OpenStack Fuel 6.0 Community - problem with HA deployment (probably floating IP issue)

asked 2015-01-22 04:36:34 -0600

anonymous user



I'm currently trying to get in touch with OpenStack so I'm deploying test enviroments to get the general feel of it. The enviroment is set as QEMU and Ubuntu.

I'm using a three physical servers, one running KVM VMs with Fuel Master, three controller nodes, two compute nodes, one telemetry node and one monitoring node, and two physical servers for CEPH storage.

VMs are using network bridges to connect with storage servers, two virtual network interfaces per VM, one for PXE and a second for public, management and storage. Storage servers are connected using two physical interfaces each via 1Gbps switch.

As for networking it's Neutron with VLAN segmentation, all VLANs are set on corresponding switchports and it seems that everything about networking is OK, at least according to Fuel network checking tool.

There is no problem with non-HA deployment, it works brilliant, but HA deployment fails every single time.

Telemetry and monitoring nodes are set up first, but when it comes to controller nodes deployment fails on activating pacemaker "vip__public" service.

If I log into a controller node during deployment there's no br-ex-hapr interface with floating ip.

I'm using default network configuration, with PXE and public networks untagged.

If you need additional informations please let me know.

Kind regards,

edit retag flag offensive close merge delete