[Ocfs2-users] How to fsck.ocfs2

[Werner Flamme]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Sunil Mushran [18.02.2011 22:34]:
> Unsure why that error does occurs when run without -f and not
> when run with -f. I will need to see a o2image of the volume to
> analyse.
> 
> fsck makes use of available memory. The timing depends on the
> size of the volume, number of objects in the fs and available
> memory.
> 
> But we do have a bug in 1.4.4 that reproduces under certain
> circumstances. It has been fixed post 1.4.4.
> 
> http://oss.oracle.com/git/?p=ocfs2-tools.git;a=commitdiff;h=b987ee22a1b47b9a1e52846a49a5af38f693c7d5

Sunil, thanks for this.

How big will the o2image grow? Can I mail it to you or is it too big? I
don't want to risk a full disk, we had this a bit too often in the last
few days :-(

The fsck.ocfs2 ist not through yet. There was no output from yesterday
15:40 to about 12:15 today. It still runs.

In the log (file size 8324318), I see:

- ---snip---
# cat fsck_sap01.txt
fsck.ocfs2 1.4.3
Checking OCFS2 filesystem in /dev/sdf:
  Label:              SAP01
  UUID:               5969C8CABB854F8EA9C17B5B8DE48EC6
  Number of blocks:   183500800
  Block size:         4096
  Number of clusters: 183500800
  Cluster size:       4096
  Number of slots:    15

** Skipping slot recovery because -n was given. **
/dev/sdf was run with -f, check forced.
Pass 0a: Checking cluster allocation chains
Pass 0b: Checking inode allocation chains
Pass 0c: Checking extent block allocation chains
Pass 1: Checking inodes and blocks.
[INODE_ALLOC_REPAIR] Inode 62 is marked as invalid but its position in
the inode allocator is marked as in use.  Fix the allocation of this and
all future inodes? n
[INODE_SIZE] Inode 161569267 has a size of 69314 but has 8192 bytes of
actual data. Correct the file size? n
Duplicate clusters detected.  Pass 1b will be run
[CLUSTER_ALLOC_BIT] Cluster 328156 is marked in the global cluster
bitmap but it isn't in use.  Clear its bit in the bitmap? n

... and so on, a real lot of lines, until...

[CLUSTER_ALLOC_BIT] Cluster 393693 is marked in the global cluster
bitmap but it isn't in use.  Clear its bit in the bitmap? n
Running additional passes to resolve clusters claimed by more than one
inode...
Pass 1b: Determining ownership of multiply-claimed clusters
Pass 1c: Determining the names of inodes owning multiply-claimed clusters
pass1c: Bad magic number in inode while finding path names in Pass 1c.
The pass will continue, but some inodes may be described by inode number
instead of name.
- ---pins---

# grep -c '[CLUSTER_ALLOC_BIT]' fsck_sap01.txt
65550

OK, there *are* some errors...

Looking at those errors, I do not see cluster-specific errors, only the
usual "marked but not in use" errors every fsck finds ;-)

Best Regards,
Werner
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.15 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org/

iEYEARECAAYFAk1iV6YACgkQk33Krq8b42OQoQCfe/w81LciokhLYeHdBYETJHv0
/HIAnAyUn39xEvHV/bF9NvaSWxx7Tj6D
=0GKz
-----END PGP SIGNATURE-----