Re: [Bug #13648] nfsd: page allocation failure

[David Rientjes]

On Tue, 7 Jul 2009, Rafael J. Wysocki wrote:

> This message has been generated automatically as a part of a report
> of regressions introduced between 2.6.29 and 2.6.30.
> 
> The following bug entry is on the current list of known regressions
> introduced between 2.6.29 and 2.6.30.  Please verify if it still should
> be listed and let me know (either way).
> 
> 
> Bug-Entry	: http://bugzilla.kernel.org/show_bug.cgi?id=13648
> Subject		: nfsd: page allocation failure
> Submitter	: Justin Piszcz <jpiszcz@xxxxxxxxxxxxxxx>
> Date		: 2009-06-22 12:08 (15 days old)
> References	: http://lkml.org/lkml/2009/6/22/309
> 

Here's the last page allocation failure in the bug report:

[415964.311165] nfsd: page allocation failure. order:0, mode:0x20
[415964.311168] Pid: 2680, comm: nfsd Not tainted 2.6.30 #2
[415964.311170] Call Trace:
[415964.311171]  <IRQ>  [<ffffffff802849ed>] ? __alloc_pages_internal+0x3dd/0x4e0
[415964.311179]  [<ffffffff802a6c77>] ? cache_alloc_refill+0x2d7/0x570
[415964.311182]  [<ffffffff802a707d>] ? kmem_cache_alloc+0x8d/0xa0
[415964.311185]  [<ffffffff805a6109>] ? __alloc_skb+0x49/0x160
[415964.311188]  [<ffffffff805ea846>] ? tcp_send_ack+0x26/0x120
[415964.311191]  [<ffffffff805e867d>] ? tcp_rcv_established+0x7bd/0x940
[415964.311193]  [<ffffffff805efb1d>] ? tcp_v4_do_rcv+0xdd/0x210
[415964.311195]  [<ffffffff805f02d6>] ? tcp_v4_rcv+0x686/0x750
...
[415964.311319] Active_anon:154810 active_file:131162 inactive_anon:33447
[415964.311320]  inactive_file:690987 unevictable:0 dirty:112116 writeback:0 unstable:0
[415964.311321]  free:8662 slab:965366 mapped:9316 pagetables:4618 bounce:0
[415964.311325] DMA free:9692kB min:16kB low:20kB high:24kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB present:8668kB pages_scanned:0 all_unreclaimable? yes
[415964.311328] lowmem_reserve[]: 0 3246 7980 7980
[415964.311333] DMA32 free:21312kB min:6656kB low:8320kB high:9984kB active_anon:118464kB inactive_anon:23908kB active_file:174708kB inactive_file:1206812kB unevictable:0kB present:3324312kB pages_scanned:0 all_unreclaimable? no
[415964.311336] lowmem_reserve[]: 0 0 4734 4734
[415964.311341] Normal free:3644kB min:9708kB low:12132kB high:14560kB active_anon:500776kB inactive_anon:109880kB active_file:349940kB inactive_file:1557136kB unevictable:0kB present:4848000kB pages_scanned:0 all_unreclaimable? no
[415964.311344] lowmem_reserve[]: 0 0 0 0
...

There's simply no memory available in ZONE_NORMAL to satisfy the atomic 
allocation; you've been able to allocate beyond the 9708K minimum 
watermark because it's GFP_ATOMIC (and then only to 3641K because of 
ALLOC_HIGH and ALLOC_HARDER).

112116 pages, or 438M of memory, is dirty.

[415964.311369] 827035 total pagecache pages
[415964.311371] 4728 pages in swap cache
[415964.311373] Swap cache stats: add 12746, delete 8018, find 16878/17480
[415964.311374] Free swap  = 16756356kB
[415964.311375] Total swap = 16787768kB
[415964.312141] 2277376 pages RAM
[415964.312141] 252254 pages reserved
[415964.312141] 546309 pages shared
[415964.312141] 1520221 pages non-shared

And since you have an 8G machine, that value is only about half of the 
default dirty_background_ratio setting of 10 to start pdflush from writing 
it out.

What is suspect is almost half of your system's memory is consumed by 
slab.  Have you been able to collect slabtop -o when these failures happen 
to determine whether this is a duplicate of 
http://bugzilla.kernel.org/show_bug.cgi?id=13518 ?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/