[Linux-cluster] Fsck on GFS2

Bob Peterson rpeterso at redhat.com
Fri Feb 25 13:43:33 UTC 2011

----- Original Message -----
| Hi !
| We are in the testing phase of our setup, and I had to grow a GFS2
| filesystem under RH6, wich was full. It didn't work because of a bug,
| so I
| deleted some files to make room, and the the grow worked, but still I
| couldn't write more data. So I took the FS offline and made a FSCK on
| it.
| And it took something like 2 and a half hours to complete. The FS is
| now 941
| gigs on on a 1.8 tb lvm on a shared raid rack (active storage), via a
| 4 gb
| fibre channel switch. It contains millions (I don't have an exact
| count, but
| 3 might be a good estimate) files.
| So is it normal for fsck on a FS that size to took so long? On our
| curent
| XSan setup, xserve-raid, older drives and 2 gb fc, for the same size
| and
| file count it takes about 40 or 50 minutes to complete.


The speed and accuracy of fsck.gfs2 varies based on a lot of things,
including the amount of metadata, if there is any damage to fix,
and which version of fsck.gfs2 you are using.  The tool is improving
all the time.

I try to put the latest fixes for x86_64 on my people page:


All of the fixes going into RHEL5.7 are in that version, and
it is faster and more accurate than the version shipped with RHEL5.6.
In fact we've made about 75+ patches since RHEL5.6 to improve it.
The upstream git repos for gfs2-utils and cluster have the latest
source code as well.


Bob Peterson
Red Hat File Systems

More information about the Linux-cluster mailing list