Menu
▾
▴
Re: [OSR-users] Problem with bond configuration
|
From: Marc G. <gr...@at...> - 2009-05-26 14:29:50
|
Hi Stefano, sorry for the late response. On Wednesday 13 May 2009 10:51:04 Stefano Elmopi wrote: > Hi Marc, > > very very thank for your answer.......... I was becoming crazy !!! > I tried and now it works. > I take advantage of this email to ask for other informations. > > - It's possible configure two network interfaces of the server in > order to > have two interfaces Bond, for example bond0.11 and bond0.22, > one for communication intra-cluster and the other for the service > configured on the cluster ?? yes. http://www.open-sharedroot.org/documentation/administrators-handbook/part-viii-cluster-administration/ That should be self explanatory. VLANs and no vlans are described. If you want to use vlans you also need to install the rpm comoonics-bootimage-extras-network and build a new initrd. If not vconfig is not added to the initrd. > > > > - The ip address configured for the service given by the cluster, for > my files cluster.conf: > > <resources> > <ip address="10.43.100.204" monitor_link="1"/> > <script file="/etc/init.d/httpd" name="httpd"/> > </resources> > <service autostart="0" domain="failover" name="HTTPD"> > <ip ref="10.43.100.204"/> > <script ref="httpd"/> > </service> > > I have to configure on a node of the cluster > > <clusternode name="clu01" votes="1" nodeid="1"> > <com_info> > <syslog name="clu01"/> > <rootvolume name="/dev/sda2" fstype="ocfs2"/> > <eth name="eth0" ip="10.43.100.203" mac="00:15:60:56:75:FD"/> > <eth name="eth1" ip="10.43.100.204" mac="00:15:60:56:75:FC"/> > </com_info> > </clusternode> > > If I want to relocate the service on the node_2 when the node_1 fails, > I must also configure the ip (10.43.100.204) in the section <com_info> > of the node_2 ?? No the ip to be failovered needs to be exclusively setup in the resource section. > > > - It's possible to configure a service, for example httpd, > to be active simultaneously on multiple nodes instead > of having it running on one node and move it when the node falls ?? Yes. But you will need a loadbalancer "in front" or "integrated in the cluster". -- Gruss / Regards, Marc Grimme http://www.atix.de/ http://www.open-sharedroot.org/ |
