fsck errors with fsstress

fsck errors with fsstress

[Aneesh Kumar K.V]

Hi,

On x86_64 after fsstress with fallocate is run with 

fsstress -c -n1000 -p1000 -d /mnt/tmp/fsstress -f dwrite=0

e2fsprogs from next branch with HEAD at v1.40.8-163-g60dc00b

/usr/local/e2fsprogs/sbin/e2fsck -fnv /dev/sda5
e2fsck 1.40.8 (13-Mar-2008)
Pass 1: Checking inodes, blocks, and sizes
Inode 401326 has an invalid extent
	(logical block 76, physical block 2441549, invalid len 711)
Clear? no

Inode 401326, i_blocks is 1112, should be 0.  Fix? no

Inode 475743 has an invalid extent
	(logical block 226, physical block 2441673, invalid len 245)
Clear? no

Inode 475743, i_blocks is 328, should be 0.  Fix? no

Inode 603723 has an invalid extent
	(logical block 34, physical block 2441671, invalid len 316)
Clear? no

Inode 603723, i_blocks is 456, should be 0.  Fix? no

Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences:  -(118168--118173) -(174537--174541) -(371445--371462) -(592857--592859) -937329 -(961360--961363) -(968529--968547) -(1238135--1238143) -(1285426--1285432) -(1415800--1415801) -(1576655--1576681) -(2075764--2075776) -(2364415--2364420) -(2368402--2368410) -(2432102--2432111) -2432247 -(2433343--2433346) -(2435405--2435432) -(2441075--2441105) -(2441265--2441289) -(2441348--2441353) -2441549 -2441671 -2441673
Fix? no


/dev/sda5: ********** WARNING: Filesystem still has errors **********

Re: fsck errors with fsstress

[Theodore Tso]

On Tue, Apr 15, 2008 at 12:17:33PM +0530, Aneesh Kumar K.V wrote:
> 
> On x86_64 after fsstress with fallocate is run with 
> 
> fsstress -c -n1000 -p1000 -d /mnt/tmp/fsstress -f dwrite=0
> 
> e2fsprogs from next branch with HEAD at v1.40.8-163-g60dc00b
> 
> /usr/local/e2fsprogs/sbin/e2fsck -fnv /dev/sda5
> e2fsck 1.40.8 (13-Mar-2008)
> Pass 1: Checking inodes, blocks, and sizes
> Inode 401326 has an invalid extent
> 	(logical block 76, physical block 2441549, invalid len 711)
> Clear? no

I think this was fixed by Eric's patch:

    e2fsck: Only check PR_1_EXTENT_ENDS_BEYOND for leaf nodes

I've integrated this and pushed it out into the latest 'pu' branch.

          		     	 	   - Ted