0 Replies Latest reply on Apr 11, 2014 9:58 AM by u6006

    What will trigger 82599 VF known issue "Rx Page Allocation Errors"

    u6006

      Hi,

       

      I got this logged in my kern.log:

       

      warning kernel: swapper: page allocation failure. order:0, mode:0x20
      ...
      warning kernel: Call Trace:
      warning kernel: [<ffffffff810f9bdc>] ? fallback_alloc+0x16e/0x20b
      warning kernel: <IRQ>  [<ffffffff810f9d2a>] ? ____cache_alloc_node+0xb1/0xf8
      warning kernel: [<ffffffff810ca6c8>] ? __alloc_pages_nodemask+0x6f4/0x780
      warning kernel: [<ffffffff810f9716>] ? kmem_cache_alloc_node+0xdd/0x178
      warning kernel: [<ffffffff812ad4c3>] ? __alloc_skb+0x59/0x159
      warning kernel: [<ffffffff810f94d8>] ? kmem_getpages+0x52/0x11a
      warning kernel: [<ffffffff810f9bdc>] ? fallback_alloc+0x16e/0x20b
      warning kernel: [<ffffffffa022677a>] ? ixgbevf_alloc_rx_buffers+0x5e/0x18e [ixgbevf]
      warning kernel: [<ffffffff810f9d2a>] ? ____cache_alloc_node+0xb1/0xf8
      warning kernel: [<ffffffffa0228dda>] ? ixgbevf_poll+0x4e8/0x601 [ixgbevf]
      warning kernel: [<ffffffff810f9716>] ? kmem_cache_alloc_node+0xdd/0x178
      warning kernel: [<ffffffff812b7dd7>] ? net_rx_action+0xd1/0x29a
      warning kernel: [<ffffffff812ad4c3>] ? __alloc_skb+0x59/0x159
      warning kernel: [<ffffffff8105d694>] ? lock_timer_base+0x26/0x4c
      warning kernel: [<ffffffff8105686c>] ? __do_softirq+0xd9/0x1a6
      warning kernel: [<ffffffffa022677a>] ? ixgbevf_alloc_rx_buffers+0x5e/0x18e [ixgbevf]
      warning kernel: [<ffffffff8100ad6c>] ? call_softirq+0x1c/0x30
      warning kernel: [<ffffffffa0228d7f>] ? ixgbevf_poll+0x48d/0x601 [ixgbevf]
      warning kernel: [<ffffffff8100c635>] ? do_softirq+0x5f/0xbe
      warning kernel: [<ffffffff812b7dd7>] ? net_rx_action+0xd1/0x29a
      warning kernel: [<ffffffff8105686c>] ? __do_softirq+0xd9/0x1a6
      warning kernel: [<ffffffff81056688>] ? irq_exit+0x36/0x78
      warning kernel: [<ffffffff8100ad6c>] ? call_softirq+0x1c/0x30
      warning kernel: [<ffffffff8100c635>] ? do_softirq+0x5f/0xbe
      warning kernel: [<ffffffff81056688>] ? irq_exit+0x36/0x78
      warning kernel: [<ffffffff8100bc83>] ? do_IRQ+0xa0/0xb6
      warning kernel: [<ffffffff8100bc83>] ? do_IRQ+0xa0/0xb6
      warning kernel: [<ffffffff8100a593>] ? ret_from_intr+0x0/0x11
      ...

       

      After a while, the system crashed for out-of-memory (kernel panics). It only happened on some servers (low end Intel Xeon) but not on the others. So I got 3 questions for you guys:

       

      - Is this exactly the VF known issue mentioned in the subject?

      - If yes, what will trigger it?

      - Why the memory was used up when this happened?