Re: [linux-lvm] multiple snapshots

[Imre Gergely <imre.gergely@astral.ro>]

anybody anything? i still have this problem... the main question is: can
one have multiple snapshots of the same LV, without problems? with one
snapshot e ok, but as soon as i create another snapshot, and try to do
something on the origin LV, it hangs, it oops-es.

Imre Gergely wrote:
> i've got a log entry for one of the oops-es:
> 
> Unable to handle kernel paging request at virtual address 00100108
>  printing eip:
> c025e142
> *pde = 00000000
> Oops: 0000 [#1]
> PREEMPT
> Modules linked in: pcnet32
> CPU:    0
> EIP:    0060:[<c025e142>]    Not tainted VLI
> EFLAGS: 00010213   (2.6.12.5)
> EIP is at __origin_write+0x62/0x230
> eax: 001000e0   ebx: c59a9c60   ecx: c7c915c0   edx: ffff0001
> esi: 001000e0   edi: c568a7cc   ebp: 00000000   esp: c7189cd0
> ds: 007b   es: 007b   ss: 0068
> Process kjournald (pid: 991, threadinfo=c7188000 task=c7c915c0)
> Stack: c70b6580 c7189d04 00000001 c7189cf4 00000000 c025dbc0 c7c5fe3c
> c48b4cec
>        00000000 c6285460 00001340 00000010 c61b6b40 c5f0c220 00002be0
> 00000010
>        c1bca9e0 c037ae7c 00000001 c32e9aec c7189da4 c025e357 c568a7cc
> c1bca560
> Call Trace:
>  [<c025dbc0>] copy_callback+0x0/0x50
>  [<c025e357>] do_origin+0x47/0x70
>  [<c0251449>] __map_bio+0x49/0x120
>  [<c02518c1>] __clone_and_map+0x2a1/0x2b0
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c0251978>] __split_bio+0xa8/0x130
>  [<c0251a6e>] dm_request+0x6e/0xa0
>  [<c0205c67>] generic_make_request+0x147/0x1e0
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c0205d62>] submit_bio+0x62/0x100
>  [<c0153c44>] bio_alloc_bioset+0xe4/0x1c0
>  [<c0153d40>] bio_alloc+0x20/0x30
>  [<c0153572>] submit_bh+0xd2/0x120
>  [<c01c2371>] journal_commit_transaction+0xd51/0x1220
>  [<c01048be>] do_IRQ+0x1e/0x30
>  [<c02b51c7>] schedule+0x347/0x5e0
>  [<c01c4ab6>] kjournald+0xd6/0x250
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c0127470>] autoremove_wake_function+0x0/0x60
>  [<c01029ae>] ret_from_fork+0x6/0x14
>  [<c01c49c0>] commit_timeout+0x0/0x10
>  [<c01c49e0>] kjournald+0x0/0x250
>  [<c0100c31>] kernel_thread_helper+0x5/0x14
> Code: 85 d2 0f 85 46 01 00 00 8b 53 18 8d 42 e8 89 c3 8b 40 18 0f 18 00
> 90 39 fa 75 e2 8b 44 24 1c 85 c0 0f 84 fc 00 00 00 8b 74 24 1c <8b> 46
> 28 ba 01 00 ff ff 0f c1 10 85 d2 0f 85 2e 05 00 00 85 ed
> 
> at this point the system still works but i cannot do reading/writing on
> that mounted partition. looks like ext3's journal has some problems with
> the snapshots?
> 
> i did a restart, i removed, then recreated the two snapshots, did some
> reading/writing, then another oops popped up:
> 
> Aug 16 10:32:21 fc2 kernel: Unable to handle kernel NULL pointer
> dereference at
> virtual address 00000000
> Aug 16 10:32:21 fc2 kernel:  printing eip:
> Aug 16 10:32:21 fc2 kernel: c025e196
> Aug 16 10:32:21 fc2 kernel: *pde = 00000000
> Aug 16 10:32:21 fc2 kernel: Oops: 0002 [#1]
> Aug 16 10:32:21 fc2 kernel: PREEMPT
> Aug 16 10:32:21 fc2 kernel: Modules linked in: pcnet32
> Aug 16 10:32:21 fc2 kernel: CPU:    0
> Aug 16 10:32:21 fc2 kernel: EIP:    0060:[<c025e196>]    Not tainted VLI
> Aug 16 10:32:21 fc2 kernel: EFLAGS: 00010246   (2.6.12.5)
> Aug 16 10:32:21 fc2 kernel: EIP is at __origin_write+0xb6/0x230
> Aug 16 10:32:21 fc2 kernel: eax: 00000000   ebx: c583c0e0   ecx:
> c7fc2560   edx:
>  0000ffff
> Aug 16 10:32:21 fc2 kernel: esi: c57a9b0c   edi: c507498c   ebp:
> 00000000   esp:
>  c7fe5b14
> Aug 16 10:32:21 fc2 kernel: ds: 007b   es: 007b   ss: 0068
> Aug 16 10:32:21 fc2 kernel: Process pdflush (pid: 8, threadinfo=c7fe4000
> task=c7
> fc2560)
> Aug 16 10:32:21 fc2 kernel: Stack: c4c796a0 c7fe5b48 00000001 c7fe5b38
> 00000000
> c025dbc0 c57a9b0c c7c5f1dc
> Aug 16 10:32:21 fc2 kernel:        00000000 c5332dc0 000009e0 00000010
> c50ed330
> c7ee0520 000044e0 00000010
> Aug 16 10:32:21 fc2 kernel:        c73e8620 c037ae7c 00000001 c51a4dbc
> c7fe5be8
> c025e357 c507498c c73e85c0
> Aug 16 10:32:21 fc2 kernel: Call Trace:
> Aug 16 10:32:22 fc2 kernel:  [<c025dbc0>] copy_callback+0x0/0x50
> Aug 16 10:32:22 fc2 kernel:  [<c025e357>] do_origin+0x47/0x70
> Aug 16 10:32:22 fc2 kernel:  [<c0251449>] __map_bio+0x49/0x120
> Aug 16 10:32:22 fc2 kernel:  [<c02518c1>] __clone_and_map+0x2a1/0x2b0
> Aug 16 10:32:22 fc2 kernel:  [<c0127470>] autoremove_wake_function+0x0/0x60
> Aug 16 10:32:22 fc2 kernel:  [<c0251978>] __split_bio+0xa8/0x130
> Aug 16 10:32:22 fc2 kernel:  [<c0251a6e>] dm_request+0x6e/0xa0
> Aug 16 10:32:22 fc2 kernel:  [<c0205c67>] generic_make_request+0x147/0x1e0
> Aug 16 10:32:22 fc2 kernel:  [<c0127470>] autoremove_wake_function+0x0/0x60
> Aug 16 10:32:22 fc2 last message repeated 2 times
> Aug 16 10:32:22 fc2 kernel:  [<c0205d62>] submit_bio+0x62/0x100
> Aug 16 10:32:22 fc2 kernel:  [<c0153c44>] bio_alloc_bioset+0xe4/0x1c0
> Aug 16 10:32:22 fc2 kernel:  [<c0153d40>] bio_alloc+0x20/0x30
> Aug 16 10:32:22 fc2 kernel:  [<c0153572>] submit_bh+0xd2/0x120
> Aug 16 10:32:22 fc2 kernel:  [<c0151c90>]
> __block_write_full_page+0x150/0x320
> Aug 16 10:32:22 fc2 kernel:  [<c01533dd>] block_write_full_page+0xcd/0x100
> Aug 16 10:32:22 fc2 kernel:  [<c01b2620>] ext3_get_block+0x0/0xa0
> Aug 16 10:32:22 fc2 kernel:  [<c01b3292>] ext3_ordered_writepage+0xd2/0x1c0
> Aug 16 10:32:22 fc2 kernel:  [<c01b2620>] ext3_get_block+0x0/0xa0
> Aug 16 10:32:22 fc2 kernel:  [<c01b3180>] bget_one+0x0/0x10
> Aug 16 10:32:22 fc2 kernel:  [<c0173842>] mpage_writepages+0x262/0x3e0
> Aug 16 10:32:22 fc2 kernel:  [<c01b31c0>] ext3_ordered_writepage+0x0/0x1c0
> Aug 16 10:32:22 fc2 kernel:  [<c013739d>] do_writepages+0x3d/0x50
> Aug 16 10:32:22 fc2 kernel:  [<c0171d61>] __sync_single_inode+0x71/0x210
> Aug 16 10:32:22 fc2 kernel:  [<c0171f67>]
> __writeback_single_inode+0x67/0x150
> Aug 16 10:32:22 fc2 kernel:  [<c0251b70>] dm_any_congested+0x30/0x60
> Aug 16 10:32:22 fc2 kernel:  [<c0253e8d>] dm_table_any_congested+0x5d/0x60
> Aug 16 10:32:22 fc2 kernel:  [<c0251b70>] dm_any_congested+0x30/0x60
> Aug 16 10:32:22 fc2 kernel:  [<c01721e7>] sync_sb_inodes+0x197/0x2a0
> Aug 16 10:32:22 fc2 kernel:  [<c01723c4>] writeback_inodes+0xd4/0xf0
> Aug 16 10:32:22 fc2 kernel:  [<c0137053>] background_writeout+0x73/0xc0
> Aug 16 10:32:22 fc2 kernel:  [<c0137afb>] __pdflush+0xbb/0x1a0
> Aug 16 10:32:22 fc2 kernel:  [<c0137be0>] pdflush+0x0/0x30
> Aug 16 10:32:22 fc2 kernel:  [<c0137c08>] pdflush+0x28/0x30
> Aug 16 10:32:22 fc2 kernel:  [<c0136fe0>] background_writeout+0x0/0xc0
> Aug 16 10:32:22 fc2 kernel:  [<c0137be0>] pdflush+0x0/0x30
> Aug 16 10:32:22 fc2 kernel:  [<c0126fa5>] kthread+0xa5/0xb0
> Aug 16 10:32:22 fc2 kernel:  [<c0126f00>] kthread+0x0/0xb0
> Aug 16 10:32:22 fc2 kernel:  [<c0100c31>] kernel_thread_helper+0x5/0x14
> Aug 16 10:32:22 fc2 kernel: Code: c0 0f 84 e7 00 00 00 89 48 04 8b 4c 24
> 5c 89 4
> a 04 8b 46 2c 85 c0 0f 85 bf 00 00 00 c7 46 2c 01 00 00 00 8b 46 28 ba
> ff ff 00
> 00 <0f> c1 10 0f 85 f0 04 00 00 8b 5e 28 8b 43 10 8b 48 10 8b 41 04
> 
> after this the same thing... i cannot read nor write to the partition,
> and eventually have to reboot.
> 
> (now there's a little detail i didn't mention... this "computer" i'm
> testing on is in a vmware installed fedora core 2. but i don't think
> that could be a problem. or could it?)
> 
> Imre Gergely wrote:
> 
>>mmm, i forgot:
>>
>>[root@fc2 root]# uname -r
>>2.6.12.5
>>
>>lvm> version
>>  LVM version:     2.01.13-cvs (2005-06-14)
>>  Library version: 1.01.01 (2005-03-29)
>>  Driver version:  4.4.0
>>
>>
>>Imre Gergely wrote:
>>
>>
>>>hi
>>>
>>>i tried to create 2 snapshots of a LV, and i got in trouble ;) with one
>>>snapshot there are apparently no problems, with two the kernel gives me
>>>random oops-es and hangs. first, i tried with the root LV, then i tried
>>>with a simple partition, the result were the same. i created the two
>>>snapshots, after that i did a vgdisplay, there was everything alright, i
>>>mounted the origin LV, i copied some files, i watched the percentage
>>>increasing on the snapshots, then i did a 'ls -la' on the mounted
>>>directory and then the kernel said bye-bye... i don't see any kernel
>>>panic, just the usual register and memory dump, and the hang.
>>>
>>>any ideas? don't tell me i cannot have more than one snapshot on one
>>>given LV.
>>>
>>>thx.
>>>
>>>_______________________________________________
>>>linux-lvm mailing list
>>>linux-lvm@redhat.com
>>>https://www.redhat.com/mailman/listinfo/linux-lvm
>>>read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>>>
>>
>>
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@redhat.com
> https://www.redhat.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>