Hi, Exchange Topology 3 node Active-Active-Passive cluster Two exchange servers 2003 Ent Sp2 Weve seen few occurrences where exchange server loop into persistent failovers and upon research we found that its because of node (instead exchange server) IP in exchange virtual directory(@ inetmgr) that caused this. When we manually select the exchange servers IP we see that resources come online properly. In the event log we see below error all the time when issue happens Event Type: Error Event Source: MSExchangeCluster Event Category: Services Event ID: 1005 Date: 10/31/2009 Time: 8:48:00 PM User: N/A Computer: Node Description: Exchange HTTP Virtual Server Instance 100 (exchange_server): The IsAlive check for this resource failed. Why does IP get change at the first place? Regards, Pushkal MishrA

Never heard of IPs can get changed on its own. Any AV or any other other unwanted activites on the server?Raj

The way in which failover happens IP do get change. Nodes:node1,node2, node3 Exchange servers: exserver1(on node1), exserver2(node2) In the above scenario, exchange virtual server directory (on IIS manager) on node 1 would point to the IP of exserver1 and exchange virtual server directory on node 2 would point to the IP of exserver2. In the event of failover (now exserver1 on node2 and exserver2 on node1), exchange virtual server directory on node 2 would point to exserver1 and exchange virtual server directory on node 1 would point to exserver2. The problem in my case is that SOMETIMES IP of exchange virtual server directory points to node IP and that causes persistent failovers, HTTP timeouts etc. Not sure though but may be incase of any issues(may be resources failed or on passive node) virtual directory might assume the IP of node...........We definitely have an AV(Symantec Endpoint and Group Shield) but I didn't find any unwanted activities...... Regards, Pushkal MishrA

Exactly. The explaination that you gave is normal working of cluster.Raj

Interestingly, The IP that VS dir was pointing belonged to the server's old IP(I hadnt realized that) and it was picking that up from host-file. Once I cleared the entry and performed couple of failovers, it appears fine to me now. I need to monitor this but I am certain that was the cause. Regards, Pushkal MishrA