Skip to contentSkip to conversation

dm-crypt / Software Raid5 issues on Ubuntu Lucid 64bit

dm-crypt Thorsten Peter 5 participants 11 messages Sep 3, 2010 – Sep 6, 2010

lore
Share & tools
lore
#0 lore raw

Hey folks,

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5 array. Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm: md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow
Aug 29 02:54:24 liberty24 kernel: [221558.020638] RIP: 0010:[<ffffffff814353b8>]  [<ffffffff814353b8>] clone_endio+0x38/0xe0
Aug 29 02:54:24 liberty24 kernel: [221558.020638] RSP: 0018:ffff880069531c40  EFLAGS: 00010246
Aug 29 02:54:24 liberty24 kernel: [221558.020638] RAX: ffffffffa023f760 RBX: 0000000000000000 RCX: 0100000000000081
Aug 29 02:54:24 liberty24 kernel: [221558.020638] RDX: 0000000000000017 RSI: 0000000000000000 RDI: ffffc9001247e040
Aug 29 02:54:24 liberty24 kernel: [221558.035928] RBP: ffff880069531c70 R08: 0000000000000000 R09: 010000000000282c
Aug 29 02:54:24 liberty24 kernel: [221558.035928] R10: 0000000000000000 R11: 0000000000000001 R12: ffff88000ca0ed80
Aug 29 02:54:24 liberty24 kernel: [221558.035928] R13: ffff88000df85ed0 R14: ffff8800695ca600 R15: ffbf8800116cb348
Aug 29 02:54:24 liberty24 kernel: [221558.035928] FS:  00007fb75e7f4700(0000) GS:ffff880001c40000(0000) knlGS:0000000000000000
Aug 29 02:54:24 liberty24 kernel: [221558.035928] CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
Aug 29 02:54:24 liberty24 kernel: [221558.035928] CR2: 00007f2d10190000 CR3: 0000000053ae8000 CR4: 00000000000006e0
Aug 29 02:54:24 liberty24 kernel: [221558.035928] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
Aug 29 02:54:24 liberty24 kernel: [221558.035928] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Aug 29 02:54:24 liberty24 kernel: [221558.035928] Process md0_raid5 (pid: 400, threadinfo ffff880069530000, task ffff88006a08dbc0)
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  ffff88001be85ed8 0000000000000000 0000000000000000 ffff88000ca0ed80
Aug 29 02:54:24 liberty24 kernel: [221558.035928] <0> ffff8800695ca600 ffff88001be85ed8 ffff880069531c80 ffffffff8117159d
Aug 29 02:54:24 liberty24 kernel: [221558.035928] <0> ffff880069531cb0 ffffffffa023d8c9 ffff880069531ca0 ffff88004d8312b0
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff8117159d>] bio_endio+0x1d/0x40
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffffa023d8c9>] crypt_dec_pending+0x69/0x90 [dm_crypt]
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffffa023daa8>] crypt_endio+0x68/0x150 [dm_crypt]
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff8117159d>] bio_endio+0x1d/0x40
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffffa0200add>] handle_stripe5+0x46d/0x9a0 [raid456]
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffffa0201028>] handle_stripe+0x18/0x30 [raid456]
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffffa0201432>] raid5d+0x202/0x320 [raid456]
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff814298dc>] md_thread+0x5c/0x130
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff81085090>] ? autoremove_wake_function+0x0/0x40
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff81429880>] ? md_thread+0x0/0x130
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff81084d16>] kthread+0x96/0xa0
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff810141ea>] child_rip+0xa/0x20
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff81084c80>] ? kthread+0x0/0xa0
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  [<ffffffff810141e0>] ? child_rip+0x0/0x20
Aug 29 02:54:24 liberty24 kernel: [221558.035928]  RSP <ffff880069531c40>
Aug 29 02:54:24 liberty24 kernel: [221558.133125] ---[ end trace 29f7fd9a7bbb5a00 ]---

Using the array uncrypted, data copy to it works just fine, copied over the full 900 GB now 6 times without a hickup.

Can anybody help me with this?

thanks

Thorsten

#1 lore raw
Quoted 5 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5 array. Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm: 

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

#2 lore raw

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am totally new to dm-crypt. Just used the first method that I found in the howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa repo.

Thanks,

Thorsten

Quoted 30 lines · Milan Broz
Quoted 2 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

Signature · 6 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
#3 lore raw

Sorry man, sent one of my msgs using a false E-Mail address, please reply to my other address if possible: [email protected]. That's the one that is registered in the list.

thx

Thorsten

Quoted 56 lines · Thorsten Peter

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am totally new to dm-crypt. Just used the first method that I found in the howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa repo.

Thanks,

Thorsten

Quoted 21 lines · Milan Broz
Quoted 2 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

Quoted 10 lines

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never
reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

--
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------

Signature · 6 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
#4 lore raw

I would recomend one read-thrugh of the FAQ. Just posted it today and it is also in the web here:
http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions

The most important thing to know is that if you trash the LUKS header, you are screwed ;-)

Arno

Quoted 82 lines · Thorsten Peter

Sorry man, sent one of my msgs using a false E-Mail address, please reply to my other address if possible: [email protected]. That's the one that is registered in the list.

thx

Thorsten

Quoted 56 lines · Thorsten Peter

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am totally new to dm-crypt. Just used the first method that I found in the howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa repo.

Thanks,

Thorsten

Quoted 21 lines · Milan Broz
Quoted 2 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

Quoted 10 lines

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never
reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

--
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------

Signature · 11 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------

_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt
Signature · 7 lines
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: [email protected] 
GnuPG:  ID: 1E25338F  FP: 0C30 5782 9D93 F785 E79C  0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

If it's in the news, don't worry about it.  The very definition of 
"news" is "something that hardly ever happens." -- Bruce Schneier 
#5 lore raw

Alright,

here's what I did now:

- upgraded Kernel from kernel ppa to 2.6.35-19-server #25~lucid1-Ubuntu
- recreated my encrypted device using the LUKS Encryption howto

On the first try, copying back my data to the encrypted array the machine froze totally on me at some point of time. Second try, I get the same kinda lockup as before with the following trace in log:

[36531.299878] Pid: 375, comm: md0_raid5 Not tainted 2.6.35-19-server #25~lucid1-Ubuntu RS780-SB700/Unknow
[36531.299961] RIP: 0010:[<ffffffff81467cc8>]  [<ffffffff81467cc8>] clone_endio+0x38/0xe0
[36531.300041] RSP: 0018:ffff880069587c40  EFLAGS: 00010246
[36531.300090] RAX: ffffffffa023f800 RBX: 0000000000000000 RCX: ffff88006d415490
[36531.300153] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffc9001247b040
[36531.300217] RBP: ffff880069587c70 R08: 0000000000000004 R09: 0000000000000000
[36531.300281] R10: ffff8800000180d8 R11: 0000000000000000 R12: ffff88004f7a5c00
[36531.300345] R13: ffff88000df85ed0 R14: ffff88006d2ca600 R15: ffbf88001dc318c0
[36531.300409] FS:  00007f00109f7700(0000) GS:ffff880001e40000(0000) knlGS:0000000000000000
[36531.300481] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[36531.300533] CR2: 00000000008fb308 CR3: 0000000068063000 CR4: 00000000000006e0
[36531.300597] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[36531.300661] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[36531.300725] Process md0_raid5 (pid: 375, threadinfo ffff880069586000, task ffff88006a3644a0)
[36531.300799] Stack:
[36531.300818]  ffff880050b4bab0 0000000000000000 0000000000000000 ffff88004f7a5c00
[36531.300891] <0> ffff88006d2ca600 ffff880050b4bab0 ffff880069587c80 ffffffff8117f4fd
[36531.300970] <0> ffff880069587cb0 ffffffffa023d959 ffff880069587ca0 ffff880015832670
[36531.301050] Call Trace:
[36531.301079]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301129]  [<ffffffffa023d959>] crypt_dec_pending+0x69/0x90 [dm_crypt]
[36531.301192]  [<ffffffffa023db38>] crypt_endio+0x68/0x150 [dm_crypt]
[36531.301250]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301300]  [<ffffffffa0200d3d>] handle_stripe5+0x4cd/0xa00 [raid456]
[36531.301362]  [<ffffffffa0201288>] handle_stripe+0x18/0x30 [raid456]
[36531.301421]  [<ffffffffa02016a2>] raid5d+0x212/0x330 [raid456]
[36531.301477]  [<ffffffff81035dc9>] ? default_spin_lock_flags+0x9/0x10
[36531.301538]  [<ffffffff8145be6c>] md_thread+0x5c/0x130
[36531.301586]  [<ffffffff8107dfe0>] ? autoremove_wake_function+0x0/0x40
[36531.301646]  [<ffffffff8145be10>] ? md_thread+0x0/0x130
[36531.301694]  [<ffffffff8107da86>] kthread+0x96/0xa0
[36531.301740]  [<ffffffff8100aee4>] kernel_thread_helper+0x4/0x10
[36531.301795]  [<ffffffff8107d9f0>] ? kthread+0x0/0xa0
[36531.301841]  [<ffffffff8100aee0>] ? kernel_thread_helper+0x0/0x10
[36531.301895] Code: 89 65 e0 4c 89 6d e8 4c 89 75 f0 4c 89 7d f8 0f 1f 44 00 00 4c 8b 6f 58 49 89 fc 85 f6 89 f3 49 8b 7d 08 4d 8b 7d 00 48 8b 47 08 <4d> 8b 37 48 8b 40 48 75 0e 41 f6 44 24 18 01 ba fb ff ff ff 0f
[36531.302215] RIP  [<ffffffff81467cc8>] clone_endio+0x38/0xe0
[36531.302269]  RSP <ffff880069587c40>
[36531.316119] ---[ end trace 2508e8b8766e359a ]---

I am really running out of options here, what else could be wrong?

Thorsten

Quoted 88 lines · Arno Wagner

I would recomend one read-thrugh of the FAQ. Just posted it today and it is also in the web here:
http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions

The most important thing to know is that if you trash the LUKS header, you are screwed ;-)

Arno

Quoted 77 lines · Thorsten Peter

Sorry man, sent one of my msgs using a false E-Mail address, please reply to my other address if possible: [email protected]. That's the one that is registered in the list.

thx

Thorsten

Quoted 53 lines · Thorsten Peter

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am totally new to dm-crypt. Just used the first method that I found in the howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa repo.

Thanks,

Thorsten

Quoted 19 lines · Milan Broz
Quoted 2 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

Quoted 10 lines

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never
reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

--
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------

Signature · 10 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt
#6 lore raw

Alright guys, I think I identified my issue. Just wanted to let you
know ....

A bad memory module :-(

Didn't recognize this until I ran Memtest86 yesterday after running out of ideas. Right now I am running 1x1GB DDR2-800 and am still testing to write to the array without lockups or other errors so far.

Thanks anyway for your help

Thorsten

Quoted 142 lines · Thorsten Peter

Alright,

here's what I did now:

- upgraded Kernel from kernel ppa to 2.6.35-19-server #25~lucid1-Ubuntu
- recreated my encrypted device using the LUKS Encryption howto

On the first try, copying back my data to the encrypted array the machine froze totally on me at some point of time. Second try, I get the same kinda lockup as before with the following trace in log:

[36531.299878] Pid: 375, comm: md0_raid5 Not tainted 2.6.35-19-server #25~lucid1-Ubuntu RS780-SB700/Unknow
[36531.299961] RIP: 0010:[<ffffffff81467cc8>]  [<ffffffff81467cc8>] clone_endio+0x38/0xe0
[36531.300041] RSP: 0018:ffff880069587c40  EFLAGS: 00010246
[36531.300090] RAX: ffffffffa023f800 RBX: 0000000000000000 RCX: ffff88006d415490
[36531.300153] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffc9001247b040
[36531.300217] RBP: ffff880069587c70 R08: 0000000000000004 R09: 0000000000000000
[36531.300281] R10: ffff8800000180d8 R11: 0000000000000000 R12: ffff88004f7a5c00
[36531.300345] R13: ffff88000df85ed0 R14: ffff88006d2ca600 R15: ffbf88001dc318c0
[36531.300409] FS:  00007f00109f7700(0000) GS:ffff880001e40000(0000) knlGS:0000000000000000
[36531.300481] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[36531.300533] CR2: 00000000008fb308 CR3: 0000000068063000 CR4: 00000000000006e0
[36531.300597] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[36531.300661] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[36531.300725] Process md0_raid5 (pid: 375, threadinfo ffff880069586000, task ffff88006a3644a0)
[36531.300799] Stack:
[36531.300818]  ffff880050b4bab0 0000000000000000 0000000000000000 ffff88004f7a5c00
[36531.300891] <0> ffff88006d2ca600 ffff880050b4bab0 ffff880069587c80 ffffffff8117f4fd
[36531.300970] <0> ffff880069587cb0 ffffffffa023d959 ffff880069587ca0 ffff880015832670
[36531.301050] Call Trace:
[36531.301079]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301129]  [<ffffffffa023d959>] crypt_dec_pending+0x69/0x90 [dm_crypt]
[36531.301192]  [<ffffffffa023db38>] crypt_endio+0x68/0x150 [dm_crypt]
[36531.301250]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301300]  [<ffffffffa0200d3d>] handle_stripe5+0x4cd/0xa00 [raid456]
[36531.301362]  [<ffffffffa0201288>] handle_stripe+0x18/0x30 [raid456]
[36531.301421]  [<ffffffffa02016a2>] raid5d+0x212/0x330 [raid456]
[36531.301477]  [<ffffffff81035dc9>] ? default_spin_lock_flags+0x9/0x10
[36531.301538]  [<ffffffff8145be6c>] md_thread+0x5c/0x130
[36531.301586]  [<ffffffff8107dfe0>] ? autoremove_wake_function+0x0/0x40
[36531.301646]  [<ffffffff8145be10>] ? md_thread+0x0/0x130
[36531.301694]  [<ffffffff8107da86>] kthread+0x96/0xa0
[36531.301740]  [<ffffffff8100aee4>] kernel_thread_helper+0x4/0x10
[36531.301795]  [<ffffffff8107d9f0>] ? kthread+0x0/0xa0
[36531.301841]  [<ffffffff8100aee0>] ? kernel_thread_helper+0x0/0x10
[36531.301895] Code: 89 65 e0 4c 89 6d e8 4c 89 75 f0 4c 89 7d f8 0f 1f 44 00 00 4c 8b 6f 58 49 89 fc 85 f6 89 f3 49 8b 7d 08 4d 8b 7d 00 48 8b 47 08 <4d> 8b 37 48 8b 40 48 75 0e 41 f6 44 24 18 01 ba fb ff ff ff 0f
[36531.302215] RIP  [<ffffffff81467cc8>] clone_endio+0x38/0xe0
[36531.302269]  RSP <ffff880069587c40>
[36531.316119] ---[ end trace 2508e8b8766e359a ]---

I am really running out of options here, what else could be wrong?

Thorsten

Quoted 88 lines · Arno Wagner

I would recomend one read-thrugh of the FAQ. Just posted it today and it is also in the web here:
http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions

The most important thing to know is that if you trash the LUKS header, you are screwed ;-)

Arno

Quoted 77 lines · Thorsten Peter

Sorry man, sent one of my msgs using a false E-Mail address, please reply to my other address if possible: [email protected]. That's the one that is registered in the list.

thx

Thorsten

Quoted 53 lines · Thorsten Peter

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am totally new to dm-crypt. Just used the first method that I found in the howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa repo.

Thanks,

Thorsten

Quoted 29 lines · Milan Broz
Quoted 2 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 9 lines

Format went fine, but when I try to copy back large amount of data to the encrypted array (about 900gig) my copy thread always hangs itself at some point during the process. I can't access the array anymore, though the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok.
When the problem occures I see the following trace in log:

Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never
reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

Signature · 6 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
Signature · 10 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt
#7 lore raw

Thanks for the info.

Arno

Quoted 190 lines · BOBA FETT

Alright guys, I think I identified my issue. Just wanted to let you
know ....

A bad memory module :-(

Didn't recognize this until I ran Memtest86 yesterday after running out of ideas. Right now I am running 1x1GB DDR2-800 and am still testing to write to the array without lockups or other errors so far.

Thanks anyway for your help

Thorsten

Quoted 70 lines · Thorsten Peter

Alright,

here's what I did now:

- upgraded Kernel from kernel ppa to 2.6.35-19-server #25~lucid1-Ubuntu
- recreated my encrypted device using the LUKS Encryption howto

On the first try, copying back my data to the encrypted array the machine froze totally on me at some point of time.
Second try, I get the same kinda lockup as before with the following trace in log:

[36531.299878] Pid: 375, comm: md0_raid5 Not tainted 2.6.35-19-server  
#25~lucid1-Ubuntu RS780-SB700/Unknow
[36531.299961] RIP: 0010:[<ffffffff81467cc8>]  [<ffffffff81467cc8>]  
clone_endio+0x38/0xe0
[36531.300041] RSP: 0018:ffff880069587c40  EFLAGS: 00010246
[36531.300090] RAX: ffffffffa023f800 RBX: 0000000000000000 RCX:  
ffff88006d415490
[36531.300153] RDX: 0000000000000000 RSI: 0000000000000000 RDI:  
ffffc9001247b040
[36531.300217] RBP: ffff880069587c70 R08: 0000000000000004 R09:  
0000000000000000
[36531.300281] R10: ffff8800000180d8 R11: 0000000000000000 R12:  
ffff88004f7a5c00
[36531.300345] R13: ffff88000df85ed0 R14: ffff88006d2ca600 R15:  
ffbf88001dc318c0
[36531.300409] FS:  00007f00109f7700(0000) GS:ffff880001e40000(0000)  
knlGS:0000000000000000
[36531.300481] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[36531.300533] CR2: 00000000008fb308 CR3: 0000000068063000 CR4:  
00000000000006e0
[36531.300597] DR0: 0000000000000000 DR1: 0000000000000000 DR2:  
0000000000000000
[36531.300661] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:  
0000000000000400
[36531.300725] Process md0_raid5 (pid: 375, threadinfo  
ffff880069586000, task ffff88006a3644a0)
[36531.300799] Stack:
[36531.300818]  ffff880050b4bab0 0000000000000000 0000000000000000  
ffff88004f7a5c00
[36531.300891] <0> ffff88006d2ca600 ffff880050b4bab0 ffff880069587c80  
ffffffff8117f4fd
[36531.300970] <0> ffff880069587cb0 ffffffffa023d959 ffff880069587ca0  
ffff880015832670
[36531.301050] Call Trace:
[36531.301079]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301129]  [<ffffffffa023d959>] crypt_dec_pending+0x69/0x90  
[dm_crypt]
[36531.301192]  [<ffffffffa023db38>] crypt_endio+0x68/0x150 [dm_crypt]
[36531.301250]  [<ffffffff8117f4fd>] bio_endio+0x1d/0x40
[36531.301300]  [<ffffffffa0200d3d>] handle_stripe5+0x4cd/0xa00 [raid456]
[36531.301362]  [<ffffffffa0201288>] handle_stripe+0x18/0x30 [raid456]
[36531.301421]  [<ffffffffa02016a2>] raid5d+0x212/0x330 [raid456]
[36531.301477]  [<ffffffff81035dc9>] ? default_spin_lock_flags+0x9/0x10
[36531.301538]  [<ffffffff8145be6c>] md_thread+0x5c/0x130
[36531.301586]  [<ffffffff8107dfe0>] ? autoremove_wake_function+0x0/0x40
[36531.301646]  [<ffffffff8145be10>] ? md_thread+0x0/0x130
[36531.301694]  [<ffffffff8107da86>] kthread+0x96/0xa0
[36531.301740]  [<ffffffff8100aee4>] kernel_thread_helper+0x4/0x10
[36531.301795]  [<ffffffff8107d9f0>] ? kthread+0x0/0xa0
[36531.301841]  [<ffffffff8100aee0>] ? kernel_thread_helper+0x0/0x10
[36531.301895] Code: 89 65 e0 4c 89 6d e8 4c 89 75 f0 4c 89 7d f8 0f  
1f 44 00 00 4c 8b 6f 58 49 89 fc 85 f6 89 f3 49 8b 7d 08 4d 8b 7d 00  
48 8b 47 08 <4d> 8b 37 48 8b 40 48 75 0e 41 f6 44 24 18 01 ba fb ff ff  
ff 0f
[36531.302215] RIP  [<ffffffff81467cc8>] clone_endio+0x38/0xe0
[36531.302269]  RSP <ffff880069587c40>
[36531.316119] ---[ end trace 2508e8b8766e359a ]---
Quoted 101 lines

I am really running out of options here, what else could be wrong?

Thorsten

Quoted 96 lines · Arno Wagner

I would recomend one read-thrugh of the FAQ. Just posted it today and it is also in the web here:
http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions

The most important thing to know is that if you trash the LUKS header, you are screwed ;-)

Arno

Quoted 85 lines · Thorsten Peter

Sorry man, sent one of my msgs using a false E-Mail address, please reply to
my other address if possible: [email protected]. That's the one that is registered in the list.

thx

Thorsten

Quoted 60 lines · Thorsten Peter

Hey Milan,

could you give me a pointer on how to setup the encryption using LUKS? I am
totally new to dm-crypt. Just used the first method that I found in the
howto ....
Right now I am following Ricks advice and upgrading to 2.6.35 from a ppa
repo.

Thanks,

Thorsten

Quoted 33 lines · Milan Broz
Quoted 3 lines · Thorsten Peter

I am switching from loop-aes to dm_crypt. I am trying to use it on a fresh Ubuntu 64bit server installation with my 6x750gb Software raid5

array.

Quoted 3 lines

Array is clean and freshly created, I encrypted it using

cryptsetup -y -c aes -h ripemd160 -s 256 create STUFF /dev/md0

Please use LUKS instead here to get persistent header.

Quoted 10 lines

Format went fine, but when I try to copy back large amount of data to
the encrypted array (about 900gig) my copy thread always hangs itself at
some point during the process. I can't access the array anymore, though
the raid itself is fine, no drives are kicked out or anything. Reboot gives me back access and the array is still ok. When the problem occures I see the following trace in log:

Quoted 2 lines
Aug 29 02:54:24 liberty24 kernel: [221558.020638] Pid: 400, comm:

md0_raid5 Not tainted 2.6.32-24-server #41-Ubuntu Unknow

If it is reproducible with upstream kernel, please send me OOPs, if not, please use Ubuntu bugzilla - I saw strange reports which were never
reproducible on upstream.

(The bug seems like something we fixed long time ago.)

Milan

Signature · 6 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
Signature · 10 lines
--------------------------------
Dipl.-Inf. (FH) Thorsten Peter

E-Mail: [email protected]
Mobile: 01795322057
--------------------------------
_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt

_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt

Signature · 7 lines
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: [email protected] 
GnuPG:  ID: 1E25338F  FP: 0C30 5782 9D93 F785 E79C  0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

If it's in the news, don't worry about it.  The very definition of 
"news" is "something that hardly ever happens." -- Bruce Schneier 
#8 lore raw
In reply to BOBA FETT
Quoted 4 lines · BOBA FETT

Alright guys, I think I identified my issue. Just wanted to let you
know ....

A bad memory module :-(

welcome to the club - this is at least 5th report which was caused by broken RAM :-)

I think we should add this to FAQ.
raid5 + dm-crypt (+ XFS) seems to be nice hw stress test...

Milan

#9 lore raw
Quoted 6 lines · Milan Broz
Quoted 4 lines · BOBA FETT

Alright guys, I think I identified my issue. Just wanted to let you
know ....

A bad memory module :-(

Quoted 8 lines

welcome to the club - this is at least 5th report which was caused by broken RAM :-)

I think we should add this to FAQ.
raid5 + dm-crypt (+ XFS) seems to be nice hw stress test...

Milan

I will add an item. Fortunately I have not been hit myself by RAM problems it 7 years, but before I had one module with a weak bit in a 22 machine cluster with process migration. That was fun!

Arno

Quoted 5 lines

_______________________________________________
dm-crypt mailing list
[email protected]
http://www.saout.de/mailman/listinfo/dm-crypt

Signature · 7 lines
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: [email protected] 
GnuPG:  ID: 1E25338F  FP: 0C30 5782 9D93 F785 E79C  0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

If it's in the news, don't worry about it.  The very definition of 
"news" is "something that hardly ever happens." -- Bruce Schneier 
#10 lore raw
In reply to Milan Broz
Quoted 2 lines · Milan Broz

welcome to the club - this is at least 5th report which was caused by broken RAM :-)

A lot of new Gigabyte boards have something called "dynamic overclocking". It's difficult to discover at this is enabled by default. This leads to higher voltage for the CPU and RAM, and of course more GHz, which can give the same symptoms as bad RAM (and get the memtest86 to fail in the same way).

I have a Gigabyte MA770-UD3 myself, and it's a quite fantastic board, rockstable and reliable, AFTER I set the clockspeed and voltage for both CPU and RAM to "manual" :-)

I had the same symptoms before...