Skip to contentSkip to conversation

Oops: dm_mirror on 2.6.14

DISCUSSION Dieter Stüken 2 participants 2 messages Nov 12, 2005 – Jan 10, 2006 Device mapper

lore
Share & tools
lore
#0 lore raw

I run into trouble while moving some LVs. Looks very much the same as http://www.redhat.com/archives/dm-devel/2005-June/msg00162.html
to me. The first Oops arise accessing some data during pvmove, as I unfortunately did not unmount the LVs. Meanwhile I recognized my kernel (2.6.12.2) still suffered from the bio_clone bug
http://www.redhat.com/archives/dm-devel/2005-July/msg00086.html,
but I can't tell if this really caused the problem. After all, I get an Oops now, whenever I try to activate my lvm, as dm-mirror is still active because the pvmove did not finish :-(

The interesting thing is, that even my new 2.6.14 kernel fails to activate the mirrored volume. I may either abort the pvmove or restore the VG to its previous state, but then the opportunity to analyze and fix the problem may be lost.

Here comes the Oops of my current 2.6.14 kernel (x86/PIII):

Unable to handle kernel paging request at virtual address d0a47000
 printing eip:
d0a339a7
*pde = 0127b067
Oops: 0000 [#1]
Modules linked in: dm_mirror dm_mod sata_promise libata pdc202xx_old via82cxxx ...
CPU:    0
EIP:    0060:[<d0a339a7>]    Not tainted VLI
EFLAGS: 00010246   (2.6.14-default-gf2c84c0e)

EIP is at core_in_sync+0x7/0x20 [dm_mirror]
eax: c958bdc0 ebx: d0a41000 ecx: 00000000 edx: 00030000
esi: cffc2900 edi: c97657e0 ebp: 00000000 esp: c84fda3c
ds: 007b es: 007b ss: 0068
Process udev_volume_id (pid: 12044, threadinfo=c84fc000 task=c554e5c0)

Stack: d0a38380 d0a35651 00000000 0000000a 00030000 00000000 c01608a9 c6a3f740
       c958be40 c6a3f738 cffc2900 d0a310b8 d087837c 0c000000 00000000 cffc2180
       c6a3f748 c84fdae8 d0878618 00000000 00000001 00000008 00000000 00000000
Call Trace:
 [<d0a35651>] mirror_map+0xa1/0x150 [dm_mirror]
 [<c01608a9>] bio_clone+0xc9/0xd0
 [<d087837c>] __map_bio+0x3c/0x100 [dm_mod]
 [<d0878618>] __clone_and_map+0xc8/0x320 [dm_mod]
 [<c0140f1a>] mempool_alloc+0x2a/0xc0
 [<d0878917>] __split_bio+0xa7/0x120 [dm_mod]
 [<d08789f6>] dm_request+0x66/0x90 [dm_mod]
 [<c025772b>] generic_make_request+0xeb/0x170
 [<c01608a9>] bio_clone+0xc9/0xd0
 [<d087837c>] __map_bio+0x3c/0x100 [dm_mod]
 [<d0878618>] __clone_and_map+0xc8/0x320 [dm_mod]
 [<c0140f1a>] mempool_alloc+0x2a/0xc0
 [<d0878917>] __split_bio+0xa7/0x120 [dm_mod]
 [<d08789f6>] dm_request+0x66/0x90 [dm_mod]
 [<c025772b>] generic_make_request+0xeb/0x170
 [<c0257800>] submit_bio+0x50/0xe0
 [<c01605e9>] bio_alloc_bioset+0xe9/0x1e0
 [<c015ff6d>] submit_bh+0x13d/0x190
 [<c015ef2a>] block_read_full_page+0x30a/0x320
 [<c0162e90>] blkdev_get_block+0x0/0x80
 [<c0144aec>] read_pages+0x4c/0x100
 [<c014247c>] __alloc_pages+0x17c/0x410
 [<c0144c44>] __do_page_cache_readahead+0xa4/0x100
 [<c0144df1>] blockable_page_cache_readahead+0x51/0xd0
 [<c0145031>] page_cache_readahead+0x101/0x1a0
 [<c013e516>] do_generic_mapping_read+0x326/0x490
 [<c013e912>] __generic_file_aio_read+0x1a2/0x210
 [<c013e680>] file_read_actor+0x0/0xf0
 [<c013ea85>] generic_file_read+0x95/0xc0
 [<c012ea20>] autoremove_wake_function+0x0/0x50
 [<c015b947>] vfs_read+0xd7/0x180
 [<c015bcd1>] sys_read+0x41/0x70
 [<c0102f29>] syscall_call+0x7/0xb
Signature · 4 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]
    http://www.conterra.de/
    (0)251-7474-501
#1 lore raw
Quoted 10 lines · Dieter Stüken

I run into trouble while moving some LVs. Looks very much the same as http://www.redhat.com/archives/dm-devel/2005-June/msg00162.html
to me. The first Oops arise accessing some data during pvmove, as I unfortunately did not unmount the LVs. Meanwhile I recognized my kernel (2.6.12.2) still suffered from the bio_clone bug
http://www.redhat.com/archives/dm-devel/2005-July/msg00086.html,
but I can't tell if this really caused the problem. After all, I get an Oops now, whenever I try to activate my lvm, as dm-mirror is still active because the pvmove did not finish :-(

Quoted 6 lines

The interesting thing is, that even my new 2.6.14 kernel fails to activate the mirrored volume. I may either abort the pvmove or restore the VG to its previous state, but then the opportunity to analyze and fix the problem may be lost.

Here comes the Oops of my current 2.6.14 kernel (x86/PIII):

Is this problem still present in kernel 2.6.15?

Quoted 47 lines
Unable to handle kernel paging request at virtual address d0a47000
 printing eip:
d0a339a7
*pde = 0127b067
Oops: 0000 [#1]
Modules linked in: dm_mirror dm_mod sata_promise libata pdc202xx_old 
via82cxxx ...
CPU:    0
EIP:    0060:[<d0a339a7>]    Not tainted VLI
EFLAGS: 00010246   (2.6.14-default-gf2c84c0e)

EIP is at core_in_sync+0x7/0x20 [dm_mirror]
eax: c958bdc0 ebx: d0a41000 ecx: 00000000 edx: 00030000
esi: cffc2900 edi: c97657e0 ebp: 00000000 esp: c84fda3c
ds: 007b es: 007b ss: 0068
Process udev_volume_id (pid: 12044, threadinfo=c84fc000 task=c554e5c0)

Stack: d0a38380 d0a35651 00000000 0000000a 00030000 00000000 c01608a9 
c6a3f740
       c958be40 c6a3f738 cffc2900 d0a310b8 d087837c 0c000000 00000000 
       cffc2180
       c6a3f748 c84fdae8 d0878618 00000000 00000001 00000008 00000000 
       00000000
Call Trace:
 [<d0a35651>] mirror_map+0xa1/0x150 [dm_mirror]
 [<c01608a9>] bio_clone+0xc9/0xd0
 [<d087837c>] __map_bio+0x3c/0x100 [dm_mod]
 [<d0878618>] __clone_and_map+0xc8/0x320 [dm_mod]
 [<c0140f1a>] mempool_alloc+0x2a/0xc0
 [<d0878917>] __split_bio+0xa7/0x120 [dm_mod]
 [<d08789f6>] dm_request+0x66/0x90 [dm_mod]
 [<c025772b>] generic_make_request+0xeb/0x170
 [<c01608a9>] bio_clone+0xc9/0xd0
 [<d087837c>] __map_bio+0x3c/0x100 [dm_mod]
 [<d0878618>] __clone_and_map+0xc8/0x320 [dm_mod]
 [<c0140f1a>] mempool_alloc+0x2a/0xc0
 [<d0878917>] __split_bio+0xa7/0x120 [dm_mod]
 [<d08789f6>] dm_request+0x66/0x90 [dm_mod]
 [<c025772b>] generic_make_request+0xeb/0x170
 [<c0257800>] submit_bio+0x50/0xe0
 [<c01605e9>] bio_alloc_bioset+0xe9/0x1e0
 [<c015ff6d>] submit_bh+0x13d/0x190
 [<c015ef2a>] block_read_full_page+0x30a/0x320
 [<c0162e90>] blkdev_get_block+0x0/0x80
 [<c0144aec>] read_pages+0x4c/0x100
 [<c014247c>] __alloc_pages+0x17c/0x410
 [<c0144c44>] __do_page_cache_readahead+0xa4/0x100
 [<c0144df1>] blockable_page_cache_readahead+0x51/0xd0
 [<c0145031>] page_cache_readahead+0x101/0x1a0
Quoted 21 lines
 [<c013e516>] do_generic_mapping_read+0x326/0x490
 [<c013e912>] __generic_file_aio_read+0x1a2/0x210
 [<c013e680>] file_read_actor+0x0/0xf0
 [<c013ea85>] generic_file_read+0x95/0xc0
 [<c012ea20>] autoremove_wake_function+0x0/0x50
 [<c015b947>] vfs_read+0xd7/0x180
 [<c015bcd1>] sys_read+0x41/0x70
 [<c0102f29>] syscall_call+0x7/0xb
Signature · 9 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]
    http://www.conterra.de/
    (0)251-7474-501
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed