Re: JBoss Clustering Compatibility

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





Lon Hohberger wrote:
On Mon, 2006-08-21 at 14:13 -0400, Sutterfield, Geary L. wrote:
Hello all,
Does anyone have any information or experience about using GFS and the
Cluster Suite and JBoss Clustering on the same cluster? I've been
looking for a couple of days, but I can't seem to find any information
on that configuration. Our goal is to create a J2EE cluster with a
shared, global filesystem. GFS seemed like a logical choice. I'm speculating that if Cluster Suite was only used for GFS, and not
for any other services, then it might work okay with JBoss Clustering.
Cluster Suite would handle the GFS clustering, and JBoss Clustering the
J2EE app server clustering. But I also wonder about conflicts or
incompatibilities. And I especially wonder how fencing would impact
JBoss. Has anyone tried this configuration, or heard of anyone who has?

I have not tried it, but here are some of the 'gotchas':

(a) Disagreements in the JBoss cluster view vs. the GFS cluster view.
GFS cluster view should always win in the end, because it has fencing.
So, the GFS/RHCS cluster should be configured to converge on membership
changes before JBoss (whether that means increase deadnode_timeout, or
tweaking JBoss is up to you).

(b) Cluster quorum.  I don't know much about JBoss clustering (yet), but
remember that cluster suite / GFS need a majority of votes (normally,
there's a 1:1 correlation between nodes and votes) available in order to
form a quorum.  Without a quorum, access to GFS will be halted.

(c) Fencing.  You will probably need to use power-based fencing, because
storage/SAN level fencing will not drop a node out of the JBoss cluster
by itself.  Since I don't currently know much about JBoss clustering
generally, I also don't know the specific implications of trying to run
JBoss on a given node after the GFS volumes have been cut off from that
node.  I *assume* that JBoss would take that node out of the JBoss
cluster.


If there any problems, I suppose they will be worked out eventually,
since Red Hat has acquired JBoss. But we don't want to waste a lot of
time going down a path with known problems. Any thoughts or feedback
would be appreciated.

-- Lon

What is the process for initiating/requesting a white paper to be written about a particular topic? JBoss on Red Hat Clustering with GFS might make for a very good topic for a white paper

Riaan

begin:vcard
fn:Riaan van Niekerk
n:van Niekerk;Riaan
org:Obsidian Systems;Obsidian Red Hat Consulting
email;internet:riaan@xxxxxxxxxxxxxx
title:Systems Architect
tel;work:+27 11 792 6500
tel;fax:+27 11 792 6522
tel;cell:+27 82 921 8768
x-mozilla-html:FALSE
url:http://www.obsidian.co.za
version:2.1
end:vcard

--
Linux-cluster mailing list
Linux-cluster@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/linux-cluster

[Index of Archives]     [Corosync Cluster Engine]     [GFS]     [Linux Virtualization]     [Centos Virtualization]     [Centos]     [Linux RAID]     [Fedora Users]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite Camping]

  Powered by Linux