[Pacemaker] Does this config make sense? (OpenVZ, DRBD, Corosync, Pacemaker, OCFS2) Active/Active/Pasive

Wessel Louwris wessel at techtribe.nl
Thu Jun 17 09:41:09 UTC 2010


Hello,

I tried openvz and HA containers also, but the problem seems to be
ClusterIP. HA uses this to decide which network packet is going to which
servers as far as I understood (http://www.linux-ha.org/ClusterIP). 

You could change the IPAddr scripts of HA on your containers so that
they will pass some iptables command's to the host system (I found some
scripts on this list or somewhere else to do that) but since the OpenVZ
kernel does not support ipt_CLUSTERIP this did not help me.

But any feedback on this is welcome, it could still be an interesting
combination I think.

greetings, Wessel


On 17-6-2010 11:20, marc genou wrote:
> I am testing some configs and I am not really sure if this goal can be
> achieved.
>
> As main target I would like to deploy a HA containers (OpenVZ) system
> which should be able to run this containers, performing failover in a
> transparent way for the final admin/user of this container.
> I mean, not to be noticed about a failed container and be able to keep
> working transparently on the "second" active.
>
> Is this possible? 
>
>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
>   


-- 
Wessel Louwris, software developer/system admin
TECHTRIBE
Prinshendriklaan 29, 1075 AZ  Amsterdam 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20100617/fde17af4/attachment-0002.htm>


More information about the Pacemaker mailing list