2002-04-21 16:29:20

by Jan Niehusmann

[permalink] [raw]
Subject: ext3: i_blocks is xx, should be yy

On an (otherwise clean, and cleanly unmounted) ext3 partition, I got
some ext2fs errors like the following:

Inode 295208, i_blocks is 112, should be 8. Fix<y>?
Inode 295342, i_blocks is 8, should be 0. Fix<y>?
Inode 295343, i_blocks is 8, should be 0. Fix<y>?
Inode 295344, i_blocks is 8, should be 0. Fix<y>?
[...]

I looked up the inodes with debug2fs and noticed that all of them belong
to files which probably had failed write attemps because the fs was
full.

This happened with 2.4.18-rc2-ac1, neither the latest one nor a
'release' version, but I don't remember reading about such a problem on
linux-kernel.

Jan


2002-04-21 16:35:45

by Russell King

[permalink] [raw]
Subject: Re: ext3: i_blocks is xx, should be yy

On Sun, Apr 21, 2002 at 06:29:15PM +0200, Jan Niehusmann wrote:
> On an (otherwise clean, and cleanly unmounted) ext3 partition, I got
> some ext2fs errors like the following:
>
> Inode 295208, i_blocks is 112, should be 8. Fix<y>?

Did the filesystem run out of free blocks at any point?

If so, the following could explain it:

http://marc.theaimsgroup.com/?l=linux-kernel&m=101778284030725&w=2

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2002-04-21 16:39:15

by Jan Niehusmann

[permalink] [raw]
Subject: Re: ext3: i_blocks is xx, should be yy

On Sun, Apr 21, 2002 at 05:35:35PM +0100, Russell King wrote:
> Did the filesystem run out of free blocks at any point?
> If so, the following could explain it:
> http://marc.theaimsgroup.com/?l=linux-kernel&m=101778284030725&w=2

Ah, yes, thanks - i missed this mail when I searched the archives.
This mail describes exactly what I have seen.

Jan