Re: Auto Checking Raid 6 crashes my system

[Dawning Sky <the.dawning.sky@gmail.com>]

I tried to reproduce the crash.  First I echoed "check" into
/sys/block/md127/md/sync_action under the single mode and the array
finished checking.  Then I booted into init 3 and repeated the same
thing and got the following error on the screen and the computer hung.
 I had to manually type the error since I had no way to copy/paste, so
there might be a few typos.


UG: unable to handle kernel paging request at 000000006a312c50
IP: [<ffffffff810402a8>] task_rq_lock_0x3c/0x7e
PGD 0
Oops: 0000 [#1] SMP
last sysfs file: /sys/devices/virtual/block/md127/md/sync_speed
CPU 1
Modules linked: fuse ipt_MASQUERADE iptable_nat nf_nat bridge stp llc
nfsd lockd nfs_acl auth_rpcgss exportfs autofs4 hwmon_vid sunrpc
cpufreq_ondemand powernow_k8 freq_table ipv6 nf_conntrack_netbios_ns
ext2 kvm_amd kvm uinput snd_hda_codec_analog usblp snd_hda_intel
snd_hda_codec snd_hwdep snd_seq snd_seq_device snd_pcm snd_timer snd
forcedeth amd64_edac_mod ppdev soundcore edac_core i2c_nforce2
snd_page_alloc parport_pc k8temp serio_raw parport asusatk0110 raid456
raid6_pq async_xor async_tx xor dm_multipath ata_generic pata_jmicron
firewire_ohci firewire_core crc_itu_t pata_amd pata_acpi sata_nv
usb_storage nouveau ttm drm_kms_helper drm i2c_algo_bit i2c_core [last
unloaded: scsi_wait_scan]
Pid: 523, comm: md127_raid5 Not tainted 2.6.31.12-174.2.3.fc12.x86_64
#1 System Product Name
RIP: 0010:[<ffffffff810402a8>] [<ffffffff810402a8>] task_rq_lock+0x3c/0x7e
RSP: 0010:ffff8801155d5c10 EFLAGS: 00010046
RAX: 000000001d1836b0 RBX: 0000000000015600 RCX: 0000000000000000
RDX: 0000000000000046 RSI: ffff8801155d5c58 RDI: ffff8800dc4c2f00
RBP: ffff8801155d5c30 R08: 0000000000000000 R09: 0000000000000001
R10: ffff880115cb64e8 R11: 6db6db6db6db6db7 R12: ffff8801155d5c58
R13: ffff8800dc4c2f00 R14: 0000000000015600 R15: 0000000000000000
FS: 00007f66cc00c780(0000) GS: ffff88002803c000(0000) knlGS: 0000000000000000
CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 000000006a312c50 CR3: 0000000001001000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process md127_raid5 (pid: 523, threadinfo ffff8801155d4000, task
ffff880115daaf00)
Stack:
 ffff8800dc4c2f00 0000000000000000 0000000000000003 ffff8801155d5c58
<0> ffff8801155d5c90 ffffffff8104ae7d
[drm] nouveau 0000:07:00.0: GPU lockup - switching to software fbcon
 ffff880115fc4c00 ffff880115a45400
<0> ffff88011809e400 00000000000000046 ffff888115cb64e8 0000000000000001
Call Trace:
 [<ffffffff8104ae7d>] try_to_wake_up+0x9a/0x2de
 [<ffffffff8104b0d3>] default_wake_function+0x12/0x14
 [<ffffffff8103c237>] __wake_up_common+0x4e/0x84
 [<ffffffff810400cd>] __wake_up+0x39/0x4d
 [<ffffffffa017ce72>] __release_stripe+0x115/0x147 [raid456]
 [<ffffffffa017ced9>] release_stripe+0x35/0x49 [raid456]
 [<ffffffffa0182cd8>] raid5d+0x44e/0x563 [raid456]
 [<ffffffff8141c285>] ? schedule_timeout+0xb3/0xe3
 [<ffffffff8105c236>] ? process_timeout+0x0/0x10
 [<ffffffff8133deef>] md_thread+0xf1/0x10f
 [<ffffffff81067b37>] ? autoremove_wake_function+0x0/0x39
 [<ffffffff8133ddfe>] ? md_thread+0x0/0x10f
 [<ffffffff810677b5>] kthread+0x91/0x99
 [<ffffffff81022daa>] child_rip+0xa/0x20
 [<ffffffff81067724>] ? kthread+0x0/0x99
 [<ffffffff81012da0>] ? child_rip+0x0/0x20
Code: c7 c3 00 56 01 00 49 89 fd 49 89 f4 9c 58 66 66 90 66 90 48 89
c2 fa 66 66 90 66 66 90 49 89 14 24 49 8b 45 08 49 89 de 8b 40 18 <4c>
03 34 c5 d0 76 6f 81 4c 89 f7 e8 ac d3 3d 00 49 8b 45 08 8b
RIP [<ffffffff810402a8>] task_rq_lock+0x3c/0x7e
 RSP <ffff8801155d5c10>
CR2: 000000006a312c50
---[ end trace 6c5abd1701cc36a0 ]---
BUG: unable to handle kernel


On Tue, Feb 16, 2010 at 9:16 AM, Dawning Sky <the.dawning.sky@gmail.com> wrote:
> On Tue, Feb 16, 2010 at 5:37 AM, Kristleifur Daðason
> <kristleifur@gmail.com> wrote:
>> On Tue, Feb 16, 2010 at 5:04 AM, Dawning Sky <the.dawning.sky@gmail.com> wrote:
>>> Hi,
>>>
>>> I just build a brand new md raid-6 with 5 disks.  And on Fedora 12,
>>> the auto checking of md devices via a weekly cron job is enabled by
>>> default.  It performs the checking by echo "check" into
>>> /sys/block/mdX/md/sync_action.  But after a while, the kernel just
>>> crashes, without finishing checking the raid device or leaving
>>> anything in the messages file.
>>>
>>> For now, I've disable the raid-check cron job.  The kernel version I'm
>>> running is 2.6.31.12-174.2.3.fc12.x86_64.
>>>
>>> Thanks,
>>>
>>> DS
>>
>>
>> Hi,
>>
>> I encountered similar issues because of a faulty mvsas driver that
>> flaked out under the load. What hardware do you have?
>
> I'm using the onboard sata chip, which I believe is nVidia, and the
> driver is nv_sata.
>
>>
>> -- Kristleifur
>>
>
> DS
>
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html