Created attachment 1168748 [details] ironic inspector log Description of problem: I tried an OpenStack deployment but I cannot register nodes. During the introspection process, the following line appear in openstack-ironic-inspector log units : Jun 16 07:48:21 undercloud.example.com ironic-inspector[654]: 2016-06-16 07:48:21.811 654 WARNING ironic_inspector.plugins.standard [-] The following interfaces were invalid or not eligible in introspection data for node with BMC and were excluded: {u'eth0': {u'ip': u'192.168.150.127', u'mac': u'52:54:00:44:1c:d0'}} It then try to use eth1 but it doesn't work as this interface is not in the provisioning network. My environment is based on KVM (standard ISO install from the git bootiso repo). My OSP nodes have two network interfaces, eth0 in the provisioning net, eth1 in the public net. I have search about this error message and the only reference I came across is this bug (which doesn't seems related) : https://bugzilla.redhat.com/show_bug.cgi?id=1308714 Version-Release number of selected component (if applicable): QCI-1.2-RHEL-7-20160613.t.0-QCI-x86_64-dvd1.iso QCIOOO-8.0-RHEL-7-20160527.n.0-QCIOOO-x86_64-dvd1.iso How reproducible: Always. Steps to Reproduce: 1. Install QCI and QCIOOO 2. Start two OSP nodes VM 3. Start an OSP deployment 4. Try to register nodes Actual results: Fail with error message above. Expected results: Nodes register correctly. Additional info:
Created attachment 1168749 [details] messages log from undercloud
Created attachment 1168750 [details] foreman production log
What OS is your hypervisor running? If RHEL 7 this is likely a known bug with ipxe-qemu(-roms) that is now mostly fixed. The one place it remains is for VM's running on a RHEL 7 hypervisor. https://bugzilla.redhat.com/show_bug.cgi?id=1302420 If this is the case can you try using an updated ipxe-roms-qemu package or following the instructions in the last comment.
My base host was RHEL 7.2 and updating the IPXE roms effectively solve the problem. I now use Fedora 24 as my base host and the problem is not there too.
*** This bug has been marked as a duplicate of bug 1302420 ***