[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#603508: kernel crashes during raid6 resync



On Sun, 2010-11-14 at 21:06 +0100, Andreas Florath wrote:
> Package: linux-2.6
> Version: 2.6.32-27
> File: /boot/vmlinuz-2.6.32-5-amd64
> Severity: normal
> 
> *** Please type your report below this line ***
> After installation of four new disks and configuring RAID 6 using
> those four disks, the kernel keeps crashing every some minutes.
> When using programs which exercise the disk (like bonnie++) it only
> takes seconds to get an Oops.
[...]
> Nov 14 18:50:10 peleus kernel: [  989.048138] JBD2: Detected IO errors while flushing file data on dm-0-8
> Nov 14 18:50:14 peleus kernel: [  992.624892] BUG: unable to handle kernel NULL pointer dereference at 0000000000000008
> Nov 14 18:50:14 peleus kernel: [  992.625081] IP: [<ffffffff810b5799>] mempool_free+0x14/0x7e
> Nov 14 18:50:14 peleus kernel: [  992.625238] PGD 7a5be067 PUD 77886067 PMD 0
> Nov 14 18:50:14 peleus kernel: [  992.625458] Oops: 0000 [#1] SMP
> Nov 14 18:50:14 peleus kernel: [  992.625613] last sysfs file: /sys/devices/virtual/block/dm-0/dm/name
> Nov 14 18:50:14 peleus kernel: [  992.625684] CPU 1
> Nov 14 18:50:14 peleus kernel: [  992.625788] Modules linked in: ext4 jbd2 crc16 dm_mod loop snd_pcm snd_timer i2c_i801 i2c_core snd soundcore snd_page_alloc pcspkr evdev joydev button processor ext3 jbd mbcache raid10 raid456 async_raid6_recov async_pq raid6_pq async_xor xor async_memcpy async_tx raid1 raid0 multipath linear md_mod sd_mod crc_t10dif usbhid hid usb_storage ahci libata uhci_hcd thermal scsi_mod e1000e ehci_hcd thermal_sys usbcore nls_base [last unloaded: scsi_wait_scan]
> Nov 14 18:50:14 peleus kernel: [  992.628146] Pid: 335, comm: md127_raid6 Tainted: G    B      2.6.32-5-amd64 #1 X7SPA-HF
[...]

The 'B' taint flag above indicates that the kernel already ran into a
serious error.  We need to see the earlier error message beginning with
'BUG: Bad page ...'

Also, please consider testing the system RAM using memtest86+.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: