[Starlingx-discuss] Correct way of configuring bondings / tagged vlans

[bnovickovs at weecodelab.com]

We were trying to deploy STX 4.0.1.

Installation: standard controllers with dedicated storage nodes.

Followed the guide: 
https://docs.starlingx.io/deploy_install_guides/r4_release/bare_metal/dedicated_storage_install_kubernetes.html

Hardware: 2 NIC, 500GB on controllers, 64GB RAM

pxeboot is using default vlan cause it doesnt support tagging.

pxeboot_subnet: 192.168.20.0/24
pxeboot_start_address: 192.168.20.11
pxeboot_end_address: 192.168.20.100

management_subnet: 192.168.30.0/24
management_start_address: 192.168.30.11
management_end_address: 192.168.30.100

cluster_host_subnet: 192.168.40.0/24
cluster_host_start_address: 192.168.40.11
cluster_host_end_address: 192.168.40.100

external_oam_subnet: 192.168.10.0/24
external_oam_gateway_address: 192.168.10.250
external_oam_floating_address: 192.168.10.101
external_oam_start_address: 192.168.10.12
external_oam_end_address: 192.168.10.100
external_oam_node_0_address: 192.168.10.11
external_oam_node_1_address: 192.168.10.12

controller-0

system host-if-add controller-0 -m 1500 -a active_standby bond0 ae eno1 
eno2
system host-if-add controller-0 -V 30 -c platform mgmt vlan bond0
system host-if-add controller-0 -V 40 -c platform clusterhost vlan bond0
system host-if-add controller-0 -V 10 -c platform oam vlan bond0
system interface-network-assign controller-0 oam oam
system interface-network-assign controller-0 mgmt mgmt
system interface-network-assign controller-0 clusterhost cluster-host
system interface-network-assign controller-0 bond0 pxeboot
system ntp-modify ntpservers=0.pool.ntp.org,1.pool.ntp.org
system storage-backend-add ceph --confirmed
system host-label-assign controller-0 openstack-control-plane=enabled
system host-unlock controller-0

system host-update 2 personality=controller install_output=text 
console=ttyS0

controller-1
system host-if-add controller-1 -m 1500 -a active_standby bond0 ae eno1 
eno2
system host-if-add controller-1 -V 30 -c platform mgmt vlan bond0
system host-if-add controller-1 -V 40 -c platform clusterhost vlan bond0
system host-if-add controller-1 -V 10 -c platform oam vlan bond0
system interface-network-assign controller-1 oam oam
system interface-network-assign controller-1 mgmt mgmt
system interface-network-assign controller-1 clusterhost cluster-host
system interface-network-assign controller-1 bond0 pxeboot

Behavior: Controller-1 constantly rebooting trying to auto-recover 
itself for no particular reason...There are alarms for vlan 30 and vlan 
40 "failure detected with peer over port" when both controllers are 
unlocked and enabled, kube cluster is healthy. Even suppressing those, 
does not help to stop reboots. Screenshot of alarms - 
https://prnt.sc/vo0cl0

Please suggest what else can be done / provided logs to troubleshoot the 
issue. Thank you