On two different machines running two different intel cards, I'm seeing a very similar issue. One machine has the Intel 82572EI and the other is 82571EB. After installing irqbalance on both machines, the messages have gone away for _only_ the 82571EB machine (before it was multiple times daily).
Here is the last message from the 82571EB machine: May 23 15:59:17 node01 kernel: [36728.009353] e1000e 0000:04:00.0: eth2: Detected Hardware Unit Hang: May 23 15:59:17 node01 kernel: [36728.009355] TDH <39> May 23 15:59:17 node01 kernel: [36728.009356] TDT <3b> May 23 15:59:17 node01 kernel: [36728.009357] next_to_use <3b> May 23 15:59:17 node01 kernel: [36728.009357] next_to_clean <39> May 23 15:59:17 node01 kernel: [36728.009358] buffer_info[next_to_clean]: May 23 15:59:17 node01 kernel: [36728.009359] time_stamp <1008af3ae> May 23 15:59:17 node01 kernel: [36728.009359] next_to_watch <39> May 23 15:59:17 node01 kernel: [36728.009360] jiffies <1008af4f9> May 23 15:59:17 node01 kernel: [36728.009361] next_to_watch.status <0> May 23 15:59:17 node01 kernel: [36728.009362] MAC Status <80383> May 23 15:59:17 node01 kernel: [36728.009362] PHY Status <792d> May 23 15:59:17 node01 kernel: [36728.009363] PHY 1000BASE-T Status <7800> May 23 15:59:17 node01 kernel: [36728.009364] PHY Extended Status <3000> May 23 15:59:17 node01 kernel: [36728.009365] PCI Status <10> And from the 82572EI machine (still seeing the messages on this machine): May 26 16:09:27 host02 kernel: [2389975.056960] e1000e 0000:07:00.0: eth2: Detected Hardware Unit Hang: May 26 16:09:27 host02 kernel: [2389975.056961] TDH <38> May 26 16:09:27 host02 kernel: [2389975.056962] TDT <3b> May 26 16:09:27 host02 kernel: [2389975.056963] next_to_use <3b> May 26 16:09:27 host02 kernel: [2389975.056963] next_to_clean <38> May 26 16:09:27 host02 kernel: [2389975.056964] buffer_info[next_to_clean]: May 26 16:09:27 host02 kernel: [2389975.056965] time_stamp <1239be208> May 26 16:09:27 host02 kernel: [2389975.056966] next_to_watch <38> May 26 16:09:27 host02 kernel: [2389975.056966] jiffies <1239be30c> May 26 16:09:27 host02 kernel: [2389975.056967] next_to_watch.status <0> May 26 16:09:27 host02 kernel: [2389975.056968] MAC Status <80383> May 26 16:09:27 host02 kernel: [2389975.056968] PHY Status <792d> May 26 16:09:27 host02 kernel: [2389975.056969] PHY 1000BASE-T Status <7800> May 26 16:09:27 host02 kernel: [2389975.056970] PHY Extended Status <3000> May 26 16:09:27 host02 kernel: [2389975.056970] PCI Status <10> Searching around, this may already be fixed elsewhere. See the following reports: https://bugzilla.redhat.com/show_bug.cgi?id=785806 http://www.mail-archive.com/e1000-devel@lists.sourceforge.net/msg04658.html Let me know if you need more specific information related to my systems. -- Matth
