Hide Forgot
Description of problem: Fresh install of Packstack on top of fresh install of Fedora 19. Packstack fails to complete because neutron has problems starting up. See the attached log files. Version-Release number of selected component (if applicable): openstack-neutron-2013.2-2.fc20.noarch (I don't know why "fc20" is there, it's definitely an F19 host). How reproducible: At least once.
Created attachment 833142 [details] openstack-setup.log
Created attachment 833143 [details] 192.168.0.134_neutron.pp.log
Created attachment 833145 [details] 192.168.0.134_keystone.pp.log Strange .. this affects keystone too (even though keystone was working before).
> (I don't know why "fc20" is there, it's definitely an F19 host). RDO Havana for F19 == regular F20 Havana builds really in a convenience repo We'll do that until N and N+1 Fedora keep compatible enough Python runtimes.
Manually starting the zfs-fuse service (ie. service zfs-fuse start) has removed these warnings/errors.
The ZFS related errors can be ignored. That's chatter produced by Puppet that has no operational impact. Looking at 192.168.0.134_neutron.pp.log, I see: Error: Could not start Service[openvswitch]: Execution of '/sbin/service openvswitch start' returned 1: Error: /Stage[main]/Vswitch::Ovs/Service[openvswitch]/ensure: change from stopped to running failed: Could not start Service[openvswitch]: Execution of '/sbin/service openvswitch start' returned 1: ...which causes everything else to fail that has a dependency on Service[openvswitch]. Do you have any logs that indicate why openvswitch failed to start (possibly /var/log/messages)? For Keystone, all the errors are of the form: Error: /Stage[main]/Keystone::Roles::Admin/Keystone_role[_member_]: Could not evaluate: Execution of '/u sr/bin/keystone --endpoint http://127.0.0.1:35357/v2.0/ role-list' returned 1: Service Unavailable (HTTP 503) ...which suggests that the keystone service failed to start. Is there anything of interest in /var/log/keystone/keystone.log?
As I don't have this machine any longer, I'm closing this bug.