2015-07-22 14:18:08

by Pavel Machek

Subject: two fsck.ext4's needed to clean filesystem

Hi!

pavel@amd:~/misc$ /sbin/fsck.ext4 -y -f delme
...
delme: ***** FILE SYSTEM WAS MODIFIED *****
delme: 2054/25064 files (0.1% non-contiguous), 98231/100000 blocks
pavel@amd:~/misc$ /sbin/fsck.ext4 -y -f delme
e2fsck 1.42.12 (29-Aug-2014)
Backing up journal inode block information.

Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information

delme: ***** FILE SYSTEM WAS MODIFIED *****
delme: 2054/25064 files (0.5% non-contiguous), 98231/100000 blocks
pavel@amd:~/misc$ cp delme.fsck.original-damaged delme.report-1
pavel@amd:~/misc$ /sbin/fsck.ext4 -V
e2fsck 1.42.12 (29-Aug-2014)
Using EXT2FS Library version 1.42.12, 29-Aug-2014
pavel@amd:~/misc$

Is that considered ok? Should fsck restart itself or report
'filesystem still has errors' in this case?

Thanks and best regards,
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2015-07-22 14:22:01

by Pavel Machek

Subject: Re: two fsck.ext4's needed to clean filesystem

Hi!

> Is that considered ok? Should fsck restart itself or report
> 'filesystem still has errors' in this case?

It definitely should in this case:

Free inodes count wrong (22999, counted=23011).
Fix? yes


delme.fsck.damaged: ***** FILE SYSTEM WAS MODIFIED *****
delme.fsck.damaged: 2053/25064 files (0.1% non-contiguous),
97825/100000 blocks
Second check...
e2fsck 1.42.12 (29-Aug-2014)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Block bitmap differences: +18015
Fix? yes

Free blocks count wrong for group #2 (36, counted=35).
Fix? yes

Free blocks count wrong (2175, counted=2174).
Fix? yes


delme.fsck.damaged: ***** FILE SYSTEM WAS MODIFIED *****
delme.fsck.damaged: 2053/25064 files (0.2% non-contiguous),
97826/100000 blocks
Fsck lied about its success (result = 1)
pavel@amd:~/misc$ cp delme.fsck.original-damaged delme.report-2

(I can make the images public. Yes, it is artificial test.)

Best regards,
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

2015-07-22 16:33:01

by Eric Sandeen

Subject: Re: two fsck.ext4's needed to clean filesystem

On 7/22/15 9:21 AM, Pavel Machek wrote:
> Hi!
>
>> Is that considered ok? Should fsck restart itself or report
>> 'filesystem still has errors' in this case?

It should.

> It definitely should in this case:
>
> Free inodes count wrong (22999, counted=23011).
> Fix? yes
>
>
> delme.fsck.damaged: ***** FILE SYSTEM WAS MODIFIED *****
> delme.fsck.damaged: 2053/25064 files (0.1% non-contiguous),
> 97825/100000 blocks
> Second check...
> e2fsck 1.42.12 (29-Aug-2014)
> Pass 1: Checking inodes, blocks, and sizes
> Pass 2: Checking directory structure
> Pass 3: Checking directory connectivity
> Pass 4: Checking reference counts
> Pass 5: Checking group summary information
> Block bitmap differences: +18015
> Fix? yes
>
> Free blocks count wrong for group #2 (36, counted=35).
> Fix? yes
>
> Free blocks count wrong (2175, counted=2174).
> Fix? yes
>
>
> delme.fsck.damaged: ***** FILE SYSTEM WAS MODIFIED *****
> delme.fsck.damaged: 2053/25064 files (0.2% non-contiguous),
> 97826/100000 blocks
> Fsck lied about its success (result = 1)
> pavel@amd:~/misc$ cp delme.fsck.original-damaged delme.report-2
>
> (I can make the images public. Yes, it is artificial test.)

Please do - you're right, it should fix it on the first pass.

Thanks,
-Eric

2015-07-24 14:27:44

by Pavel Machek

Subject: Re: two fsck.ext4's needed to clean filesystem

Hi!

> >> Is that considered ok? Should fsck restart itself or report
> >> 'filesystem still has errors' in this case?
>
> It should.

Ok, good.

> > delme.fsck.damaged: ***** FILE SYSTEM WAS MODIFIED *****
> > delme.fsck.damaged: 2053/25064 files (0.2% non-contiguous),
> > 97826/100000 blocks
> > Fsck lied about its success (result = 1)
> > pavel@amd:~/misc$ cp delme.fsck.original-damaged delme.report-2
> >
> > (I can make the images public. Yes, it is artificial test.)
>
> Please do - you're right, it should fix it on the first pass.

I put them on

http://atrey.karlin.mff.cuni.cz/~pavel/outgoing/delme.report-1
http://atrey.karlin.mff.cuni.cz/~pavel/outgoing/delme.report-2

. Let me know if you can reproduce it.

Thanks and best regards,
Pavel
--
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html