Ask Your Question
0

OpenStack on OpenStack deployment

asked 2015-05-26 06:52:39 -0600

Praveen N gravatar image

updated 2015-05-27 00:43:37 -0600

Am working on OpenStack on OpenStack deployment. On Baremetal machine(16GB RAM, 500 GB Hard Drive, i3 processor) with Centos 7 Genome Desktop, OpenStack-packstack(juno) is deployed and provisioned 2 Instances(aio and compute) using the image CentOS-7-x86_64-GenericCloud-20140917_02.qcow2,each VM is with 2 eth interfaces,and the network topology is shown in the diagram. C:\fakepath\Network_Topology_aio_Packstack.JPG

On these OpenStack instances aio and compute i deployed OpenStack(juno release) manually(multi node setup:1 aio and 1 compute).Then provisioned 2VMs (test-aio and test-compute) on aio host and compute host respectively using nova boot command. But there is no communication between the VM on aio host to VM on compute host even though the VMs are in active state and assigned with fixed IP. All other openstack services are up & active.Also am able to assign floating IPs to these VMs. Only am able to ping and do ssh login to VM launched on aio host from aio host but not to the compute host VM. And test-aio to test-compute ping is not working.

Observations made from logs in nova-compute, nova-scheduler,neutron-server, openvswitch are quoted below just for information (WARNING and ERROR here are single lines picked up from logs and listed below which seemed to be unusual and may be source of clue for fixing the pinging issue between VMs).

WARNING nova.compute.manager [-] Bandwidth usage not supported by hypervisor.

WARNING oslo.messaging._drivers.amqpdriver [req-68821b3a-7efc-4a00-b678-96cb74a94fe4 ] No calling threads waiting for msg_id 

WARNING neutron.openstack.common.loopingcall [-] task run outlasted 

neutron.openstack.common.loopingcall

ERROR oslo.messaging._drivers.impl_rabbit [-] Failed to consume message from queue: [Errno 104] Connection reset by peer

ERROR neutron.agent.linux.utils

Also i noticed the br-tun entering to promiscuous mode as observerd in aio instance console as depicted in screen capture.C:\fakepath\aio_promiscuous.png

Looking for inputs to fix this pinging issue between VMs .

edit retag flag offensive close merge delete

Comments

I would try AIO setup packstack , then update generated answer-file with EXCLUDE_SERVERS=AIO_IP and ran again to add Compute Node. It worked hundreds of times.Specifically in initial phase of testing RDO Kilo ( I mean RC2 Devs release)

dbaxps gravatar imagedbaxps ( 2015-05-27 01:10:25 -0600 )edit
dbaxps gravatar imagedbaxps ( 2015-05-27 01:11:52 -0600 )edit

Attaching two private nets ton same neutron-router as I could guess is supposed to provide routing between nets ? Am I correct ?

dbaxps gravatar imagedbaxps ( 2015-05-27 01:22:38 -0600 )edit

Hi dbaxps Thanks for the quick reply!!..Yes 2 private nets attached to a same router providing routing between nets.

Praveen N gravatar imagePraveen N ( 2015-05-27 05:02:23 -0600 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2015-08-07 07:15:44 -0600

Praveen N gravatar image

updated 2015-08-07 07:17:15 -0600

The Work around which provided the quick fix is: The br-ex was not assigned with the IP during the process of deployment. The IP for br-ex was assigned manually on Controller node and Compute node with the command,

ifconfig br-ex  y.y.y.y netmask x.x.x.x

Then VMs were able to ping and communicate to each other.

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: 2015-05-26 06:52:39 -0600

Seen: 726 times

Last updated: Aug 07 '15