[CLOUDSTACK-6739] [OVS] Migrating vm within the cluster deletes the existing flows from the ovs bridge

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Duplicate
• Affects Version/s: 4.4.0
• Fix Version/s: 4.4.0
• Component/s: Management Server, Network Controller
• Security Level: Public (Anyone can view this level - this is the default.)
• Labels: ovs
• Environment: Latest build from 4.4 with commit e6961fd21bb6d793302c234d0f409f66dc498072
• Assignee: Murali Reddy
• Reporter: Sanjeev N
• Created: 21/May/14 11:13
• Updated: 17/Jun/14 09:20
• Resolved: 17/Jun/14 09:20

Description

[OVS] Migrating vm within the cluster deletes the existing flows from the ovs bridge

Steps to reproduce:
================
1.Bring up CS in advanced zone with a xen cluster with two hosts
2.Create physical network with GRE isolation
3.Create an isolated network offering with virtual networking service and OVS as the service provider.
4.Deploy two vms in the network and make sure that all the vms and VR are on host1 (So that no tunnel port will be created at this time)
5.Verify ovs flow-table entries for the bridge created at this stage on host1
6.Migrate one of the guest vms to host2 (So that tunnel ports will be created on both the hosts)
7.Verify flow-table entries on both the hosts

Actual Result:
===========
After vm migration on host1 only default entry is present in the bridge flow table where as on host2 flow entries are properly created for the bridge

Impact:
======
With flow entries bieng in this state if we deploy addition vm and if it comes up on host2 , the vm would not get the ip address

Attaching MS log file and ovstunnel log files from both the hosts.
Please look for job-158 in the MS log and xapi7 in ovstunnel for flow related info.

Issue Links

• Is contained by: CLOUDSTACK-6779 [OVS] Expunging VM (deleting vif) deletes all the rules from ovs bridge flow table