I want to consider our current cluster - 3 nodes for implementing VLAN while it is live.
This is a critical cluster contain PROD environment. I want to do this without impact or down time.
current the interface file looks like this:
auto lo
iface lo inet loopback
iface enp3s0 inet manual
pre-up /sbin/ethtool -s enp3s0 speed 1000 duplex full
iface enp4s0 inet manual
pre-up /sbin/ethtool -s enp4s0 speed 1000 duplex full
auto bond0
iface bond0 inet manual
slaves enp3s0 enp4s0
bond_mode 1
bond_miimon 100
auto vmbr0
iface vmbr0 inet static
address xxx.xxx.29.1
netmask 255.255.252.0
gateway xxx.xxx.28.254
bridge_ports bond0
bridge_stp off
bridge_fd 0
just for example sake, I have 2 VLANs 100 and 200 id.
VLAN 100 - 192.168.50.0/24
VLAN 200 - 172.200.10.0/22
What should the interface file look like after the changes are ready.
I know we should move all the VM's off the node first. and split them between the last 2 nodes.
Then repeat the process for the other 2 nodes.
If I could NOT move an VM's around other than adding new nodes to this which isn't possible at this time.
What other alternatives would I have other than downtime.
Phillip
This is a critical cluster contain PROD environment. I want to do this without impact or down time.
current the interface file looks like this:
auto lo
iface lo inet loopback
iface enp3s0 inet manual
pre-up /sbin/ethtool -s enp3s0 speed 1000 duplex full
iface enp4s0 inet manual
pre-up /sbin/ethtool -s enp4s0 speed 1000 duplex full
auto bond0
iface bond0 inet manual
slaves enp3s0 enp4s0
bond_mode 1
bond_miimon 100
auto vmbr0
iface vmbr0 inet static
address xxx.xxx.29.1
netmask 255.255.252.0
gateway xxx.xxx.28.254
bridge_ports bond0
bridge_stp off
bridge_fd 0
just for example sake, I have 2 VLANs 100 and 200 id.
VLAN 100 - 192.168.50.0/24
VLAN 200 - 172.200.10.0/22
What should the interface file look like after the changes are ready.
I know we should move all the VM's off the node first. and split them between the last 2 nodes.
Then repeat the process for the other 2 nodes.
If I could NOT move an VM's around other than adding new nodes to this which isn't possible at this time.
What other alternatives would I have other than downtime.
Phillip