2.5.13 error: kswapd: page allocation failure. order:0 mode:0x20

Suddenly, after 36 hours uptime, the console was flooded with this message.

May  6 06:54:24 alpha -- MARK --
May  6 07:11:28 alpha kernel: apd: page allocation failure. order:0, mode:0x20
May  6 07:11:28 alpha kernel: kswapd: page allocation failure. order:0, mode:0x20
May  6 07:11:28 alpha last message repeated 296 times
May  6 07:11:37 alpha kernel: x20
May  6 07:11:37 alpha kernel: ssh: page allocation failure. order:0, mode:0x20
May  6 07:11:37 alpha last message repeated 314 times
May  6 07:11:39 alpha kernel: apd: page allocation failure. order:0, mode:0x20
May  6 07:11:39 alpha kernel: kswapd: page allocation failure. order:0, mode:0x20
May  6 07:11:39 alpha last message repeated 296 times
May  6 07:11:39 alpha kernel: ssh: page allocation failure. order:0, mode:0x20
May  6 07:11:40 alpha last message repeated 73 times
May  6 07:11:46 alpha kernel: Kernel logging (proc) stopped.
May  6 07:11:46 alpha kernel: Kernel log daemon terminating.
May  6 07:11:46 alpha exiting on signal 15
May  6 07:13:35 alpha syslogd 1.4.1#10: restart.

Given the fact that this alpha has 1.5 Gb memory and 2 Gb swap, I feel 'out-of-memory' shouldn't happen (I was running mutt, slrn, some ssh sessions - nothing big).

after reboot:

MemTotal:      1553808 kB
MemFree:       1397688 kB
MemShared:           0 kB
Buffers:         71584 kB
Cached:          23848 kB
SwapCached:          0 kB
Active:          25064 kB
Inactive:        92616 kB
HighTotal:           0 kB
HighFree:            0 kB
LowTotal:      1553808 kB
LowFree:       1397688 kB
SwapTotal:     2292720 kB
SwapFree:      2292720 kB
Dirty:           14760 kB
Writeback:           0 kB

Is this report useful to someone? If more info is needed, just ask me.

Jurriaan

Cadets were humping in the corridors, a party of outraged Senators roamed
Krane barracks, everyone, including me, questioned orders, ...
Law and order.
	David Feintuch - Fisherman's Hope.
Debian GNU/Linux 2.5.13 on Alpha 988 bogomips load:0.00 0.00 0.00

36 hours is probably a world record ;)

It's possible that the machine really did run out of memory. There are some scenarios with memory-mapped holey files and with swap which can result in unfreeable memory. 2.5.14 is better.

It's just a warning. 2.5.14 will calm it down.

-