Hello,
I have a new physical HP server running Server 2008 R2 with Whats Up Gold. Recently, the server started developing an issue roughly every 48 hours where it cannot ping or use ICMP to the internet. The ability to ping to the internet returns roughly two hours later after the start of the symptoms or if I perform a reboot. We know this because we have some stuff we monitor over the internet. It has been in production since November and started this behavior in early January for some reason. Nothing has changed to my knowledge that would cause this issue. It doesn't lose the ability to use ICMP internally just going out to the internet. The local firewall is off and IPv6 is not disabled. I have seen this happen live and have been able to troubleshoot with our network engineers. They are supposedly not seeing any denials of traffic on their firewalls or load balancers.
So far I have:
1. Changed cabling and switch ports with no luck
2. Worked with HP to replaced the NIC card itself and make sure drivers, BIOS, and firmware are updated
3. Rebuilt the TCP/IP stack on the local server
4. Taken the server out of production, changed the network segment and local IP (to confirm the problem actually isn't a network firewall issue), and the server still isn't able to ping to the internet.
The discovery of #4 leads to me believe it is the server itself. Like I said in the above paragraph, the server's ability returns usually within two hours or immediately after a reboot.
I am on the verge of either rebuilding or demanding HP replace the server. The server is fully up to date on patches and drivers. I just really haven't seen behavior like this that didn't involve changing out the NIC. I am not sure if changing the system board would do anything per say. I just want to know what else I could be missing?