Hi Aimee, The network settings file is not used by the clean, only by the deploy. It will determine what IPs to configure on your undercloud VM, your overcloud nodes, and indicate which NICs to configure them on. It will also dictate which IPs are assigned to the OVS bridges on your host.
If your clean isn't working, then we should fix that first. Can you ping me tmrw in #opnfv-apex and we can setup a tmate and debug it? Thanks, Tim Rozet Red Hat SDN Team ----- Original Message ----- From: "Aimee Ukasick" <[email protected]> To: [email protected] Sent: Thursday, June 1, 2017 6:06:52 PM Subject: [opnfv-tech-discuss] [Apex] What exactly does Apex do to the network settings... Apex Geniuses, I tried to do a 3-node, bare-metal install in a lab, using a single-nic HP DL-160 as the jumphost. The lab guy installed Centos and configured the networking. The installation failed rather quickly, then opnfv-clean failed.. and long story short, the networking is foobarred something awful. Here's what I did: 1. backed up /etc/sysconfig/network-scripts 2. followed instructions on http://docs.opnfv.org/en/stable-danube/submodules/apex/docs/release/installation/baremetal.html 2-1. used the individual RPMs instead of the release-danube one (the same version of RPM that I used to successfully install a 3-node bare metal at home a couple of days ago on 1 NUC and 2 SuperMicros) 3. sudo opnfv-deploy -n network_settings.yaml -i inventory.yaml -d os-nosdn-nofeature-noha.yaml --debug opnfv-deploy failed pretty quickly, something I thought about network but I can't locate a log file that says why it failed (doesn't help the annoying HP iLO console doesn't work half the time) - should be in var/log/apex right? opnfv-clean failed unable to ping gateway restart network service, eventually rebooted unable to ping gateway "network unreachable" replaced /etc/sysconfig/network-scripts with the originals from before I started the deployment restart network service unable to ping gateway "network unreachable" so reboot - same thing route, netstat always show gateway as 0.0.0.0 gateway is configured in ifcfg-eno1 unable to add default gateway "network unreachable" if config shows 2 active interfaces: eno1 and virbr0 eno1 has correct ip, gateway, but netmask is 255.255.255.255; virbr0 shows no gateway but netmask of 255.255.255.0 I used virsh to delete its default network (the IP in virbr0), stopped libvirtd, restarted networking... you get the idea - nothing has worked Then the server guy got another server guy to spend a few more hours - he doesn't know why the networking isn't working either. So, what exactly does Apex do to the networking files? I looked in opnfv-deploy and opnfv-clean but I'm still not clear. Is the HP networking problem somehow related to the virsh default network? Any ideas? I'm out of my depth, but if you point me in a direction, I'm sure I can figure it out. Any and all suggestions very much appreciated. Thanks in advance! aimee (every developer's worst nightmare) -- Aimee Ukasick Open Source Engagement OPNFV, OpenStack, ONAP IRC: aimeeu _______________________________________________ opnfv-tech-discuss mailing list [email protected] https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss _______________________________________________ opnfv-tech-discuss mailing list [email protected] https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss
