Sunday, February 17, 2019

1 server 2 nics 1 ip goes "down"

Basic scenario here I've been brought in as a consultant to help resolve this problem so far it's a head scratcher. There are server and client VLANS, L3 Switch in the middle serving as router. Switch is an HP-5412zl.

There are multiple servers but the one that is serving as primary DNS is what got them to bring me in. The server has a dual onboard nics which are both in use. One of the IP addresses will inexplicably and intermittently go offline. Going offline here means quit responding to ping and DNS is unavailable as well. The other nic stays online. I've already swapped the switch ports that they are attached to and the problem follows the problem nic. When the issue occours unplugging the nic and plugging it back in immediately fixes the issue.

I was leaning towards this being the one nic in the server or possibly the cable causing the switch to shut down the port. However, I started a trace to those IPs with ping plotter from the client VLAN and after I added a different server I found that it was "going offline" at the exact time as the problem NIC on the DNS server. When reseating the original problem nic the second server still remains offline though the interface with the issue comes back online.

I had to leave due to a sick kiddo shortly after this latest discovery so I'll be heading back to the site this week. I'm wondering about looking into the switch and its MAC table and doing a packet capture in the server VLAN to see if the switch is arping around for the MAC. But this issue is just feels very strange.

EDIT: For spelling and clarity



No comments:

Post a Comment