to use Continuously available functionality on a user data share, disable the offline files option on that share to prevent users from pinning data to their clients.
Article ID: 2981863 - Last Review: 1 Apr 2015 - Revision: 1