What's the difference between the LooksAlive and IsAlive checks that the Microsoft Cluster service performs on Exchange services?
The cluster resource manager is in charge of testing Exchange services running in a cluster. To determine whether a resource is available, the Cluster service uses two methods: a quick check (LooksAlive), then a more detailed check (IsAlive), which verifies that the resource is available. However, the LooksAlive API never calls Exchange objects because Exchange 2000 Server implements a more advanced monitoring technique in which the services notify the cluster resource manager when a resource fails. To determine whether the Information Store (IS) can pass an IsAlive check, the IS service implements a remote procedure call (RPC) that the Cluster service uses. The Service Control Manager (SCM) monitors the System Attendant, Routing Engine, and Message Transfer Agent (MTA) services and tells the Cluster service whether those services are running. And the cluster resource manager tests the HTTP, IMAP4, POP3, and SMTP protocol handlers to determine whether they answer connections on their respective ports.