https://jira.opendaylight.org This file is an XML representation of an issue en-us 8.20.10 820010 22-06-2022 https://jira.opendaylight.org/browse/NETVIRT-966 netvirt <p>We found the following issue when running test with ODL Carbon SR2.<br/> In conntrack mode, SNAT communication fails if route gateway is set before virtual NW  creation or VM creation in a fresh environment.<br/> The following WARN karaf log message shows when configuring router gateway and interface.<br/>    2017-10-24 10:09:47,079 | WARN | nPool-1-worker-0 | ExternalNetworksChangeListener | 369<br/>    org.opendaylight.netvirt.natservice-impl - 0.4.2.Carbon | associateExternalNetworkWithVPN :<br/>    primary napt Switch not found for router Uuid <span class="error">&#91;_value=e81d29f5-19cb-412b-aa7e-4c9377e00ee4&#93;</span> on dpn: 0<br/>    2017-10-24 10:11:10,857 | WARN | eChangeHandler-0 | AbstractSnatService | 369<br/>    org.opendaylight.netvirt.natservice-impl - 0.4.2.Carbon | getTunnelInterfaceName :<br/>    RPC Call to getTunnelInterfaceId returned with Errors []</p> <p>○Possible cause<br/> 　By checking both OVS and ODL logs, we think that the cause might be as follows<br/> 　　It seems that in controller node the patch ports of br-int br-ex are created when creating the first virtual NW.<br/> 　　And in compute node, the patch ports are created when creating the first VW in this node.<br/> 　　The processes are not finished correctly when router gateway is set before patch ports are created on each node.</p> <p>○Reproduction steps:<br/> 1. neutron-server and related agent restart (controller node)<br/> 2. OVS initialization (controller node and compute node)<br/> 3. ODL initialization (all 3 odl nodes）<br/> 4. ODL 3 node cluster start<br/> 5. OVS configuration (controller node and compute node)<br/> 6. External NW creation,　External NW subnet creation<br/> 7. Virtual NW creation<br/> 8. Router creation <br/> 9. Router GW configuration<br/> 10. Router IF configuration<br/> 11. VM creation <br/> SNAT communication confirm<br/> 12. All resources deletion<br/> 13. OVS stop (controller node and compute node)<br/> 14. ODL stop</p> <p>○log files<br/> We will upload them if necessary.</p> <p>Environment Details:<br/> OpenStack Version: stable/ocata<br/> ODL Version: Carbon-SR2<br/> Node:<br/> OpenStack Controller Node: 1 node<br/> OpenStack Compute Node: 1 node<br/> ODL Node: 3 nodes(cluster)</p> NETVIRT-966

In conntrack mode, SNAT communication fails if router gateway is set before virtual NW creation or VM creation in a fresh environment.

Bug Medium Resolved Cannot Reproduce Bertrand Low Ran Xiao Tue, 24 Oct 2017 11:01:27 +0000 Thu, 5 Apr 2018 21:28:23 +0000 Thu, 5 Apr 2018 21:28:23 +0000 Carbon Carbon General 0 7 <p>Hi there,</p> <p>the bug reproduction steps appear to contradict the bug title and description. Can you confirm if step 7 (Virtual NW creation) is in the correct order, or if it should come after step 9 or even 10 (as per title "...SNAT communication fails if router gateway is set before creating virtual NW...")?</p> <p>thanks,</p> <p>Bertrand</p> <p>Hi Bertrand,<br/> The bug reproduction steps are correct, as the issue is that SNAT communication only works when router gateway is set after both Virtual NW and VM creation.<br/> The steps we shared above is one of resources creation orders that SNAT can not work.<br/> Sorry for making you confused.<br/> I've modified the bug title and description.</p> <p> </p> <p>Hi Ran,</p> <p>thanks for the clarification. How do you verify the SNAT communication?</p> <p>Hi Bert,</p> <p>To verify SNAT communication ,please follow the below steps:</p> <p>1. ODL initialization (all 3 odl nodes)<br/> 2. ODL 3 node cluster start<br/> 3. neutron-server and related agent restart (controller node)<br/> 4. OVS configuration (controller node and compute node)<br/> 5. Create Internal Network and Subnet (net1)<br/> 6. Create 2 VM's for this network.(vm1)<br/> 7. Create external network and subnet (externl_net1)<br/> 8. Create a router and set gateway to the external network and add a interface to the internal network (net1)<br/> 9. Check for the communication from vm1 to external host by Pinging from vm1 to the external host. In conntrack mode SNAT can be verified using PING.<br/>  </p> <p>Hi Yogalakshmi,</p> <p>Thanks for your quick reply to Bertrand's question.</p> <p>Hi Bertrand,</p> <p>The verify steps I executed are the same as Yogalakshmi shared.<br/> If there is any other information i can provide, please let me know.</p> <p> </p> <p>Hi Ran,</p> <p>I have followed the reproduction steps indicated in this bug report, and I have used both ssh and ping to test the SNAT connection from the VM instances to the external host.</p> <p>I can verify that the bug is no longer reproduced using the following snapshots:</p> <ul class="alternate" type="square"> <li>distribution-karaf-0.6.3-20171128.170008-250.tar.gz (Carbon)</li> </ul> <p>- karaf-0.7.1-20171122.210509-337.tar.gz (Nitrogen)</p> <p>SNAT communication is successful.</p> <p>I recommend closing this bug as it is no longer reproducible.</p> <p>thanks,</p> <p>Bertrand</p> Development Rank 0|i036t3: