[RAID1 Bug] bio too big device md0 (248 > 200) (2.6.9-rc2-mm1)

[RAID1 Bug] bio too big device md0 (248 > 200) (2.6.9-rc2-mm1)

[Raphaël Rigo]

Hello,
kernel version : 2.6.9-rc2-mm1
i'm using a RAID1 array over 2 disks : one ATA, and another one in 
"SATA" (if we can call the ICH5 a real SATA controller).
During array synchronisation, I get "bio too big device md0 (248 > 200)" 
error, which I fixed in doing
//#define RESYNC_BLOCK_SIZE (64*1024)
#define RESYNC_BLOCK_SIZE PAGE_SIZE
in raid1.c, following the instruction on an old thread (for kernel 2.6.0).
I would like to know if there is any better fix now, else then this mail 
  will act as a remainder ;)

Raphaël Rigo

Re: [RAID1 Bug] bio too big device md0 (248 > 200) (2.6.9-rc2-mm1)

[Neil Brown]

On Monday September 20, raphael.rigo@twilight-hall.net wrote:
> Hello,
> kernel version : 2.6.9-rc2-mm1
> i'm using a RAID1 array over 2 disks : one ATA, and another one in 
> "SATA" (if we can call the ICH5 a real SATA controller).
> During array synchronisation, I get "bio too big device md0 (248 > 200)" 
> error, which I fixed in doing
> //#define RESYNC_BLOCK_SIZE (64*1024)
> #define RESYNC_BLOCK_SIZE PAGE_SIZE
> in raid1.c, following the instruction on an old thread (for kernel 2.6.0).
> I would like to know if there is any better fix now, else then this mail 
>   will act as a remainder ;)

This is not (as far as I can tell) a raid1 bug.

   bio too big device md0

means that someone sent a request to md0 that was too large. (124K
instead of the max 100K).

raid1 never does that.  It send requests to the underlying devices.
So if you make your raid1 from hde and sda, then a message like
   bio to big device sda
might indicate a problem with raid1.

Are you sure you quoted the error message correctly?
If you, how was the array being used? What filesystem?

NeilBrown

Re: [RAID1 Bug] bio too big device md0 (248 > 200) (2.6.9-rc2-mm1)

[Raphaël Rigo]

Neil Brown wrote:
> On Monday September 20, raphael.rigo@twilight-hall.net wrote:
> 
>>Hello,
>>kernel version : 2.6.9-rc2-mm1
>>i'm using a RAID1 array over 2 disks : one ATA, and another one in 
>>"SATA" (if we can call the ICH5 a real SATA controller).
>>During array synchronisation, I get "bio too big device md0 (248 > 200)" 
>>error, which I fixed in doing
>>//#define RESYNC_BLOCK_SIZE (64*1024)
>>#define RESYNC_BLOCK_SIZE PAGE_SIZE
>>in raid1.c, following the instruction on an old thread (for kernel 2.6.0).
>>I would like to know if there is any better fix now, else then this mail 
>>  will act as a remainder ;)
> 
> 
> This is not (as far as I can tell) a raid1 bug.
> 
>    bio too big device md0
> 
> means that someone sent a request to md0 that was too large. (124K
> instead of the max 100K).
> 
> raid1 never does that.  It send requests to the underlying devices.
> So if you make your raid1 from hde and sda, then a message like
>    bio to big device sda
> might indicate a problem with raid1.
> 
> Are you sure you quoted the error message correctly?
> If you, how was the array being used? What filesystem?
> 
> NeilBrown
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/


Is I am sure I did the right quote (copy paste from kern.log).
The array was being recovered (it is a newly created array), copying 
from hda to sda. The filesystem used is ext3.
Btw the first number in the message did vary.

I did the modification in raid1.c based on this mail you wrote :
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&frame=right&th=69e566d020ab2da6&seekm=Pine.LNX.4.58.0312232253140.7841%40infocalypse.jimlawson.org.lucky.linux.kernel#link5

Hope this helps,

Raphaël Rigo