[OVSDB-215] East-west traffic (with different network)does not work after a compute node is unstacked and re-stacked Created: 06/Oct/15  Updated: 29/May/18  Resolved: 15/Mar/16

Status: Resolved
Project: ovsdb
Component/s: openstack.net-virt
Affects Version/s: unspecified
Fix Version/s: None

Type: Bug
Reporter: Vasanthan Balasubramaniyan Assignee: Josh Hershberg
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All


Issue Links:
Duplicate
duplicates OVSDB-213 Inter-network East-west traffic does ... Resolved
External issue ID: 4417
Priority: High

 Description   

Build#: Lithium-SR1 Build
Openstack: stable/kilo
ODL config (on etc/custom.properties)
===========
ovsdb.of.version=1.3
ovsdb.l3.fwd.enabled=yes

local.conf snip
=======
enable_plugin networking-odl https://github.com/stackforge/networking-odl stable/kilo
Q_PLUGIN=ml2
ENABLE_TENANT_TUNNELS=True
Q_ML2_TENANT_NETWORK_TYPE=vxlan
PUBLIC_INTERFACE=em3
ODL_MODE=compute
ODL_MGR_IP=<ODL_IP>
ODL_LOCAL_IP=<DATA_NET_IP>
ODL_PROVIDER_MAPPINGS=br-ex:<EXT_INT>
disable_service q-l3
Q_L3_ENABLED=True
ODL_L3=True

Openstack setup
=================
eth0 => Mgmt Network eth1=> Data Network eth3 => External Network
1 openstack controller & 3 compute node

Problem:
============
East-West traffic not working in between subnets in specific restack scenario.

Step to Recreate:
===============
1.Router Gateway contains interface of both 1.1.1.x & 2.1.1.x with 8.1.1.1 External Gateway.
2.Spawn 2 VM's each 1.1.1.x & 2.1.1.x network on both compute nodes.
Compute Node1 => VM1: 1.1.1.3 VM3: 2.1.1.3
Compute Node2 => VM2: 1.1.1.4 VM4: 2.1.1.4
3.Associate Floating IP Address on following way
Compute Node1 => VM1: 1.1.1.3 Floating IP : 8.1.1.3
Compute Node2 => VM2: 1.1.1.4 Floating IP : 8.1.1.4

5.Disassociated floating IP,Deleted all VM's spwaned on compute node-2 and un-stacked

6.Restacked and provisioned new VM(s) with private ip on 2.1.1.x network and associated the floating IP address.
Compute Node2 => VM5: 2.1.1.5 Floating IP Address=> 8.1.1.5

Try following test:
1.ping between 2.1.1.5(compute node 2) to 2.1.1.3 (compute node 1) - successful
2.ping between 2.1.1.5 private ip /8.1.1.5-floating ip to 8.1.1.1 (ext Gateway),8.1.1.3(compute node 1 VM) successful
3.ping between 2.1.1.5 (compute node2) to 1.1.1.3 (compute node1) -Unsuccessful.

3rd Test needs to passed via Private IP of VM as well.Since RouterGateway Interface contains 1.1.1.x Network.

Workaround:
==========
Provision new VM on restacked compute node with 1.1.1.x network,then 1.1.1.x network will be reachable from 2.1.1.5 VM

Associated defect #4408,4346



 Comments   
Comment by Vasanthan Balasubramaniyan [ 01/Feb/16 ]

On Bery-RC1 Build,Openstack-Kilo

While re-stack the compute node and create new VM's on existing network and associate the floating IP address,Unable to communicate via private IP's.However its managed to communicated via Floating IP to other compute node VM's floating IP address.

Until hard-reboot of existing VM's(with existing network),VM cannot communicate with each-other with same private network IP's (spawn across different compute node).

Hence this defect visible,with workaround as "reboot of existing VM"

Generated at Wed Feb 07 20:35:48 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.