Bug #1456
closedUDP multicast does not work properly
0%
Description
NFD commit c1d6a6b089489a41f30355829435e86eb559d088
Steps to reproduce:
Follow the steps in task 1390 or run the attached script.
The script I am using is attached. Please place them in the same directory with same path on each node.
When I test with ethernet multicast face, the test passed. The command to test with ethernet face is:
./multicast-test.sh eth1 10.0.0.2 10.0.0.3
However, when I test with udp multicast face (./multicast-test.sh 10.0.0.1 10.0.0.2 10.0.0.3
), it failed because no data come back for ndn:/test-mcast/B/1. ndn-traffic-server log on host B shows that it saw the interest.
Additional information:
I have run traceroute from all three nodes and they are able to reach the other two nodes within one hop.
I have tested tcp-udp-tunnel scenario using A and B from above. The test passed.
In multicast-test.sh I tried to comment out the lines that starts nfd and ndn-traffic-server on C. I still cannot get data back for ndn:/test-mcast/B/1.
Files
Updated by Yi Huang over 10 years ago
- Tracker changed from Task to Bug
- Target version set to v0.1
- Start date deleted (
04/09/2014)
Updated by Junxiao Shi over 10 years ago
- Category set to Faces
- Status changed from New to Rejected
I cannot reproduce this bug with NFD commit edae353eae13ff94c2d780489b36b832cced2212.
Attempted steps:
start NFD on hostA and hostB (two hosts are enough because report says bug persists if hostC doesn't run NFD)
on hostA, execute
build/bin/nfdc set-strategy ndn:/ ndn:/localhost/nfd/strategy/broadcast for faceid in `build/bin/nfd-status -f | grep -e 'udp4://' -e ':56363' | awk '{ print $1 }' | cut -d= -f2` do build/bin/nfdc add-nexthop ndn:/test-mcast $faceid done
on hostB, start
ndnpingserver /test-mcast
on hostA, execute
ndnping /test-mcast
ndnping is able to receive Data in the last step.
I believe these simplified steps are equivalent to those listed in #1390 for testing UDP multicast between two hosts.
Updated by Yi Huang over 10 years ago
That is why I didn't understand and thought it might be a VM specific issue. Actually host C works as expected. Only ndn:/test-mcast/B/1 cannot get data back.
When I run my script with nfd and traffic generator disabled on B, the test passed as well.
However, when I disable nfd and traffic generator on C, the test failed. I cannot even get data for ndn:/test-mcast/Z/1. server log shows that the server saw the interest.
Updated by Junxiao Shi over 10 years ago
Please report these from host B:
- In nfd.conf, what's the value of udp.mcast?
- Does tcpdump show a Data packet from B's unicast IP to the multicast group IP? What's the size of this packet, compared to MTU?
- Set log level of Forwarder to TRACE. Does the log show outgoing Data pipeline is entered for a Data toward the multicast face?
Also try using 10.0.0.3 as B, and 10.0.0.2 as C. Does the problem persist on B?