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-1471 netvirt <p>1. TROUBLE DESCRIPTION<br/> 1.1. Trouble summary<br/> To ping6 across VMs , ipv6 stateful mode subnet need to be added to the router or else ping6 will fail <br/> since MAC addreess getting resolved only adding  to the router. Looks like design need to be change<br/>  ostinato@ostinato:~$ ifconfig<br/> ens3      Link encap:Ethernet  HWaddr fa:16:3e:fa:22:9c  <br/>           inet addr:10.10.10.8  Bcast:10.10.10.255  Mask:255.255.255.0<br/>           inet6 addr: 2001:db8:aaaa::f/128 Scope:Global<br/>           inet6 addr: fe80::f816:3eff:fefa:229c/64 Scope:Link<br/>           UP BROADCAST RUNNING MULTICAST  MTU:9000  Metric:1<br/>           RX packets:13341 errors:0 dropped:0 overruns:0 frame:0<br/>           TX packets:13350 errors:0 dropped:0 overruns:0 carrier:0<br/>           collisions:0 txqueuelen:1000 <br/>           RX bytes:1169148 (1.1 MB)  TX bytes:1169279 (1.1 MB)</p> <p>lo        Link encap:Local Loopback  <br/>           inet addr:127.0.0.1  Mask:255.0.0.0<br/>           inet6 addr: ::1/128 Scope:Host<br/>           UP LOOPBACK RUNNING  MTU:65536  Metric:1<br/>           RX packets:160 errors:0 dropped:0 overruns:0 frame:0<br/>           TX packets:160 errors:0 dropped:0 overruns:0 carrier:0<br/>           collisions:0 txqueuelen:1 <br/>           RX bytes:11840 (11.8 KB)  TX bytes:11840 (11.8 KB)</p> <p>ostinato@ostinato:~$ ping6 2001:db8:aaaa::a<br/> PING 2001:db8:aaaa::a(2001:db8:aaaa::a) 56 data bytes<br/> 64 bytes from 2001:db8:aaaa::a: icmp_seq=1 ttl=64 time=0.411 ms<br/> 64 bytes from 2001:db8:aaaa::a: icmp_seq=2 ttl=64 time=0.280 ms<br/> ^C</p> <p>1.3. Frequency of fault<br/> Seen multiple times</p> <p>1.4. Workaround and Recovery<br/> Need to add to the router</p> <p>1.5. System impact</p> <p>ipv6 stateful mode is not working as expected </p> <p>1.6. Known problem within the OpenStack community?<br/> NA</p> <p>2. Configuration data<br/> 2.1. HW Configuration details<br/> EBS, HP or DELL:<br/> Number of subracks:<br/> Number of blades:<br/> EMC (Yes or No): <br/> Switches used (X670V, X770 or BD):</p> <p>2.2. SW Configuration details<br/> CEE-  CEE8RC2 <br/> CSC - 19.1.1.0-269<br/> CSS -2.8.2.css5.R2C</p> NETVIRT-1471

ipv6 stateful mode subnet need to be added to the router for pinging across VMs

Bug High Resolved Won't Do Karthikeyan Krishnan RajaRajan Manickam Wed, 24 Oct 2018 11:26:44 +0000 Tue, 30 Oct 2018 05:16:45 +0000 Tue, 30 Oct 2018 05:03:23 +0000 Neon ipv6service 0 3 <p><a href="https://jira.opendaylight.org/secure/ViewProfile.jspa?name=rajamk" class="user-hover" rel="rajamk">rajamk</a>, in the above logs the VM IPv6 address is shown as "2001:db8:aaaa::f/128"<br/> May i know how you created the IPv6 Subnet? Also, if you do not associate a router to the subnet there will not be any Router Advertisements sent over the network. So, please share the detailed steps.</p> <p>Have tested the IPv6 L2 data traffic with <b>DHCPv6 state-full address mode</b> and below are the our observations/findings. </p> <ul> <li> After booted the dual stack VMs <b>(Host OS: Ubuntu 16.04.3  LTS)</b>, explicitly initiated the DHCPv6 request(solicit) on the VM console using the command "dhclient -6 &lt;eth0&gt;"</li> <li> After the dhcpv6 client request, VMs are got the IPv6 GUA addresses via openstack q-dhcp service(Advertise/Request/Reply).</li> </ul> <p> We are observing IPv6 default route information is not configured by DHCPv6 process. It is taken care of IPv6 GUA address assignment only. Please find attached pcap file for booted VM.</p> <p> <b>Manually added "default route" information on the booted VMs from the same subnet, after that L2 data traffic is working fine.</b></p> <p> <b>Conclusion:</b></p> <ul> <li>   <b>DHCPv6 State-full mode:</b> IPAM only handled by DHCPv6 server. Adding default route information to the client is done by via "router advertisement" packet.</li> <li>   <b>DHCPv6 State-less Mode:</b> Other configurations like DNS Domain Suffix, etc is handled by DHCPv6 server. Both IPAM and default route configuration to the client is done by via "router advertisement" packet.</li> <li>   <b>SLAAC Mode:</b> IPAM and default route configuration to the client is done by via "router advertisement" packet.</li> </ul> <p> In all of the supported IPv6 address mode router advertisement is mandatory to work properly L2 traffic.</p> <ul> <li>   <b>DHCPv6 State-full mode:</b> IPAM only handled by DHCPv6 server. Adding default route information to the client is done by via "router advertisement" packet.</li> <li>   <b>DHCPv6 State-less Mode:</b> Other configurations like DNS Domain Suffix, etc is handled by DHCPv6 server. Both IPAM and default route configuration to the client is done by via "router advertisement" packet.</li> <li>   <b>SLAAC Mode:</b> IPAM and default route configuration to the client is done by via "router advertisement" packet.</li> </ul> <p> In all of the supported IPv6 address mode router advertisement is mandatory to work properly L2 traffic.</p> <p> </p> <p>i.e: DHCP for IPv4 = functionality DHCPv6 + RA (via router).</p> <p> </p> <p>Hence closing this bug now.</p> <p>Yes in IPv6 use-cases, the RA plays an important role and VMs learn the default route via the RA.</p> Development Rank 0|i03ju7: