Hi,
I have a weird problem. The scenario is as below:
There are two Windows Server 2008 R2 servers in Fail Over Cluster. Three cluster services are running there.
One for Tomcat, one for Apache and another one for MySQL. Anyway, the cluster services are working very well.
I configured one NIC (Broadcom) for Production Network and one NIC (Intel) for Heartbeat Network. No teaming is configured.
Servers are IBM X 3650 M3.
The problem is.. there is email alter setting in application which use the email server to send out the alert email. But when the application try to connect to email server, the connection time out intermittently. Actually email server (exchange 2010 ) and
cluster servers in same VLAN.
So Network team troubleshoot and said that it may be the ARP issue. MAC address table in the router cannot resolve the MAC address of cluster IP.
I think the MAC address of cluster IP is that of physical NIC of active node. But when the cluster services fail over to another node, MAC address will change to new one, right? So I need to give them two MAC addresses for cluster IP. But I am not sure it is
workable or not.
Another weird thing is ..when I tried to ping to email server or even gateway, every first ping says "Destination host unreachalbe" and subsequence ping are successful. and the same thing.. when I tried to telnet to email server port 25, every first attempt
fail saying "cannot open the host on port 25" but subsequence attempt is successful.
One more thin, when I rung the continuous ping to email server and tried to telent to port 25 is successful without any error.
So I'm now very confusing where the problem is coming from.
Really appreciate for any advice from all.
Thanks
Thu Ya
Thura