Re: problem on __alloc_pages

[Marcelo Tosatti <marcelo.tosatti@cyclades.com>]

On Thu, 5 Feb 2004, Evaldo Gardenali wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Marcelo Tosatti wrote:
> |
> | On Thu, 5 Feb 2004, Evaldo Gardenali wrote:
> |
> |
> |>-----BEGIN PGP SIGNED MESSAGE-----
> |>Hash: SHA1
> |>
> |>Hi.
> |>I am a newbie to kernel memory alloc, and got this on my server.
> |>
> |>Feb  5 11:09:39 server1 kernel: __alloc_pages: 1-order allocation failed
> |>(gfp=0x1f0/0)
> |>Feb  5 11:09:39 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0x1f0/0)
> |>Feb  5 11:10:36 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0x1d2/0)
> |>Feb  5 11:11:18 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0x1d2/0)
> |>Feb  5 11:11:18 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0x1d2/0)
> |>Feb  5 11:11:18 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0xf0/0)
> |>Feb  5 11:11:18 server1 kernel: __alloc_pages: 0-order allocation failed
> |>(gfp=0x1d2/0)
> |>Fev  5 11:11:39 server1 gconfd (evaldo-2337): Recebido sinal 15,
> |>desligando corretamente
> |>Fev  5 11:11:40 server1 gconfd (evaldo-2337): Terminando
> |>Feb  5 11:11:52 server1 /usr/sbin/gpm[437]: imps2: Auto-detected
> |
> |
> | Hi Evaldo,
> |
> | Do you have swap space available when this happens?
> |
>
> Absolutely. this box never gets 500M swap used. highest I ever saw is
> 400M on VERY HIGH activity (which happened only once, when we had a task
> force going on here). right now, swap usage is 3M
> I just added 500M more swap to it, just in case (swap partition)
>
> Oh. and I had memtest86 run the 'full test' for 4 days non-stop before
> going production ;)

This is probably happening because you ran out of swap space. Adding more
will help.

Check for application memory leaks.