[gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Joshua Hoblitt]

[-- Attachment #1: Type: text/plain, Size: 1023 bytes --]

After reading in the kernel changelog that 2.6.19 should support ext3
filesystems up to 16TB in size, I tried to create an 11TB filesystem.
Attempting to do so gets this error from mke2fs:

# mke2fs -j -J size=400 -O sparse_super,dir_index -i32768 -m1 /dev/sda4 
mke2fs 1.39 (29-May-2006)
mke2fs: Filesystem too large.  No more than 2**31-1 blocks
         (8TB using a blocksize of 4k) are currently supported.
# mke2fs -j /dev/sda4
mke2fs 1.39 (29-May-2006)
mke2fs: Filesystem too large.  No more than 2**31-1 blocks
         (8TB using a blocksize of 4k) are currently supported.

This is on a system running 2.6.19-gentoo-r5 with sys-fs/e2fsprogs-1.39

ChangeLog for in kernel 16TB ext3 support:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=855565e81ad8940cc645b5110ec2c7f124a76d23

Release notes for e2fsprogs claiming 16TB support (2**32 blocks).

http://e2fsprogs.sourceforge.net/e2fsprogs-release.html#1.39

Any idea what I am missing?

Cheers,

-J

--

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Joshua Hoblitt]

[-- Attachment #1: Type: text/plain, Size: 1291 bytes --]

Ps.  I've looked through e2fsprogs configure.in and it doesn't look like
there's a magic flag to enable this support either.

-J

--
On Mon, Apr 09, 2007 at 02:09:10PM -1000, Joshua Hoblitt wrote:
> After reading in the kernel changelog that 2.6.19 should support ext3
> filesystems up to 16TB in size, I tried to create an 11TB filesystem.
> Attempting to do so gets this error from mke2fs:
> 
> # mke2fs -j -J size=400 -O sparse_super,dir_index -i32768 -m1 /dev/sda4 
> mke2fs 1.39 (29-May-2006)
> mke2fs: Filesystem too large.  No more than 2**31-1 blocks
>          (8TB using a blocksize of 4k) are currently supported.
> # mke2fs -j /dev/sda4
> mke2fs 1.39 (29-May-2006)
> mke2fs: Filesystem too large.  No more than 2**31-1 blocks
>          (8TB using a blocksize of 4k) are currently supported.
> 
> This is on a system running 2.6.19-gentoo-r5 with sys-fs/e2fsprogs-1.39
> 
> ChangeLog for in kernel 16TB ext3 support:
> 
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=855565e81ad8940cc645b5110ec2c7f124a76d23
> 
> Release notes for e2fsprogs claiming 16TB support (2**32 blocks).
> 
> http://e2fsprogs.sourceforge.net/e2fsprogs-release.html#1.39
> 
> Any idea what I am missing?
> 
> Cheers,
> 
> -J
> 
> --



[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Naga]

On Tuesday 10 April 2007 02.09.10 Joshua Hoblitt wrote:
> After reading in the kernel changelog that 2.6.19 should support ext3
> filesystems up to 16TB in size, I tried to create an 11TB filesystem.
> Attempting to do so gets this error from mke2fs:
>
> # mke2fs -j -J size=400 -O sparse_super,dir_index -i32768 -m1 /dev/sda4
> mke2fs 1.39 (29-May-2006)
> mke2fs: Filesystem too large.  No more than 2**31-1 blocks
>          (8TB using a blocksize of 4k) are currently supported.
[...]
> Any idea what I am missing?

Bigger blocksize?

-- 
Naga
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Joshua Hoblitt]

[-- Attachment #1: Type: text/plain, Size: 1889 bytes --]

After looking through various patches that were posted to LKML, I
discovered that passing "-F" to mke2fs seems to make this work as
expected.  However, this behavior seems to be undocumented and that
makes me more then a little nervous.

The resulting filesystem does seem to be mountable:

    /dev/sda4               11T   173M    11T   1% /export/ipp003.0

Is this feature undocumented for a reason or is it just unsupported?

-J

--
On Mon, Apr 09, 2007 at 02:12:43PM -1000, Joshua Hoblitt wrote:
> Ps.  I've looked through e2fsprogs configure.in and it doesn't look like
> there's a magic flag to enable this support either.
> 
> -J
> 
> --
> On Mon, Apr 09, 2007 at 02:09:10PM -1000, Joshua Hoblitt wrote:
> > After reading in the kernel changelog that 2.6.19 should support ext3
> > filesystems up to 16TB in size, I tried to create an 11TB filesystem.
> > Attempting to do so gets this error from mke2fs:
> > 
> > # mke2fs -j -J size=400 -O sparse_super,dir_index -i32768 -m1 /dev/sda4 
> > mke2fs 1.39 (29-May-2006)
> > mke2fs: Filesystem too large.  No more than 2**31-1 blocks
> >          (8TB using a blocksize of 4k) are currently supported.
> > # mke2fs -j /dev/sda4
> > mke2fs 1.39 (29-May-2006)
> > mke2fs: Filesystem too large.  No more than 2**31-1 blocks
> >          (8TB using a blocksize of 4k) are currently supported.
> > 
> > This is on a system running 2.6.19-gentoo-r5 with sys-fs/e2fsprogs-1.39
> > 
> > ChangeLog for in kernel 16TB ext3 support:
> > 
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=855565e81ad8940cc645b5110ec2c7f124a76d23
> > 
> > Release notes for e2fsprogs claiming 16TB support (2**32 blocks).
> > 
> > http://e2fsprogs.sourceforge.net/e2fsprogs-release.html#1.39
> > 
> > Any idea what I am missing?
> > 
> > Cheers,
> > 
> > -J
> > 
> > --
> 
> 



[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

[gentoo-amd64] Re: trouble creating an ext3 filesystem > 8TB

[Duncan]

Joshua Hoblitt <jhoblitt@ifa.hawaii.edu> posted
20070410211253.GA24125@ifa.hawaii.edu, excerpted below, on  Tue, 10 Apr
2007 11:12:53 -1000:

> After looking through various patches that were posted to LKML, I
> discovered that passing "-F" to mke2fs seems to make this work as
> expected.  However, this behavior seems to be undocumented and that
> makes me more then a little nervous.
> 
> The resulting filesystem does seem to be mountable:
> 
>     /dev/sda4               11T   173M    11T   1% /export/ipp003.0
> 
> Is this feature undocumented for a reason or is it just unsupported?

Hmm... no clue about e2fs (reiserfs here), but -F = force, so yeah, I can 
see why you'd be rather uncomfortable using it.

>From following the various kernel articles on LWN, including some eXfs 
ones (X=2/3/4), I could strongly speculate that they are doing it to 
enable backward compatibility.  Since the 16 T support is so new, they 
don't enable it by default, thus allowing older kernels to boot the 
resulting filesystems as well.  If one knows what they are doing and 
presumably knows not to try to load the filesystems with older kernels 
that can't properly handle 16 T partitions, there's the force switch to 
enable it, but it doesn't do it by default to ensure backward 
compatibility by default.

However, that would be simply speculation, altho I'd like to think it's 
at least somewhat informed speculation.  It's your data and your risk if 
I'm wrong, not mine, so feel free to treat this as worth exactly what you 
paid for it, zero. =8^)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Wil Reichert]

On 4/10/07, Joshua Hoblitt <jhoblitt@ifa.hawaii.edu> wrote:
> After looking through various patches that were posted to LKML, I
> discovered that passing "-F" to mke2fs seems to make this work as
> expected.  However, this behavior seems to be undocumented and that
> makes me more then a little nervous.
>
> The resulting filesystem does seem to be mountable:
>
>     /dev/sda4               11T   173M    11T   1% /export/ipp003.0
>
> Is this feature undocumented for a reason or is it just unsupported?

Any reason you are so set on ext3?  Its fine for a general purpose fs
but this might be a sign you are stretching its intended limits.
Perhaps something 64-bit like xfs or jfs would be a better option
depending on your intended usage / workload.

Wil
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Mark Haney]

Joshua Hoblitt wrote:
> After reading in the kernel changelog that 2.6.19 should support ext3
> filesystems up to 16TB in size, I tried to create an 11TB filesystem.
> Attempting to do so gets this error from mke2fs:
> 
> # mke2fs -j -J size=400 -O sparse_super,dir_index -i32768 -m1 /dev/sda4 
> mke2fs 1.39 (29-May-2006)
> mke2fs: Filesystem too large.  No more than 2**31-1 blocks
>          (8TB using a blocksize of 4k) are currently supported.
> # mke2fs -j /dev/sda4
> mke2fs 1.39 (29-May-2006)
> mke2fs: Filesystem too large.  No more than 2**31-1 blocks
>          (8TB using a blocksize of 4k) are currently supported.
> 
> This is on a system running 2.6.19-gentoo-r5 with sys-fs/e2fsprogs-1.39
> 
> ChangeLog for in kernel 16TB ext3 support:
> 
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=855565e81ad8940cc645b5110ec2c7f124a76d23
> 
> Release notes for e2fsprogs claiming 16TB support (2**32 blocks).
> 
> http://e2fsprogs.sourceforge.net/e2fsprogs-release.html#1.39
> 
> Any idea what I am missing?
> 
> Cheers,
> 
> -J
> 
> --

You know, I've been watching this thread thinking something was nagging 
me about this.  And after a night of not thinking about it I remember 
what it was.  I had a similar issue last summer.  The way I fixed it was 
to use LVM.  Create a volume group the size you want and create the 
filesystem inside it.  My volume was only slightly larger than 8TB, but 
it did create and it works great.


-- 
Ita erat quando hic adveni.

Mark Haney
Sr. Systems Administrator
ERC Broadband
(828) 350-2415
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] trouble creating an ext3 filesystem > 8TB

[Joshua Hoblitt]

[-- Attachment #1: Type: text/plain, Size: 1580 bytes --]

On Wed, Apr 11, 2007 at 02:06:44AM -0700, Wil Reichert wrote:
> On 4/10/07, Joshua Hoblitt <jhoblitt@ifa.hawaii.edu> wrote:
> >After looking through various patches that were posted to LKML, I
> >discovered that passing "-F" to mke2fs seems to make this work as
> >expected.  However, this behavior seems to be undocumented and that
> >makes me more then a little nervous.
> >
> >The resulting filesystem does seem to be mountable:
> >
> >    /dev/sda4               11T   173M    11T   1% /export/ipp003.0
> >
> >Is this feature undocumented for a reason or is it just unsupported?
> 
> Any reason you are so set on ext3?  Its fine for a general purpose fs
> but this might be a sign you are stretching its intended limits.
> Perhaps something 64-bit like xfs or jfs would be a better option
> depending on your intended usage / workload.

Yes, ext3 has proven itself, to my satisfaction, to be extremely
reliable even if it's not mounted with the journal in ordered mode.  In
fact, I've never lost data from an ext3 filesystem that wasn't the fault
of the hardware (disk/controller failure).  I've dabbled with xfs and
had some issues -- I've spoken with people that have had data corrupt
issues with xfs when the system has exhausted memory.  For my usage
pattern, xfs only gives me about a 10% performance boost and frankly I
don't need the extra bandwidth.  I don't really have any experience with
jfs.  At least to me, ext3 is clearly the most stable choice.  Even
lustrefs uses ext3 as the storage backend because of it's stability.

-J

--

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]