swapper: page allocation failure. order:1, mode:0x20
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 6 lines · Bernd Schubert
Oh no, not this page allocation problems again. In summer I already posted problems with page allocation errors with 2.6.7, but to me it seemed that nobody cared. That time we got those problems every morning during the cron jobs and our main file server always completely crashed. This time its our cluster master system and first happend after an uptime of 89 days, kernel is 2.6.9. Besides of those messages, the system still seems to run stable
I really beg for help here, so please please please help me solving this probem. What can I do to solve it?
First a (dumb) question, what does 'page allocation failure' really mean? Is it some out of memory case?
Quoted 8 lines
Feb 28 10:04:45 hitchcock kernel: swapper: page allocation failure. order:1, mode:0x20
Feb 28 10:04:45 hitchcock kernel:
Feb 28 10:04:45 hitchcock kernel: Call Trace:<IRQ> <ffffffff8015b0de>{__alloc_pages+878} <ffffffff8015b10e>{__get_free_pages+14}
Feb 28 10:04:45 hitchcock kernel: <ffffffff8015edc6>{kmem_getpages+38} <ffffffff803d064a>{ip_frag_create+26}
Feb 28 10:04:45 hitchcock kernel: <ffffffff8016061e>{cache_grow+190} <ffffffff80160e80>{cache_alloc_refill+560}
Feb 28 10:04:45 hitchcock kernel: <ffffffff801617e3>{__kmalloc+195} <ffffffff803b5680>{alloc_skb+64}
Feb 28 10:04:45 hitchcock kernel: <ffffffff8031727e>{tg3_alloc_rx_skb+222} <ffffffff80317553>{tg3_rx+371}
Feb 28 10:04:45 hitchcock kernel: <ffffffff80317977>{tg3_poll+183} <ffffffff803bc306>{net_rx_action+134}Essentially the tg3 Ethernet driver is trying to allocate memory to store a received packet, and is unable to do so. Since this is done inside interrupt context, this allocation has to be serviced from physical memory. Order 1 means it only wanted one page of memory, and since that failed it looks like the system must have been awfully short on available physical RAM.. it could be some kind of kernel memory leak or VM issue, though this condition may not be entirely unexpected in certain cases, like if the system has little physical RAM free at a certain point and then a flood of network packets arrive.
Signature · 3 lines
Robert Hancock Saskatoon, SK, Canada To email, remove "nospam" from [email protected] Home Page: http://www.roberthancock.com/
Hi.
Quoted 5 lines · Robert Hancock
Bernd Schubert wrote:
Essentially the tg3 Ethernet driver is trying to allocate memory to store a received packet, and is unable to do so. Since this is done inside interrupt context, this allocation has to be serviced from physical memory. Order 1 means it only wanted one page of memory, and
Minor point, I know, but it's 2 pages of memory. If it couldn't get an order zero page, that would be even greater hernia material!
Regards,
Nigel
Quoted 5 lines
since that failed it looks like the system must have been awfully short on available physical RAM.. it could be some kind of kernel memory leak or VM issue, though this condition may not be entirely unexpected in certain cases, like if the system has little physical RAM free at a certain point and then a flood of network packets arrive.
Signature · 6 lines
Nigel Cunningham Software Engineer, Canberra, Australia http://www.cyclades.com Bus: +61 (2) 6291 9554; Hme: +61 (2) 6292 8028; Mob: +61 (417) 100 574 Maintainer of Suspend2 Kernel Patches http://softwaresuspend.berlios.de
Quoted 10 lines · Nigel Cunningham
Hi.
Quoted 2 lines · Robert Hancock
Bernd Schubert wrote:
Essentially the tg3 Ethernet driver is trying to allocate memory to store a received packet, and is unable to do so. Since this is done inside interrupt context, this allocation has to be serviced from physical memory. Order 1 means it only wanted one page of memory, andMinor point, I know, but it's 2 pages of memory. If it couldn't get an order zero page, that would be even greater hernia material!
Indeed.. off-by-one error :-)
Quoted 7 lines · Robert Hancock
Quoted 5 lines · Bernd Schubert
Oh no, not this page allocation problems again. In summer I already posted problems with page allocation errors with 2.6.7, but to me it seemed that nobody cared. That time we got those problems every morning during the cron jobs and our main file server always completely crashed. This time its our cluster master system and first happend after an uptime of 89 days, kernel is 2.6.9. Besides of those messages, the system still seems to run stable
I really beg for help here, so please please please help me solving this probem. What can I do to solve it?
You should upgrade to the newest kernel if possible. If that's not possible, increase /proc/sys/vm/min_free_kbytes
This allocation failure really should not cause your system to crash, but increasing min_free_kbytes will make it less likely that you will see an allocation failure.
Quoted 11 lines
Quoted 9 lines
First a (dumb) question, what does 'page allocation failure' really mean? Is it some out of memory case?
Feb 28 10:04:45 hitchcock kernel: swapper: page allocation failure. order:1, mode:0x20 Feb 28 10:04:45 hitchcock kernel: Feb 28 10:04:45 hitchcock kernel: Call Trace:<IRQ> <ffffffff8015b0de>{__alloc_pages+878} <ffffffff8015b10e>{__get_free_pages+14} Feb 28 10:04:45 hitchcock kernel: <ffffffff8015edc6>{kmem_getpages+38} <ffffffff803d064a>{ip_frag_create+26} Feb 28 10:04:45 hitchcock kernel: <ffffffff8016061e>{cache_grow+190} <ffffffff80160e80>{cache_alloc_refill+560} Feb 28 10:04:45 hitchcock kernel: <ffffffff801617e3>{__kmalloc+195} <ffffffff803b5680>{alloc_skb+64} Feb 28 10:04:45 hitchcock kernel: <ffffffff8031727e>{tg3_alloc_rx_skb+222} <ffffffff80317553>{tg3_rx+371} Feb 28 10:04:45 hitchcock kernel: <ffffffff80317977>{tg3_poll+183} <ffffffff803bc306>{net_rx_action+134}
Quoted 2 lines
Essentially the tg3 Ethernet driver is trying to allocate memory to store a received packet, and is unable to do so. Since this is done inside interrupt context, this allocation has to be serviced from physical memory. Order 1 means it only wanted one page of memory, and since that failed it looks like the system must have been awfully short on available physical RAM.. it could be some kind of kernel memory leak or VM issue, though this condition may not be entirely unexpected in certain cases, like if the system has little physical RAM free at a certain point and then a flood of network packets arrive.
Yep. The reason why these failures are beeing seen is that earlier kernels did not reserve enough memory for GFP_ATOMIC allocations. Later kernels increased this, and also made higher order (ie. greater than 0) GFP_ATOMIC allocations more robust.