[ARM9] OOM with plenty of free swap space?

[ARM9] OOM with plenty of free swap space?

[Clemens Eisserer]

Hi,

I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27
(precompiled wlan driver) kernel as a small buissness server
(postgres, tor, samba, lighttp).

It works quite well, however I recently discovered that postgres was
killed by the oom killer (log below),
although plenty of free swap was available. Its a really small
database so it should easily fit in the 64mb main memory.

Any idea what could the reason for this OOM?

Thank you in advance, Clemens


[17676.783874] oom-killer: gfp_mask=0x201d2, order=0
[17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
(out_of_memory+0x40/0x1d8)
[17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
[<c0072d50>] (__alloc_pages+0x240/0x2c4)
[17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
[<c0075648>] (__do_page_cache_readahead+0x150/0x324)
[17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
[<c0075914>] (do_page_cache_readahead+0x64/0x70)
[17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
[<c006eba0>] (filemap_nopage+0x190/0x3ec)
[17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
C25E0000
[17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from
[<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
[17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
[<c0029364>] (do_page_fault+0xe4/0x214)
[17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
[<c00295e0>] (do_DataAbort+0x3c/0xa4)
[17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
(ret_from_exception+0x0/0x10)
[17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
00000001
[17676.798553]  r4 = FFFFFFFF
[17676.798583] Mem-info:
[17676.798614] DMA per-cpu:
[17676.798675] cpu 0 hot: high 18, batch 3 used:2
[17676.798706] cpu 0 cold: high 6, batch 1 used:0
[17676.798767] DMA32 per-cpu: empty
[17676.798797] Normal per-cpu: empty
[17676.798828] HighMem per-cpu: empty
[17676.798950] Free pages:        1172kB (0kB HighMem)
[17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
unstable:0 free:293 slab:1257 mapped:12129 pagetables:374
[17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
active:22304kB inactive:27260kB present:65536kB pages_scanned:91
all_unreclaimable? no
[17676.799224] lowmem_reserve[]: 0 0 0 0
[17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
[17676.799377] lowmem_reserve[]: 0 0 0 0
[17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
[17676.799530] lowmem_reserve[]: 0 0 0 0
[17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
no
[17676.799682] lowmem_reserve[]: 0 0 0 0
[17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
[17676.799896] DMA32: empty
[17676.799926] Normal: empty
[17676.799957] HighMem: empty
[17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
[17676.800079] Free swap  = 167716kB
[17676.800109] Total swap = 198272kB
[17676.800170] Free swap:       167716kB
[17676.804534] 16384 pages of RAM
[17676.804565] 638 free pages
[17676.804595] 1096 reserved pages
[17676.804626] 1257 slab pages
[17676.804656] 19580 pages shared
[17676.804718] 1091 pages swap cached
[17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
and children.
[17676.805358] Out of memory: Killed process 1537 (postgres).

Re: [ARM9] OOM with plenty of free swap space?

[Clemens Eisserer]

Does nobody have an idea what could be the cause of this OOM situation?

2009/9/5 Clemens Eisserer <linuxhippy@gmail.com>:
> Hi,
>
> I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27
> (precompiled wlan driver) kernel as a small buissness server
> (postgres, tor, samba, lighttp).
>
> It works quite well, however I recently discovered that postgres was
> killed by the oom killer (log below),
> although plenty of free swap was available. Its a really small
> database so it should easily fit in the 64mb main memory.
>
> Any idea what could the reason for this OOM?
>
> Thank you in advance, Clemens
>
>
> [17676.783874] oom-killer: gfp_mask=0x201d2, order=0
> [17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
> (out_of_memory+0x40/0x1d8)
> [17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
> [<c0072d50>] (__alloc_pages+0x240/0x2c4)
> [17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
> [<c0075648>] (__do_page_cache_readahead+0x150/0x324)
> [17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
> [<c0075914>] (do_page_cache_readahead+0x64/0x70)
> [17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
> [<c006eba0>] (filemap_nopage+0x190/0x3ec)
> [17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
> C25E0000
> [17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from
> [<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
> [17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
> [<c0029364>] (do_page_fault+0xe4/0x214)
> [17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
> [<c00295e0>] (do_DataAbort+0x3c/0xa4)
> [17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
> (ret_from_exception+0x0/0x10)
> [17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
> 00000001
> [17676.798553]  r4 = FFFFFFFF
> [17676.798583] Mem-info:
> [17676.798614] DMA per-cpu:
> [17676.798675] cpu 0 hot: high 18, batch 3 used:2
> [17676.798706] cpu 0 cold: high 6, batch 1 used:0
> [17676.798767] DMA32 per-cpu: empty
> [17676.798797] Normal per-cpu: empty
> [17676.798828] HighMem per-cpu: empty
> [17676.798950] Free pages:        1172kB (0kB HighMem)
> [17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
> unstable:0 free:293 slab:1257 mapped:12129 pagetables:374
> [17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
> active:22304kB inactive:27260kB present:65536kB pages_scanned:91
> all_unreclaimable? no
> [17676.799224] lowmem_reserve[]: 0 0 0 0
> [17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> [17676.799377] lowmem_reserve[]: 0 0 0 0
> [17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> [17676.799530] lowmem_reserve[]: 0 0 0 0
> [17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
> no
> [17676.799682] lowmem_reserve[]: 0 0 0 0
> [17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
> [17676.799896] DMA32: empty
> [17676.799926] Normal: empty
> [17676.799957] HighMem: empty
> [17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
> [17676.800079] Free swap  = 167716kB
> [17676.800109] Total swap = 198272kB
> [17676.800170] Free swap:       167716kB
> [17676.804534] 16384 pages of RAM
> [17676.804565] 638 free pages
> [17676.804595] 1096 reserved pages
> [17676.804626] 1257 slab pages
> [17676.804656] 19580 pages shared
> [17676.804718] 1091 pages swap cached
> [17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
> and children.
> [17676.805358] Out of memory: Killed process 1537 (postgres).
>

Re: [ARM9] OOM with plenty of free swap space?

[Enrik Berkhan]

Clemens Eisserer wrote:
> Does nobody have an idea what could be the cause of this OOM situation?

I guess it's too large readahead. I had this situation recently, too, 
with a raid0 of 8 disks (4MB chunks) that set the file readahead count 
to 32MB or so (on a 60MB NOMMU system).

When I tried to read a 100MB file via sendfile(), the kernel insisted on 
doing the 32MB readahead ... (in __do_page_cache_readahead, like in your 
trace).

I solved my problem by switching to dm.

Enrik

> 2009/9/5 Clemens Eisserer <linuxhippy@gmail.com>:
>> Hi,
>>
>> I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27
>> (precompiled wlan driver) kernel as a small buissness server
>> (postgres, tor, samba, lighttp).
>>
>> It works quite well, however I recently discovered that postgres was
>> killed by the oom killer (log below),
>> although plenty of free swap was available. Its a really small
>> database so it should easily fit in the 64mb main memory.
>>
>> Any idea what could the reason for this OOM?
>>
>> Thank you in advance, Clemens
>>
>>
>> [17676.783874] oom-killer: gfp_mask=0x201d2, order=0
>> [17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
>> (out_of_memory+0x40/0x1d8)
>> [17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
>> [<c0072d50>] (__alloc_pages+0x240/0x2c4)
>> [17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
>> [<c0075648>] (__do_page_cache_readahead+0x150/0x324)
>> [17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
>> [<c0075914>] (do_page_cache_readahead+0x64/0x70)
>> [17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
>> [<c006eba0>] (filemap_nopage+0x190/0x3ec)
>> [17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
>> C25E0000
>> [17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from
>> [<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
>> [17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
>> [<c0029364>] (do_page_fault+0xe4/0x214)
>> [17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
>> [<c00295e0>] (do_DataAbort+0x3c/0xa4)
>> [17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
>> (ret_from_exception+0x0/0x10)
>> [17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
>> 00000001
>> [17676.798553]  r4 = FFFFFFFF
>> [17676.798583] Mem-info:
>> [17676.798614] DMA per-cpu:
>> [17676.798675] cpu 0 hot: high 18, batch 3 used:2
>> [17676.798706] cpu 0 cold: high 6, batch 1 used:0
>> [17676.798767] DMA32 per-cpu: empty
>> [17676.798797] Normal per-cpu: empty
>> [17676.798828] HighMem per-cpu: empty
>> [17676.798950] Free pages:        1172kB (0kB HighMem)
>> [17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
>> unstable:0 free:293 slab:1257 mapped:12129 pagetables:374
>> [17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
>> active:22304kB inactive:27260kB present:65536kB pages_scanned:91
>> all_unreclaimable? no
>> [17676.799224] lowmem_reserve[]: 0 0 0 0
>> [17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
>> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
>> [17676.799377] lowmem_reserve[]: 0 0 0 0
>> [17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
>> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
>> [17676.799530] lowmem_reserve[]: 0 0 0 0
>> [17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
>> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
>> no
>> [17676.799682] lowmem_reserve[]: 0 0 0 0
>> [17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
>> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
>> [17676.799896] DMA32: empty
>> [17676.799926] Normal: empty
>> [17676.799957] HighMem: empty
>> [17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
>> [17676.800079] Free swap  = 167716kB
>> [17676.800109] Total swap = 198272kB
>> [17676.800170] Free swap:       167716kB
>> [17676.804534] 16384 pages of RAM
>> [17676.804565] 638 free pages
>> [17676.804595] 1096 reserved pages
>> [17676.804626] 1257 slab pages
>> [17676.804656] 19580 pages shared
>> [17676.804718] 1091 pages swap cached
>> [17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
>> and children.
>> [17676.805358] Out of memory: Killed process 1537 (postgres).

Re: [ARM9] OOM with plenty of free swap space?

[KOSAKI Motohiro]

Hi

> Clemens Eisserer wrote:
> > Does nobody have an idea what could be the cause of this OOM situation?
> 
> I guess it's too large readahead. I had this situation recently, too, 
> with a raid0 of 8 disks (4MB chunks) that set the file readahead count 
> to 32MB or so (on a 60MB NOMMU system).
> 
> When I tried to read a 100MB file via sendfile(), the kernel insisted on 
> doing the 32MB readahead ... (in __do_page_cache_readahead, like in your 
> trace).
> 
> I solved my problem by switching to dm.

IIRC, Wu recently changed readahead code. Wu, could you please give us comment?


> 
> Enrik
> 
> > 2009/9/5 Clemens Eisserer <linuxhippy@gmail.com>:
> >> Hi,
> >>
> >> I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27
> >> (precompiled wlan driver) kernel as a small buissness server
> >> (postgres, tor, samba, lighttp).
> >>
> >> It works quite well, however I recently discovered that postgres was
> >> killed by the oom killer (log below),
> >> although plenty of free swap was available. Its a really small
> >> database so it should easily fit in the 64mb main memory.
> >>
> >> Any idea what could the reason for this OOM?
> >>
> >> Thank you in advance, Clemens
> >>
> >>
> >> [17676.783874] oom-killer: gfp_mask=0x201d2, order=0
> >> [17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
> >> (out_of_memory+0x40/0x1d8)
> >> [17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
> >> [<c0072d50>] (__alloc_pages+0x240/0x2c4)
> >> [17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
> >> [<c0075648>] (__do_page_cache_readahead+0x150/0x324)
> >> [17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
> >> [<c0075914>] (do_page_cache_readahead+0x64/0x70)
> >> [17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
> >> [<c006eba0>] (filemap_nopage+0x190/0x3ec)
> >> [17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
> >> C25E0000
> >> [17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from
> >> [<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
> >> [17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
> >> [<c0029364>] (do_page_fault+0xe4/0x214)
> >> [17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
> >> [<c00295e0>] (do_DataAbort+0x3c/0xa4)
> >> [17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
> >> (ret_from_exception+0x0/0x10)
> >> [17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
> >> 00000001
> >> [17676.798553]  r4 = FFFFFFFF
> >> [17676.798583] Mem-info:
> >> [17676.798614] DMA per-cpu:
> >> [17676.798675] cpu 0 hot: high 18, batch 3 used:2
> >> [17676.798706] cpu 0 cold: high 6, batch 1 used:0
> >> [17676.798767] DMA32 per-cpu: empty
> >> [17676.798797] Normal per-cpu: empty
> >> [17676.798828] HighMem per-cpu: empty
> >> [17676.798950] Free pages:        1172kB (0kB HighMem)
> >> [17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
> >> unstable:0 free:293 slab:1257 mapped:12129 pagetables:374
> >> [17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
> >> active:22304kB inactive:27260kB present:65536kB pages_scanned:91
> >> all_unreclaimable? no
> >> [17676.799224] lowmem_reserve[]: 0 0 0 0
> >> [17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
> >> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> >> [17676.799377] lowmem_reserve[]: 0 0 0 0
> >> [17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
> >> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> >> [17676.799530] lowmem_reserve[]: 0 0 0 0
> >> [17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
> >> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
> >> no
> >> [17676.799682] lowmem_reserve[]: 0 0 0 0
> >> [17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
> >> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
> >> [17676.799896] DMA32: empty
> >> [17676.799926] Normal: empty
> >> [17676.799957] HighMem: empty
> >> [17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
> >> [17676.800079] Free swap  = 167716kB
> >> [17676.800109] Total swap = 198272kB
> >> [17676.800170] Free swap:       167716kB
> >> [17676.804534] 16384 pages of RAM
> >> [17676.804565] 638 free pages
> >> [17676.804595] 1096 reserved pages
> >> [17676.804626] 1257 slab pages
> >> [17676.804656] 19580 pages shared
> >> [17676.804718] 1091 pages swap cached
> >> [17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
> >> and children.
> >> [17676.805358] Out of memory: Killed process 1537 (postgres).
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [ARM9] OOM with plenty of free swap space?

[Wu Fengguang]

Hi Clemens,

On Sat, Sep 05, 2009 at 08:51:24AM -0400, Clemens Eisserer wrote:
> Hi,
> 
> I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27

2.6.16 is a pretty old kernel.

> (precompiled wlan driver) kernel as a small buissness server
> (postgres, tor, samba, lighttp).
> 
> It works quite well, however I recently discovered that postgres was
> killed by the oom killer (log below),
> although plenty of free swap was available. Its a really small
> database so it should easily fit in the 64mb main memory.
> 
> Any idea what could the reason for this OOM?
> 
> Thank you in advance, Clemens
> 
> 
> [17676.783874] oom-killer: gfp_mask=0x201d2, order=0

It's combination of HIGHMEM,WAIT,IO,FS,COLD, which is not easy to
fail.

> [17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
> (out_of_memory+0x40/0x1d8)
> [17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
> [<c0072d50>] (__alloc_pages+0x240/0x2c4)
> [17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
> [<c0075648>] (__do_page_cache_readahead+0x150/0x324)
> [17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
> [<c0075914>] (do_page_cache_readahead+0x64/0x70)
> [17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
> [<c006eba0>] (filemap_nopage+0x190/0x3ec)
> [17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
> C25E0000
> [17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from

This is a mmap read-around, its ra size would be <= 128KB.

> [<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
> [17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
> [<c0029364>] (do_page_fault+0xe4/0x214)
> [17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
> [<c00295e0>] (do_DataAbort+0x3c/0xa4)
> [17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
> (ret_from_exception+0x0/0x10)
> [17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
> 00000001
> [17676.798553]  r4 = FFFFFFFF
> [17676.798583] Mem-info:
> [17676.798614] DMA per-cpu:
> [17676.798675] cpu 0 hot: high 18, batch 3 used:2
> [17676.798706] cpu 0 cold: high 6, batch 1 used:0
> [17676.798767] DMA32 per-cpu: empty
> [17676.798797] Normal per-cpu: empty
> [17676.798828] HighMem per-cpu: empty
> [17676.798950] Free pages:        1172kB (0kB HighMem)
> [17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
> unstable:0 free:293 slab:1257 mapped:12129 pagetables:374

Most LRU pages are mapped:

  active+inactive-writeback=6815+5576-231=12160 ~= 12129=mapped

> [17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
> active:22304kB inactive:27260kB present:65536kB pages_scanned:91
> all_unreclaimable? no
> [17676.799224] lowmem_reserve[]: 0 0 0 0
> [17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> [17676.799377] lowmem_reserve[]: 0 0 0 0
> [17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
> [17676.799530] lowmem_reserve[]: 0 0 0 0
> [17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
> no
> [17676.799682] lowmem_reserve[]: 0 0 0 0
> [17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
> [17676.799896] DMA32: empty
> [17676.799926] Normal: empty
> [17676.799957] HighMem: empty
> [17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
> [17676.800079] Free swap  = 167716kB
> [17676.800109] Total swap = 198272kB
> [17676.800170] Free swap:       167716kB
> [17676.804534] 16384 pages of RAM
> [17676.804565] 638 free pages
> [17676.804595] 1096 reserved pages
> [17676.804626] 1257 slab pages
> [17676.804656] 19580 pages shared
> [17676.804718] 1091 pages swap cached

There are ~30MB pages swapped, and another ~4MB in page cache.

So the system is moderately stressed. It may be a vmscan bug,
I'd recommend to try a recent kernel and see if things improved.

Thanks,
Fengguang

> [17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
> and children.
> [17676.805358] Out of memory: Killed process 1537 (postgres).
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

[RFC] md: don't scale up readahead size if RAID chunk size >= 4MB

[Wu Fengguang]

On Thu, Sep 10, 2009 at 05:11:07PM +0200, Enrik Berkhan wrote:
> Clemens Eisserer wrote:
>> Does nobody have an idea what could be the cause of this OOM situation?
>
> I guess it's too large readahead. I had this situation recently, too,  
> with a raid0 of 8 disks (4MB chunks) that set the file readahead count  
> to 32MB or so (on a 60MB NOMMU system).

The default readahead size would be 2 * 8 * 4MB = 64MB for such an
software RAID. However max_sane_readahead() will limit runtime
readahead size to available_cache / 2, which is ~30MB for your system.

> When I tried to read a 100MB file via sendfile(), the kernel insisted on  
> doing the 32MB readahead ... (in __do_page_cache_readahead, like in your  
> trace).

You could configure readahead size with the blockdev command.

But I admit that the default 64MB ra size is insanely large.
I tended to change this long ago. Though I'm not sure if that
is exactly someone with big arrays wanted. Anyway here is the
patch. Maybe some storage gurus can drop us some hint of use
cases.

Thanks,
Fengguang
---
md: don't scale up readahead size for large RAID chunk size

Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
---
 drivers/md/raid0.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

--- linux.orig/drivers/md/raid0.c	2009-09-11 14:36:02.000000000 +0800
+++ linux/drivers/md/raid0.c	2009-09-11 14:51:13.000000000 +0800
@@ -341,12 +341,16 @@ static int raid0_run(mddev_t *mddev)
 	 * chunk size, then we will not drive that device as hard as it
 	 * wants.  We consider this a configuration error: a larger
 	 * chunksize should be used in that case.
+	 * Also don't touch readahead size if the chunk size is large enough
+	 * (4MB), so that individual devices can already receive good enough
+	 * IO sizes, and otherwise (2 * stripe) would grow too large.
 	 */
 	{
 		int stripe = mddev->raid_disks *
 			(mddev->chunk_sectors << 9) / PAGE_SIZE;
-		if (mddev->queue->backing_dev_info.ra_pages < 2* stripe)
-			mddev->queue->backing_dev_info.ra_pages = 2* stripe;
+		if (mddev->chunk_sectors < (4 * 1024 * 1024 / 512) &&
+			mddev->queue->backing_dev_info.ra_pages < 2 * stripe)
+			mddev->queue->backing_dev_info.ra_pages = 2 * stripe;
 	}
 
 	blk_queue_merge_bvec(mddev->queue, raid0_mergeable_bvec);

Re: [ARM9] OOM with plenty of free swap space?

[Dave Young]

On Fri, Sep 11, 2009 at 11:21 AM, Wu Fengguang <fengguang.wu@intel.com> wrote:
> Hi Clemens,
>
> On Sat, Sep 05, 2009 at 08:51:24AM -0400, Clemens Eisserer wrote:
>> Hi,
>>
>> I am using a Nokia-770 internet tablet (ARM9) running a 2.6.16.27
>
> 2.6.16 is a pretty old kernel.

Vanilla kernel is not usable for Nokia 770, I think he must use the
kernel provided by nokia.

I have not heard of any update for Nokia-770 for a long time because
it is not supported nowadays.

So ...

>
>> (precompiled wlan driver) kernel as a small buissness server
>> (postgres, tor, samba, lighttp).
>>
>> It works quite well, however I recently discovered that postgres was
>> killed by the oom killer (log below),
>> although plenty of free swap was available. Its a really small
>> database so it should easily fit in the 64mb main memory.
>>
>> Any idea what could the reason for this OOM?
>>
>> Thank you in advance, Clemens
>>
>>
>> [17676.783874] oom-killer: gfp_mask=0x201d2, order=0
>
> It's combination of HIGHMEM,WAIT,IO,FS,COLD, which is not easy to
> fail.
>
>> [17676.797241] [<c0026890>] (dump_stack+0x0/0x14) from [<c0071850>]
>> (out_of_memory+0x40/0x1d8)
>> [17676.797393] [<c0071810>] (out_of_memory+0x0/0x1d8) from
>> [<c0072d50>] (__alloc_pages+0x240/0x2c4)
>> [17676.797515] [<c0072b10>] (__alloc_pages+0x0/0x2c4) from
>> [<c0075648>] (__do_page_cache_readahead+0x150/0x324)
>> [17676.797637] [<c00754f8>] (__do_page_cache_readahead+0x0/0x324) from
>> [<c0075914>] (do_page_cache_readahead+0x64/0x70)
>> [17676.797760] [<c00758b0>] (do_page_cache_readahead+0x0/0x70) from
>> [<c006eba0>] (filemap_nopage+0x190/0x3ec)
>> [17676.797943]  r7 = 00000000  r6 = 00219560  r5 = 00000000  r4 =
>> C25E0000
>> [17676.798004] [<c006ea10>] (filemap_nopage+0x0/0x3ec) from
>
> This is a mmap read-around, its ra size would be <= 128KB.
>
>> [<c007cc04>] (__handle_mm_fault+0x2fc/0x96c)
>> [17676.798126] [<c007c908>] (__handle_mm_fault+0x0/0x96c) from
>> [<c0029364>] (do_page_fault+0xe4/0x214)
>> [17676.798248] [<c0029280>] (do_page_fault+0x0/0x214) from
>> [<c00295e0>] (do_DataAbort+0x3c/0xa4)
>> [17676.798339] [<c00295a4>] (do_DataAbort+0x0/0xa4) from [<c0020da8>]
>> (ret_from_exception+0x0/0x10)
>> [17676.798461]  r8 = 00000000  r7 = 40639540  r6 = 40639560  r5 =
>> 00000001
>> [17676.798553]  r4 = FFFFFFFF
>> [17676.798583] Mem-info:
>> [17676.798614] DMA per-cpu:
>> [17676.798675] cpu 0 hot: high 18, batch 3 used:2
>> [17676.798706] cpu 0 cold: high 6, batch 1 used:0
>> [17676.798767] DMA32 per-cpu: empty
>> [17676.798797] Normal per-cpu: empty
>> [17676.798828] HighMem per-cpu: empty
>> [17676.798950] Free pages:        1172kB (0kB HighMem)
>> [17676.799011] Active:5576 inactive:6815 dirty:0 writeback:231
>> unstable:0 free:293 slab:1257 mapped:12129 pagetables:374
>
> Most LRU pages are mapped:
>
>  active+inactive-writeback=6815+5576-231=12160 ~= 12129=mapped
>
>> [17676.799133] DMA free:1172kB min:1024kB low:1280kB high:1536kB
>> active:22304kB inactive:27260kB present:65536kB pages_scanned:91
>> all_unreclaimable? no
>> [17676.799224] lowmem_reserve[]: 0 0 0 0
>> [17676.799285] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
>> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
>> [17676.799377] lowmem_reserve[]: 0 0 0 0
>> [17676.799468] Normal free:0kB min:0kB low:0kB high:0kB active:0kB
>> inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
>> [17676.799530] lowmem_reserve[]: 0 0 0 0
>> [17676.799621] HighMem free:0kB min:128kB low:128kB high:128kB
>> active:0kB inactive:0kB present:0kB pages_scanned:0 all_unreclaimable?
>> no
>> [17676.799682] lowmem_reserve[]: 0 0 0 0
>> [17676.799743] DMA: 33*4kB 4*8kB 1*16kB 1*32kB 1*64kB 1*128kB 1*256kB
>> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 1172kB
>> [17676.799896] DMA32: empty
>> [17676.799926] Normal: empty
>> [17676.799957] HighMem: empty
>> [17676.800018] Swap cache: add 12847, delete 11756, find 42323/43010, race 0+0
>> [17676.800079] Free swap  = 167716kB
>> [17676.800109] Total swap = 198272kB
>> [17676.800170] Free swap:       167716kB
>> [17676.804534] 16384 pages of RAM
>> [17676.804565] 638 free pages
>> [17676.804595] 1096 reserved pages
>> [17676.804626] 1257 slab pages
>> [17676.804656] 19580 pages shared
>> [17676.804718] 1091 pages swap cached
>
> There are ~30MB pages swapped, and another ~4MB in page cache.
>
> So the system is moderately stressed. It may be a vmscan bug,
> I'd recommend to try a recent kernel and see if things improved.
>
> Thanks,
> Fengguang
>
>> [17676.805267] Out of Memory: Kill process 1535 (postgres) score 11478
>> and children.
>> [17676.805358] Out of memory: Killed process 1537 (postgres).
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>



-- 
Regards
dave