Long hangs when a node is killed, but not failed over

When running the CouchbaseAspNetSample, pointing to a 2 node cluster

1. Run
2. Create a user / log in
3. Ensure it works several requests (reloads are fine)
4. Shut down one node and it seems to tolerate for a little while
5. Notice that (after a few minutes), the requests get ridiculously slow. It appears that timeouts are not being respected (I see slowness in the Get and Upsert calls as well as other slowness I can't find). Timeouts do not seem to be respected.

You may have to shut down the alternate couchbase node or try multiple users.