That is: one should enable at least one VLAN (even unused) at the interface during nodes start-up.
To do so I have created a configuration for a fake VLAN used on `eth0':
[root@node-47 ~]# cat /etc/sysconfig/network-scripts/ifcfg-eth0.50
VLAN=yes
DEVICE=eth0.50
ONBOOT=yes
BOOTPROTO=none
And uploaded that file to all the nodes rebooting them after:
for n in 48 49 50 51; do scp ifcfg-eth0.50 node-$f:/etc/sysconfig/network-scripts; ssh node-$f reboot; done
After the cluster reboot both `network verification' and `Check network connectivity from instance via floating IP' do pass.
I think this information must be included into the release notes.
Well, there is a workaround for that problem.
That is: one should enable at least one VLAN (even unused) at the interface during nodes start-up.
To do so I have created a configuration for a fake VLAN used on `eth0': network- scripts/ ifcfg-eth0. 50
[root@node-47 ~]# cat /etc/sysconfig/
VLAN=yes
DEVICE=eth0.50
ONBOOT=yes
BOOTPROTO=none
And uploaded that file to all the nodes rebooting them after: /etc/sysconfig/ network- scripts; ssh node-$f reboot; done
for n in 48 49 50 51; do scp ifcfg-eth0.50 node-$f:
After the cluster reboot both `network verification' and `Check network connectivity from instance via floating IP' do pass.
I think this information must be included into the release notes.