Mailinglist Archive: opensuse-ha (30 mails)

< Previous Next >
Re: [opensuse-ha] ext4 failure on cluster
  • From: Goldwyn Rodrigues <rgoldwyn@xxxxxxx>
  • Date: Thu, 14 Aug 2014 06:37:33 -0500
  • Message-id: <53EC9F7D.6040403@suse.de>
Hi Steve,

On 08/14/2014 04:52 AM, steve wrote:
On Thu, 2014-08-14 at 11:02 +0200, Richard Brown wrote:
On Thu, 2014-08-14 at 10:08 +0200, steve wrote:
[post to both lists]
Hi everyone
Next month we hope to go into production with our drbd-ocfs2-ctdb
cluster. Of course, we must prove to our boss that it is a fair if not
better replacement for our single AD domain file-server which it will
replace.

A few questions remain to which we cannot find a simple answer that a
non technical person would understand. Indeed, much of what we've read
we cannot understand ourselves as self styled ha setter-uppers:
1. Where our single ext4 file server is predictable under load (it just
gets slower), with both nodes up, why does the cluster fail so badly
under ext4 but absolutely screams under ocfs2? The strange thing is that
the clustered ext4 actually performs better when only one node is up.

I'm unable to comment without a better understanding of how your
drbd/ext4 setup is?

Hi
Sorry to be unclear. Our existing file server uses ext4. For our 2 node
cluster we had to change to ocfs2:
http://linuxcostablanca.blogspot.com.es/2014/07/samba4-cluster-for-ad-drbd-ocfs2-ctdb.html
I suppose the question is, what does ocfs2 have that ext4 doesn't? But
in really simple terms.


ocfs2 is a clustered filesystem wheras ext4 is a local filesystem. ocfs2 can be mounted simultaneously on both nodes without any problems of data integrity and both nodes can simultaneously serve.
ext4 being a local filesystem must be mounted on a single node at a time, which means only one node can serve at a time.

--
Goldwyn
--
To unsubscribe, e-mail: opensuse-ha+unsubscribe@xxxxxxxxxxxx
To contact the owner, e-mail: opensuse-ha+owner@xxxxxxxxxxxx

< Previous Next >
List Navigation