On Mon, Jun 30, 2008 at 04:10:56PM +0200, Paul Slootman wrote:
> Package: e2fsprogs
> Version: 1.40.8-2
> Severity: normal
>
> /dev/mapper/vgbig-loebas is 6.1T, /dev/mapper/vg11-loebas--journal is 500M
>
> [EMAIL PROTECTED]:~# mke2fs -L loebas -I 128 -j -J
> device=/dev/mapper/vg11-loebas--journal -m 0 -O
> large_file,dir_index,filetype,resize_inode /dev/mapper/vgbig-loebas
> mke2fs 1.40.8 (13-Mar-2008)
> mke2fs: File too large while trying to determine filesystem size
>
> [EMAIL PROTECTED]:~# tune2fs -J device=/dev/mapper/vg11-loebas--journal
> /dev/mapper/vgbig-loebas
> tune2fs 1.40.8 (13-Mar-2008)
> tune2fs: Filesystem has unexpected block size
> while trying to open journal on /dev/mapper/vg11-loebas--journal
So what happened is that you created the external journal with some
blocksize other than 4k. I'm guessing 1k, perhaps? There is a
requirement that the blocksize of the external journal device matches
the blocksize of the filesystem using the external journal device. So
what mke2fs tries to do helpfully is to create new filesystem with the
blocksize used by the external journal device, if it exists. So when
you specified -J device=/dev/mapper/vg11-loebas--journal, it did an
implicit -b 1024. Unfortunately, with a 6.1T filesystem, using a 1k
blocksize, the number of blocks needed to cover 6.1T don't fit in
2**32 blocks.
You're right, we should have better error message for this case.
I'm curious though --- why did you create the external journal with a
1k blocksize? Was this deliberate? Or accidental?
- Ted
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
