mounting raid5 with different unit values

mounting raid5 with different unit values

[Raz]

Hello
I am trying to mount xfs over raid5. I am trying to mount it with
different values than
the true ones. the mount fails with a general error of :
mount -osunit=6144,swidth=18432 /dev/md1 /d1/
mount: wrong fs type, bad option, bad superblock on /dev/md1,
       or too many mounted file systems

the true values are:
xfs_info /d1/
meta-data=/dev/md1               isize=256    agcount=32, agsize=5687296 blks
         =                       sectsz=4096  attr=0
data     =                       bsize=4096   blocks=181989888, imaxpct=25
         =                       sunit=256    swidth=768 blks, unwritten=1
naming   =version 2              bsize=4096
log      =internal               bsize=4096   blocks=32768, version=2
         =                       sectsz=4096  sunit=1 blks
realtime =none                   extsz=3145728 blocks=0, rtextents=0


this same command did work on 2.6.15  while it failing on 2.6.17 .
why ?
-- 
Raz

Re: mounting raid5 with different unit values

[Justin Piszcz]

On Sun, 7 Oct 2007, Raz wrote:

> Hello
> I am trying to mount xfs over raid5. I am trying to mount it with
> different values than
> the true ones. the mount fails with a general error of :
> mount -osunit=6144,swidth=18432 /dev/md1 /d1/
> mount: wrong fs type, bad option, bad superblock on /dev/md1,
>        or too many mounted file systems
> 
> the true values are:
> xfs_info /d1/
> meta-data=/dev/md1               isize=256    agcount=32, agsize=5687296 blks
>          =                       sectsz=4096  attr=0
> data     =                       bsize=4096   blocks=181989888, imaxpct=25
>          =                       sunit=256    swidth=768 blks, unwritten=1
> naming   =version 2              bsize=4096
> log      =internal               bsize=4096   blocks=32768, version=2
>          =                       sectsz=4096  sunit=1 blks
> realtime =none                   extsz=3145728 blocks=0, rtextents=0
> 
> 
> this same command did work on 2.6.15  while it failing on 2.6.17 .
> why ?
> -- 
> Raz
> 
>

Good question but also be alert that 2.6.17-2.6.17.6 will/may corrupt your 
XFS filesystem due to a nasty bug.

Justin.

Re: mounting raid5 with different unit values

[Raz]

On 10/7/07, Justin Piszcz <jpiszcz@lucidpixels.com> wrote:
>
>
> On Sun, 7 Oct 2007, Raz wrote:
>
> > Hello
> > I am trying to mount xfs over raid5. I am trying to mount it with
> > different values than
> > the true ones. the mount fails with a general error of :
> > mount -osunit=6144,swidth=18432 /dev/md1 /d1/
> > mount: wrong fs type, bad option, bad superblock on /dev/md1,
> >        or too many mounted file systems
> >
> > the true values are:
> > xfs_info /d1/
> > meta-data=/dev/md1               isize=256    agcount=32, agsize=5687296 blks
> >          =                       sectsz=4096  attr=0
> > data     =                       bsize=4096   blocks=181989888, imaxpct=25
> >          =                       sunit=256    swidth=768 blks, unwritten=1
> > naming   =version 2              bsize=4096
> > log      =internal               bsize=4096   blocks=32768, version=2
> >          =                       sectsz=4096  sunit=1 blks
> > realtime =none                   extsz=3145728 blocks=0, rtextents=0
> >
> >
> > this same command did work on 2.6.15  while it failing on 2.6.17 .
> > why ?
> > --
> > Raz
> >
> >
>
> Good question but also be alert that 2.6.17-2.6.17.6 will/may corrupt your
> XFS filesystem due to a nasty bug.
>
> Justin.
>
I am using patched xfs ( the patch of 2.6.17.7)

-- 
Raz

Re: mounting raid5 with different unit values

[Chris Wedgwood]

On Sun, Oct 07, 2007 at 09:15:23AM +0200, Raz wrote:

> mount -osunit=6144,swidth=18432 /dev/md1 /d1/

those values are set when you make the filesystem, not mount it

see the mkfs.xfs man page

Re: mounting raid5 with different unit values

[Justin Piszcz]

On Sun, 7 Oct 2007, Chris Wedgwood wrote:

> On Sun, Oct 07, 2007 at 09:15:23AM +0200, Raz wrote:
>
>> mount -osunit=6144,swidth=18432 /dev/md1 /d1/
>
> those values are set when you make the filesystem, not mount it
>
> see the mkfs.xfs man page
>
>

?

man mount :)

        sunit=value and swidth=value
               Used to specify the stripe unit and width for a RAID device or a
               stripe volume.  value must be specified in 512-byte block units.
               If this option is not specified and the filesystem was made on a
               stripe volume or the stripe width or unit were specified for the
               RAID device at mkfs  time,  then  the  mount  system  call  will
               restore the value from the superblock.  For filesystems that are
               made directly on RAID devices, these  options  can  be  used  to
               override  the  information  in  the superblock if the underlying
               disk layout changes after the filesystem has been created.   The
               swidth  option  is  required if the sunit option has been speci-
               fied, and must be a multiple of the sunit value.

He should be able to specify those option from the command line using 
mount.

Justin.

Re: mounting raid5 with different unit values

[Chris Wedgwood]

On Sun, Oct 07, 2007 at 11:48:14AM -0400, Justin Piszcz wrote:

> man mount :)

Ah of course.

But those will be more restrictive that what you can specify when you
make the file-system (because mkfs.xfs can aligned the AGs to suit).

Re: mounting raid5 with different unit values

[Raz]

Well... this thing actually works just fine with a newer kernel (
2.6.18-8-el5 centos5 ). I managed to mount / mkfs.xfs over raid5 with
a pseudo raid5 unit size, and with the appropriate raid 5 patches and
user space access-pattern, I elimintaed in 99% cases the read penalty
.
I sincerly hope I won't be getting any crashes with this file system tunnings.
so ... first, chris and all you xfs guys, many many thanks.
Chris,  How "dangerous" these tunnings are ? Am I to expect "weird"
behaviour of the file system ?

On 10/8/07, Chris Wedgwood <cw@f00f.org> wrote:
> On Sun, Oct 07, 2007 at 11:48:14AM -0400, Justin Piszcz wrote:
>
> > man mount :)
>
> Ah of course.
>
> But those will be more restrictive that what you can specify when you
> make the file-system (because mkfs.xfs can aligned the AGs to suit).
>


-- 
Raz

Re: mounting raid5 with different unit values

[Raz]

On 10/7/07, Chris Wedgwood <cw@f00f.org> wrote:
> On Sun, Oct 07, 2007 at 09:15:23AM +0200, Raz wrote:
>
> > mount -osunit=6144,swidth=18432 /dev/md1 /d1/
>
> those values are set when you make the filesystem, not mount it
>
> see the mkfs.xfs man page
>
it was possible to mkfs the file system in different units in 2.6.15,
but it is is not possible to mkfs it with differrent unit sizes in
2.6.17.

-- 
Raz

Re: mounting raid5 with different unit values

[David Chinner]

On Sun, Oct 07, 2007 at 09:15:23AM +0200, Raz wrote:
> Hello
> I am trying to mount xfs over raid5. I am trying to mount it with
> different values than
> the true ones. the mount fails with a general error of :
> mount -osunit=6144,swidth=18432 /dev/md1 /d1/
> mount: wrong fs type, bad option, bad superblock on /dev/md1,
>        or too many mounted file systems

Try looking in your syslog or dmesg for the error that occurred
during mount. THere's the possibility that it won't give you
an error because of a bug that was fixed more recently than
2.6.17 that suppressd errors incorrectly. You shoul dreally
try a more recent kernel ;)

Most likely, the new sunit/swidth are being rejected because
of alignent reasons (e.g. the AG size is not a multiple of
the new sunit you are setting).

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group

Re: mounting raid5 with different unit values

[Eric Sandeen]

David Chinner wrote:
> On Sun, Oct 07, 2007 at 09:15:23AM +0200, Raz wrote:
>> Hello
>> I am trying to mount xfs over raid5. I am trying to mount it with
>> different values than
>> the true ones. the mount fails with a general error of :
>> mount -osunit=6144,swidth=18432 /dev/md1 /d1/
>> mount: wrong fs type, bad option, bad superblock on /dev/md1,
>>        or too many mounted file systems
> 
> Try looking in your syslog or dmesg for the error that occurred
> during mount. 

Just a general plea to everyone.  If you get that above generic mount
error ("... or planets out of alignment") you *really* need to look at
dmesg to get any idea what's actually gone wrong.

-Eric