having an issue with VMXNET3 virtual network adapter and CentOS 5.7 guest on a Dell R610.
Every day a few hundred kernel errors are logged on eth0 primary NIC, which handles the bulk of network traffic (eth1 and eth2 are on private and LAN subnets and do little traffic relative to eth0 DMZ subnet).
All 3 NICs have VMXNET3 as their adapter type.
A couple of postings on the net suggest simply ignoring the errors, which I did until yesterday, when eth0 actually went down directly after another logged kernel error entry.
The log entry looks like:
localhost kernel: eth0: tq error 0x80000000 Nov 1 08:12:49 localhost kernel: eth0: resetting Nov 1 08:12:50 localhost kernel: eth0: intr type 2, mode 0, 1 vectors allocated Nov 1 08:12:50 localhost kernel: eth0: NIC Link is Up 10000 Mbps
I upgraded VMware tools to latest version and, as per this thread, disabled IPv6:
Neither approach had any effect, kernel errors persist.
1AM local time I'll bring down the VM and try VMXNET2 as a workaround, but given that VMXNET3 apparently provides the latest and greatest, curious to know if anyone has a fix and/or idea as to what is going on.