[Linux-cluster] which is better gfs2 and ocfs2?

Bob Peterson rpeterso at redhat.com
Sat Mar 12 23:13:05 UTC 2011

----- Original Message -----
| Those users are paying for GFS installations. They have every right to
| criticize its shockingly poor performance for these operations,
| especially when it adversely impacts their ability to get work done.


Agreed.  We're abundantly aware of the performance problems,
and we're not ignoring them.  People, please bear in mind that
Red Hat is also working diligently to improve all aspects of
gfs2 performance, and we've made great strides.  Cases in point:

(1) We recently found and fixed a problem that caused the
    dlm to pass locking traffic much slower than possible.
(2) We recently increased the speed and accuracy of fsck.gfs2
    quite a bit.
(3) We also recently developed a patch that improves GFS2's
    management of cluster locks by making hold times self-tuning.
    This makes gfs2 perform much faster in many situations.
(4) We've recently developed another performance patch that
    sped up clustered deletes (unlinks) as much as 25%.
(5) We recently identified and fixed a performance problem
    related to writing large files that sped things up considerably.

These patches are in various stages of development, and most or all
have already been posted to the public cluster-devel mailing list,
of various records in bugzilla, which means they're making their
way to a kernel (or user-space) near you.

Our work continues; we're improving it every day and have more
performance improvements planned.  I don't know about ocfs2,
but there's a whole team of people at Red Hat plus the open
source community at large working to improve gfs2.


Bob Peterson
Red Hat File Systems

More information about the Linux-cluster mailing list