Ask Your Question
0

HAVANA multi-node with nova-network

asked 2014-04-07 14:36:15 -0500

nicolarb gravatar image

updated 2014-04-15 06:26:46 -0500

Hello, for a POC we installed RDO havana release with packstack all-in-one option on a RHEL6.4 with nova-network. We choose to not use NEUTRON because ma skills in networking are not very High. So all works fine with this configuration

  • i can launch an instance
  • ping the IP allocated from my controller/compute node (all-in-one seetings)
  • add floating IP which is known in the VLAN of theserver which hosts RDO (so have an external connectivity with the VM.)

To add a second compute node i editrd the "answer file" generated during the initial packstack setup. an run it again. Currently i have one server (the first on which is the the controller and a compute node at once) and a second which is only a compute node. On the first nothing changed (launch, ping , external connectivity) On the second compute node i can launch an instance


but i am not enabled to ping the allocated IP from the controler/compute node :


[root@lhat03 ~(keystone_admin)]# ping 192.168.32.3
PING 192.168.32.3 (192.168.32.3) 56(84) bytes of data.
From 192.168.32.1 icmp_seq=2 Destination Host Unreachable
From 192.168.32.1 icmp_seq=3 Destination Host Unreachable

My two servers have 2 NICS and a bridge is configured on eth1 as below

ifcfg-eth0

ONBOOT=yes
DEVICE=eth0
BOOTPROTO=static
IPADDR=10.60.xx.xxx
NETMASK=255.255.254.0
BROADCAST=10.60.xx.255
BONDING_OPTS="mode=4 miimon=100"

ifcfg-eth1

DEVICE=eth1
ONBOOT=yes
NM_CONTROLLED=no
PROMISC=yes
BRIDGE=br-ex
NAME="Static Bridged eth1"
BOOTPROTO=none

ifcfg-br-ex (added manually)

DEVICE=br-ex
DEVICETYPE=ovs
TYPE=OVSBridge
BOOTPROTO=static
IPADDR=
NETMASK=
ONBOOT=yes

Same thing for the second compute node with another iP for eth0

Please someone could help me to resolve this issue? If you need more information, please let me know

Thanks

edit retag flag offensive close merge delete

Comments

What specific changes did you make to the answer file?

larsks gravatar imagelarsks ( 2014-04-07 16:02:34 -0500 )edit

Hi,

the sepecific change i made are :

  #add the IP of the second compute node
    CONFIG_NOVA_COMPUTE_HOSTS=10.60.94.244,10.60.94.253

    #change CONFIG_NOVA_COMPUTE_PRIVIF=value from lo to eth
    CONFIG_NOVA_COMPUTE_PRIVIF=eth1

    # change CONFIG_NOVA_NETWORK_HOST by CONFIG_NOVA_NETWORK_HOSTS and provide corresponding IP of the 2 compute node
    CONFIG_NOVA_NETWORK_HOSTS=10.60.94.244,10.60.94.253

After i ran ckstack packstack --answer-file=packstack-answers.txt( instead of --allinone --os-neutron-install=n i used the first time)

In fact i followed the tutorial on http://openstack.redhat.com/Adding_a_compute_node (http://openstack.redhat.com/Adding_a_...) , but physical requirements for NICS are not clear for me. Currently i have 2 NICs on each sever, and they are all physicaly connected on the same VLAN. Is it good or not ?

Regards

nicolarb gravatar imagenicolarb ( 2014-04-08 13:29:07 -0500 )edit

Thanks for the update. It's better to add this sort of thing to your original question rather than leaving it as a comment because this makes it easier for people to find (and keeps the comments from growing so long as to be unreadable).

larsks gravatar imagelarsks ( 2014-04-10 07:59:20 -0500 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2014-04-15 06:25:48 -0500

nicolarb gravatar image

Finnaly i found the cause of this issue. I used the default network created with packstack and it is not multi_host. So i delete and create it again with the good option

nova network-create novanetwork --fixed-range-v4=192.168.40/24  --bridge=br100 --multi-host=T

and all works fine

edit flag offensive delete link more

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2014-04-07 14:36:15 -0500

Seen: 546 times

Last updated: Apr 15 '14