Unfortunately no one can be told what fun_plug is - you have to see it for yourself.
You are not logged in.
I've been getting a bunch of page allocation failures in my dmesg of late. I've run fs2chk and currently am running cleanboot. Do I need to be worried? How can I prevent them from occuring in the future?
I'm currently running FW 1.05, fun_plug 0.5 and 1 x WD 1GB Green HD.
printk: 46 messages suppressed. swapper: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 printk: 22 messages suppressed. swapper: page allocation failure. order:2, mode:0x20 pdflush: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 printk: 1 messages suppressed. swapper: page allocation failure. order:2, mode:0x20 swapper: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 transmission-da: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20 kswapd0: page allocation failure. order:2, mode:0x20
Last edited by dP21 (2009-02-16 19:54:58)
Offline
Anyone have the same with a solution? Or should I even be concerned?
Offline
Are you using 9K jumbo frames?
For each network packet sent or received, some network drivers tries to allocate memory in powers of 2, thus the message: "page allocation failure. order:2", meaning 2^2=4 pages failed. Each page has 4096 bytes, thus 4*4096=16384 bytes were requested.
When using 9K jumbo frames, the nearest power of two that fits is 16384 bytes, wasting ~7Kbytes (and I suspect that the next power of two, 32768 is really used).
As the dns has only 64MB, under heavy network traffic, BT, e.g., the memory allocator might fail at first, issuing an warning as you see in the logs. No harm. But the memory allocator has to work hard to find available memory, and that extra processing might somehow hurt performance; to further complicate issues, the sender computer might get a timeout and resends the same frame again!
I had that problem and solved it by using 8K bytes jumbo frames. The message never appears again and I noticed no network throughput problem.
Last edited by jcard (2009-02-25 06:19:59)
Offline
Yes, I was using jumbo frames. I have since reduced it to 4K (since my Intel LAN driver only supports 4/9 Jumbo frames).
So far, I haven't noticed anymore failures, but will keep monitoring.
Thanks for the help!
Offline