pata_via with software raid1: "attempt to access beyond end of device"

pata_via with software raid1: "attempt to access beyond end of device"

[Dâniel Fraga]

Linux 2.6.22
Intel(R) Celeron(R) CPU 2.26GHz
00:0f.1 IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT8233/A/C/VT8235 PIPC Bus Master IDE (rev 06)
Reiserfs3

	I have 2 PATA HDs and they always worked perfectly with
software (kernel) raid1 using the old IDE kernel drivers. Now I use
the new libata drivers (VIA), and it keep giving those erros messages:

attempt to access beyond end of device
md1: rw=0, want=155284112, limit=155284096

	Is it raid's fault? Or libata's fault? The raid1 is ok:

/dev/md1:
        Version : 00.90.03
  Creation Time : Wed Jul 11 03:38:05 2007
     Raid Level : raid1
     Array Size : 77642048 (74.05 GiB 79.51 GB)
  Used Dev Size : 77642048 (74.05 GiB 79.51 GB)
   Raid Devices : 2
  Total Devices : 2
Preferred Minor : 1
    Persistence : Superblock is persistent

    Update Time : Wed Jul 18 23:22:24 2007
          State : clean
 Active Devices : 2
Working Devices : 2
 Failed Devices : 0
  Spare Devices : 0

           UUID : 6aed7615:3574770f:268e26d7:19872c99
         Events : 0.29260

    Number   Major   Minor   RaidDevice State
       0       8        2        0      active sync   /dev/sda2
       1       8       18        1      active sync   /dev/sdb2

-- 
Linux 2.6.22: Holy Dancing Manatees, Batman!
http://www.lastfm.pt/user/danielfraga
http://u-br.net
Maroon 5 - "Shiver" (Songs About Jane)
Linux 2.6.22: Holy Dancing Manatees, Batman!
http://www.lastfm.pt/user/danielfraga
http://u-br.net
The Pointer Sisters - "Automatic" (Platinum & Gold Collection Series)

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Neil Brown]

On Wednesday July 18, fragabr@gmail.com wrote:
> Linux 2.6.22
> Intel(R) Celeron(R) CPU 2.26GHz
> 00:0f.1 IDE interface: VIA Technologies, Inc.
> VT82C586A/B/VT82C686/A/B/VT8233/A/C/VT8235 PIPC Bus Master IDE (rev 06)
> Reiserfs3
> 
> 	I have 2 PATA HDs and they always worked perfectly with
> software (kernel) raid1 using the old IDE kernel drivers. Now I use
> the new libata drivers (VIA), and it keep giving those erros messages:
> 
> attempt to access beyond end of device
> md1: rw=0, want=155284112, limit=155284096

The resierfs filesystem is trying to access beyond the end of the
raid1 array.  Maybe some indexing information in the array is
corrupted.
Did you recreate the array (mdadm --create) after changing to the new
drivers?  If so, could you have accidentally made it a little smaller?

Possibly try running fsck and see if it finds anything to fix.

NeilBrown

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Dâniel Fraga]

On Thu, 19 Jul 2007 13:54:50 +1000
Neil Brown <neilb@suse.de> wrote:

> The resierfs filesystem is trying to access beyond the end of the
> raid1 array.  Maybe some indexing information in the array is
> corrupted.
> Did you recreate the array (mdadm --create) after changing to the new
> drivers?  If so, could you have accidentally made it a little smaller?
> 
> Possibly try running fsck and see if it finds anything to fix.

	Hi Neil, I recreated the array after changing to libata. I
checked the whole filesystem with reiserfsck --rebuild-tree, then
started it again and the problem happens suddenly.

	And regarding the size of array, I just added the whole
partition to the array:

Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/md1              77639736  30127524  47512212  39% /

     Array Size : 77642048 (74.05 GiB 79.51 GB)
  Used Dev Size : 77642048 (74.05 GiB 79.51 GB)

	Can I compare both values (77639736 with 77642048)? If so,
there is a difference, if not, I don't know.

	Thank you!

	Ps: the error is difficult to appear, but I'll keep looking on
it and try to ellaborate a more detailed report. 

-- 
Linux 2.6.22: Holy Dancing Manatees, Batman!
http://www.lastfm.pt/user/danielfraga
http://u-br.net

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Neil Brown]

On Thursday July 19, fragabr@gmail.com wrote:
> On Thu, 19 Jul 2007 13:54:50 +1000
> Neil Brown <neilb@suse.de> wrote:
> 
> > The resierfs filesystem is trying to access beyond the end of the
> > raid1 array.  Maybe some indexing information in the array is
> > corrupted.
> > Did you recreate the array (mdadm --create) after changing to the new
> > drivers?  If so, could you have accidentally made it a little smaller?
> > 
> > Possibly try running fsck and see if it finds anything to fix.
> 
> 	Hi Neil, I recreated the array after changing to libata. I
> checked the whole filesystem with reiserfsck --rebuild-tree, then
> started it again and the problem happens suddenly.
> 
> 	And regarding the size of array, I just added the whole
> partition to the array:
> 
> Filesystem           1K-blocks      Used Available Use% Mounted on
> /dev/md1              77639736  30127524  47512212  39% /
> 
>      Array Size : 77642048 (74.05 GiB 79.51 GB)
>   Used Dev Size : 77642048 (74.05 GiB 79.51 GB)
> 
> 	Can I compare both values (77639736 with 77642048)? If so,
> there is a difference, if not, I don't know.

Filesystems always reserve some space for internal use, so the size
reported by df is always smaller than the total number of block on the
devices.

What does
      od -D -j 65536 -N 4 /dev/md1
show.  This is the size the reiserfs thinks it is using.  Multiply by
4 and you should get 77642048 or maybe a little less.  If you get
more, then reiserfs think the device is bigger than it is.  If don't
then I have no idea what the problem is.

NeilBrown

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Dâniel Fraga]

On Fri, 20 Jul 2007 12:14:31 +1000
Neil Brown <neilb@suse.de> wrote:

> What does
>       od -D -j 65536 -N 4 /dev/md1
> show.  This is the size the reiserfs thinks it is using.  Multiply by
> 4 and you should get 77642048 or maybe a little less.  If you get
> more, then reiserfs think the device is bigger than it is.  If don't
> then I have no idea what the problem is.

	The output is:

0200000   19410528
0200004

	so...

19410528 x 4 = 77642112

	The difference (77642048 - 77642112) is 64.

	Does it help? What could I do now? Thanks!

-- 
Linux 2.6.22: Holy Dancing Manatees, Batman!
http://www.lastfm.pt/user/danielfraga
http://u-br.net
Exodus - "I Am Abomination" (Shovel Headed Kill Machine - 2005)

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Neil Brown]

On Thursday July 19, fragabr@gmail.com wrote:
> On Fri, 20 Jul 2007 12:14:31 +1000
> Neil Brown <neilb@suse.de> wrote:
> 
> > What does
> >       od -D -j 65536 -N 4 /dev/md1
> > show.  This is the size the reiserfs thinks it is using.  Multiply by
> > 4 and you should get 77642048 or maybe a little less.  If you get
> > more, then reiserfs think the device is bigger than it is.  If don't
> > then I have no idea what the problem is.
> 
> 	The output is:
> 
> 0200000   19410528
> 0200004
> 
> 	so...
> 
> 19410528 x 4 = 77642112
> 
> 	The difference (77642048 - 77642112) is 64.

So reiserfs thinks the devices is 64K larger than it really is. I
wonder how that happened.

> 
> 	Does it help? What could I do now? Thanks!

It helps that we know why you get these errors.  It looks like you
should be able to fix it with resize_reiserfs.
Looking at the man page I have, you could
   umount /whatever
   resize_reiserfs -s -64K /dev/md1
   mount /whatever

However I recommend reading the man page yourself, as this isn't the
sort of thing you want to get wrong.

NeilBrown

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Dâniel Fraga]

On Fri, 20 Jul 2007 12:59:41 +1000
Neil Brown <neilb@suse.de> wrote:

> So reiserfs thinks the devices is 64K larger than it really is. I
> wonder how that happened.

>    resize_reiserfs -s -64K /dev/md1

	Neil, I discovered how the "64k larger" issue happened...

	It was my fault.. I created the Reiser filesystem *before* the
raid device. So when I create the raid device, it warns about a
existing reiserfs filesystem and if I want to continue. This way, the
underlying filesystem will be always 64k large than the raid device.

	If, instead, I use the correct method of creating the raid
device *before*, everything is ok, because then I can "mkreiserfs" the
raid device and the filesystem will have the correct size the raid
device expects.

	Sorry for my mistake and thanks again!

-- 
Linux 2.6.22: Holy Dancing Manatees, Batman!
http://www.lastfm.pt/user/danielfraga
http://u-br.net
Marilyn Manson - "Dried Up, Tied and Dead to the World" (Antichrist
Superstar - 1996)

Re: pata_via with software raid1: "attempt to access beyond end of device"

[Cry Regarder]

Dâniel Fraga <fragabr <at> gmail.com> writes:

> 	I have 2 PATA HDs and they always worked perfectly with
> software (kernel) raid1 using the old IDE kernel drivers. Now I use
> the new libata drivers (VIA), and it keep giving those erros messages:
> 
> attempt to access beyond end of device
> md1: rw=0, want=155284112, limit=155284096

Isn't this the standard problem that some hard drives have reserved
areas that the operating system isn't supposed to use?  The old IDE
drivers didn't honor the drive's request not to use.  libata, however,
now behaves correctly and honors it.

libata.ignore_hpa=1

https://www.redhat.com/archives/fedora-devel-list/2007-May/msg01695.html

and

http://fedoraproject.org/wiki/Bugs/F7Common


Now if someone can explain to me if it is safe to write the partition table
saying I only have one partition when I built a raid-5 on /dev/sd? instead of
/dev/sd?1

Thanks!

Cry

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html