XFS memory allocation error

[yodaz]

Hi !
I'm using a Netwok Share with only Samba daemon started. The kernel version is the official LaCie kernel 2.6.12.6-arm1. There is only one client connected to the samba daemon, and I have problem with random disconnect.

Here's the NAS log :

Code: Select all

Feb 15 21:06:06 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:08 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:10 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:12 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:13 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:15 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:17 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:19 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:21 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)
Feb 15 21:06:23 (none) user.err kernel: XFS: possible memory allocation deadlock in kmem_alloc (mode:0x2d0)

Here's my smb.conf :

Code: Select all

[global]
    workgroup = WORKGROUP
    server string = LaCie Network Space 
    security = user
    map to guest = bad user
    guest account = nobody
    log file = /var/log/samba/log.%m
    max log size = 50
    log level = 1
    socket options = IPTOS_LOWDELAY TCP_NODELAY
    interfaces = egiga0  
    dns proxy = no
    local master = yes
    os level = 20
    veto files = /lost+found/Network Trash Folder/TheVolumeSettingsFolder/Temporary Items/
    load printers = No
    unix charset = UTF-8
    dos charset = CP437
    use sendfile = no
    use mmap = yes
    Wide links = no
    unix extensions = no
[myshare]
    path = /home/myshare
    browseable = yes
    force user = root
    force group = users
    create mode = 0777
    write list = admin
    valid users = admin,
[openshare]
    path = /home/openshare
    browseable = yes
    force user = root
    force group = users
    create mode = 0777
    guest ok = yes
    read only = no
    public = yes
    writable = yes
    browseable = yes
    create mask = 0666
    directory mask = 0777

I've tried the solution described here : viewtopic.php?f=221&t=1822&p=6124, but I still have the problem.

Do you have any idea how to resolve this issue ?

Thanks in advance.

[yodaz]

I've also noticed that oom killer is still running :

Code: Select all

(none) user.warn kernel: oom-killer: gfp_mask=0x1d2                                        
(none) user.warn kernel: DMA per-cpu:                                                      
(none) user.warn kernel: cpu 0 hot: low 2, high 6, batch 1                                 
(none) user.warn kernel: cpu 0 cold: low 0, high 2, batch 1                                
(none) user.warn kernel: Normal per-cpu: empty                                             
(none) user.warn kernel: HighMem per-cpu: empty                                            
(none) user.warn kernel:                                                                   
(none) user.warn kernel: Free pages:        1200kB (0kB HighMem)                           
(none) user.warn kernel: Active:152 inactive:1194 dirty:0 writeback:0 unstable:0 free:300 slab:1065 mapped:268 pagetables:77
(none) user.warn kernel: DMA free:1200kB min:1200kB low:1500kB high:1800kB active:608kB inactive:4776kB present:16384kB pages_s
(none) user.warn kernel: lowmem_reserve[]: 0 0 0                                           
(none) user.warn kernel: Normal free:0kB min:0kB low:0kB high:0kB active:0kB inactive:0kB present:0kB pages_scanned:0 all_unrec
(none) user.warn kernel: lowmem_reserve[]: 0 0 0                                           
(none) user.warn kernel: HighMem free:0kB min:128kB low:160kB high:192kB active:0kB inactive:0kB present:0kB pages_scanned:0 al
(none) user.warn kernel: lowmem_reserve[]: 0 0 0                                           
(none) user.warn kernel: DMA: 0*4kB 0*8kB 5*16kB 31*32kB 0*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 1200kB
(none) user.warn kernel: Normal: empty                                                     
(none) user.warn kernel: HighMem: empty                                                    
(none) user.warn kernel: Swap cache: add 113875, delete 113750, find 25272/63406, race 0+1 
(none) user.warn kernel: Free swap  = 125248kB                                             
(none) user.warn kernel: Total swap = 128448kB                                             
(none) user.err kernel: Out of Memory: Killed process 1138 (smbd).                         

It seems that NAS doesn't use swap at all. My sysctl settings are :

Code: Select all

vm.swap_token_timeout = 0
vm.vfs_cache_pressure = 100
vm.block_dump = 0
vm.laptop_mode = 0
vm.max_map_count = 65536
vm.min_free_kbytes = 1200
vm.lowmem_reserve_ratio = 256   32
vm.swappiness = 80
vm.nr_pdflush_threads = 2
vm.dirty_expire_centisecs = 3000
vm.dirty_writeback_centisecs = 500
vm.dirty_ratio = 40
vm.dirty_background_ratio = 10
vm.page-cluster = 3
vm.overcommit_ratio = 50
vm.overcommit_memory = 2

Any idea what's wrong ?