IMHO EJB clustering is not the direction we should take, especially considering our IoC plans (see http://gerrit.ovirt.org/#/c/5575/).

QuartzScheduler can be easily fixed by using a DB for Quartz configuration. We had such a discussion a few weeks ago.

mod_jk is probably better for our HA.

Postgres clustering should probably be a different task, as it's more infra and less code.

Another thing to consider is locking, currently done using synchronized, and will need to be changed somehow (probably using Infinispan).

On Aug 22, 2013, at 10:22 AM, Itamar Heim wrote:

On 08/21/2013 10:27 PM, plysan wrote:

2013/8/22 Itamar Heim <iheim@redhat.com <mailto:iheim@redhat.com>>

   On 08/21/2013 11:08 AM, plysan wrote:

       2013/8/21 Laszlo Hornyak <lhornyak@redhat.com

       <mailto:lhornyak@redhat.com> <mailto:lhornyak@redhat.com

       <mailto:lhornyak@redhat.com>>>

            Hi,

            ----- Original Message -----

             > From: "plysan" <plysab@gmail.com

       <mailto:plysab@gmail.com> <mailto:plysab@gmail.com

       <mailto:plysab@gmail.com>>>

             > To: engine-devel@ovirt.org

       <mailto:engine-devel@ovirt.org> <mailto:engine-devel@ovirt.org

       <mailto:engine-devel@ovirt.org>__>

             > Sent: Wednesday, August 21, 2013 3:49:31 PM

             > Subject: [Engine-devel] Ideas about engine clustering?

             >

             > Hello, guys:

             >

             > I'm trying to make a jboss clustering environment for

       engine. But

            felt the

             > difficuty too.

             > So I want to know if any one is working on this? I

       googled it and

            haven't

             > found anything.

             >

             > I know there will be lots of things to do for engine

       migrating to

            clustering

             > environment. I just want to know what problem need to be

       solved?

             > I would like to list the things I think that is needed

       to do(or

            solve):

             > 1. run jboss in standalone-ha mode or domain mode.

             > 2. add @Clustered annotation to each EJB, I think this

       will solve the

             > replication problem in cluster, probably lots of details

       in it.

             > 3. QuartzScheduler problem, only one node running a

       scheduler is

            enough at a

             > time instead of each node running one.

             > 4. use mod_cluster to handle load balancing.

             > 5. postgresql clustering.

            For failover it sounds like a plan, just wondering if this

       could

            solve load balancing problems as well.

            Ovirt is generating a big load on the DB and I would be really

            interested if SQL DB clustering could solve the issues.

            It would be great if it could, big part of the scalablity

       issue is

            the evil things we did against that poor database.

             > 6. There should be more, but hope anyone can think about

       it :D

            There are tons of data structures that are used in oVirt

       that store

            state. You will have to hunt all of these down and replace

       with a

            data structure that is shared between the cluster memebers.

       This may

            be a big lot of work from many parts of the application.

       Indeed, and like Yair Zaslavsky said, we can put the data

       structure in

       infinispan cache, but clustered EJB can do the same thing for

       us, right?

       If all those state values are inside EJBs. (iirc)**

             >

             > Any ideas?

             >

             > Thanks!

             >

             > _________________________________________________

             > Engine-devel mailing list

             > Engine-devel@ovirt.org <mailto:Engine-devel@ovirt.org>

       <mailto:Engine-devel@ovirt.org <mailto:Engine-devel@ovirt.org>__>

             > http://lists.ovirt.org/__mailman/listinfo/engine-devel

       <http://lists.ovirt.org/mailman/listinfo/engine-devel>

             >

       _________________________________________________

       Engine-devel mailing list

       Engine-devel@ovirt.org <mailto:Engine-devel@ovirt.org>

       http://lists.ovirt.org/__mailman/listinfo/engine-devel

       <http://lists.ovirt.org/mailman/listinfo/engine-devel>

   unless you need the load balancing, just for HA, hosted-engine

   should cover your needs?

I think failover and more scalability  is what i am interested here.

Just thinking about the possibilities :)

Imagine engine would never get restarted from outside view. Cool isn't it?

active/passive failover you should get from hosted-engine (yes, with some downtime)
scalability and active-active would be great. iirc, juan looked at implications of doing this a while back
_______________________________________________
Engine-devel mailing list
Engine-devel@ovirt.org
http://lists.ovirt.org/mailman/listinfo/engine-devel