[one-users] Fault Tolerance

[fc lists]

Hi,

1) The documentation page about fault taulerance is quite explicative about
the point.

http://opennebula.org/documentation:rel3.4:ftguide

You can use the internal one hook system to deal with different kind of
failure.

Also there is a plugin on the dev communit site to deal with UNKWON states

http://dev.opennebula.org/issues/1144#change-2930

Note that i never tested either of them. I always been scared of this kind
of solutions ... Too many different variables in "my kind" of failures to
let a script deal with it. Guess it really depends on what run on your VM.

2) This can be easily achieved with an external monitoring and restart
handler for the "oned" part of the management node ... nagios would do
that.
Mysql redundancy is your choise ... multimaster replication or mysql
cluster, wich is also documented on the web

site
https://support.opennebula.pro/entries/20400286-opennebula-with-mysql-cluster-for-high-availability

The beauty of opennebula , and the reason why i choosed it over the others,
is the simplicity of its components that makes indeed so easy to manage and
adapt.

cheers
fc

On Thu, May 17, 2012 at 10:02 AM, Yount, William D <
Yount.William at menloworldwide.com> wrote:

> We have been working on implementing a cloud at work. What we want is
> basically a fault tolerant hypervisor cluster. I tried CloudStack, however
> it offered only HA but not fault tolerance. If a node goes out, then a
> human has to migrate the VM over to a new node.****
>
> ** **
>
> **1)      **One feature that has caught my eye with OpenNebula is that it
> claims to offer fault tolerance. When I hear fault tolerance, I envision
> that if the node a VM is running on goes out, that the management software
> will migrate the VM over to another node and start it back up. Without this
> feature, a simple hypervisor cluster meets our need.****
>
> ** **
>
> **2)      **Does OpenNebula provide for automatic failover of the
> management server? In CloudStack, there was no built in replication of the
> management server. If your management server went down you had to start up
> another one. I had to create two management servers and then use MySQL
> replication, KeepAlived and some other tools to get automatic failover. **
> **
>
> ** **
>
> I am just looking for as much redundancy as I can get. With the
> management/storage/node model, I would like to have as much redundancy as
> possible. ****
>
> ** **
>
> ** **
>
> ** **
>
> Thanks,****
>
> *William Yount *****
>
> ** **
>
> _______________________________________________
> Users mailing list
> Users at lists.opennebula.org
> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20120517/b79b1d59/attachment-0003.htm>