[nova/neutron/openvswitch] "No such device" after failed live migrations
Hello all, after a misconfigured setup of a couple of VMs, the live migration of those VMs failed. Now we have some "No such device" listings in openvswitch for those VMs on some compute nodes. (openvswitch-vswitchd)[root@ewos1-com1-prod /]# ovs-vsctl show b5034213-9b15-45f5-8ce0-edcf32d16c57 Manager "ptcp:6640:127.0.0.1" is_connected: true Bridge br-int Controller "tcp:127.0.0.1:6633" is_connected: true fail_mode: secure Port "qvo35487c68-23" tag: 5 Interface "qvo35487c68-23" * Port "qvo5b1aac7e-d4"** ** Interface "qvo5b1aac7e-d4"** ** error: "could not open network device qvo5b1aac7e-d4 (No such device)"** ** Port "qvod9cdff27-9c"** ** Interface "qvod9cdff27-9c"** ** error: "could not open network device qvod9cdff27-9c (No such device)"* Port "qvoed2da602-02" tag: 32 Interface "qvoed2da602-02" Port "qvoa63378f6-d9" for example qvo5b1aac7e-d4 (openvswitch-vswitchd)[root@ewos1-com1-prod /]# ovs-vsctl list Interface ca7e771b-88a0-4cf6-b9be-6be77816baff _uuid : ca7e771b-88a0-4cf6-b9be-6be77816baff admin_state : [] bfd : {} bfd_status : {} cfm_fault : [] cfm_fault_status : [] cfm_flap_count : [] cfm_health : [] cfm_mpid : [] cfm_remote_mpids : [] cfm_remote_opstate : [] duplex : [] error : "could not open network device qvo5b1aac7e-d4 (No such device)" external_ids : {attached-mac="fa:16:3e:4b:f0:4b", iface-id="5b1aac7e-d45b-4a40-8f30-1275b07ffc0b", iface-status=active, vm-uuid="c3fbcc6b-2dfe-4aa1-82bd-522b161a37a9"} ifindex : [] ingress_policing_burst: 0 ingress_policing_rate: 0 lacp_current : [] link_resets : [] link_speed : [] link_state : [] lldp : {} mac : [] mac_in_use : [] mtu : [] mtu_request : [] name : "qvo5b1aac7e-d4" ofport : -1 ofport_request : [] options : {} other_config : {} statistics : {} status : {} type : "" there are no tap devices or bridge devices on the compute node, but still the entry in openvswitch root@computenode5:~# brctl show | grep 5b1aac7e root@computenode5:~# Is that an issue? if yes, should we remove the ports manually or reboot the compute nodes? Could that lead to networking issues? kind regards Nicolas ... --
participants (1)
-
Nicolas Ghirlanda