search

sdnfv / openNetVM

A high performance container-based NFV platform from GW and UCR.
http://sdnfv.github.io/onvm/
Other
261 stars 134 forks source link

Packet's src & dst IPs in bridge and RX size. #244

Open MauriceMedhat opened 4 years ago

MauriceMedhat commented 4 years ago

My topology is 3 nodes on cloudlab, pcktgen on node1 and bridge on node2. When I start generating packets from node1 to node3, on the onvm manager it appears that all packets were received and transmitted by it but on node3 only 255 appears in RX packets and the rest are on RX missed, is there a way to increase the port to receive more than 255 packets only? At the bridge the packets appear to have the same IP as destination and source, although I set the src IP with a different one at the config file of pcktgen. How does it appear like that on the bridge? Thanks in advance!

Inkednode3_LI Inkednode2bridge_LI

Yongeverhuang commented 4 years ago

Currently i try the experiment on node 1 and 2, Pktgen generating packets from node1 to node2 through directly connected port. I had the same problem in the beginning, after that someone give me advice : in node 1 testpmd uses set fwd io to make packets forward between ports, then it works nomal. Wish it help you.

EthanBaron14 commented 4 years ago

Hi @MauriceMedhat, thanks for posting this. Over the last few days, I have been working on a guide for configuring a CloudLab three-node topology using ONVM Manager and Bridge NF on the intermediate node (node2). The first step that I would recommend for troubleshooting the issue that you're having would be to run through that. It's not entirely ready to be added to a Wiki on here (just some basic editing work that has to get done), but it should still serve its purpose well. You can view the guide here. If you're still not able to get this to work, let me know and we can troubleshoot further.

twood02 commented 4 years ago

If the link above is showing you broken image files, here is a PDF version that might work better. ONVM_Documentation.pdf

MauriceMedhat commented 4 years ago

Currently i try the experiment on node 1 and 2, Pktgen generating packets from node1 to node2 through directly connected port. I had the same problem in the beginning, after that someone give me advice : in node 1 testpmd uses set fwd io to make packets forward between ports, then it works nomal. Wish it help you.

My topology is with 3 nodes not 2 because Im testing the bridge but I have tried it in my topology and didnt send from node 1 to node 3 with set fwd io.

Hi @MauriceMedhat, thanks for posting this. Over the last few days, I have been working on a guide for configuring a CloudLab three-node topology using ONVM Manager and Bridge NF on the intermediate node (node2). The first step that I would recommend for troubleshooting the issue that you're having would be to run through that. It's not entirely ready to be added to a Wiki on here (just some basic editing work that has to get done), but it should still serve its purpose well. You can view the guide here. If you're still not able to get this to work, let me know and we can troubleshoot further.

The documentation was helpful but didn't answer my questions.

twood02 commented 4 years ago

@MauriceMedhat - you report that node3 is reporting a high number for rx-missed. What is running on node3? Is it just pktgen? If that host isn't running ONVM it seems like it must be an issue with your DPDK or pktgen configuration.

Does the ONVM manager on node2 report that it is receiving and sending out the NIC port at the correct rates?

MauriceMedhat commented 4 years ago

@MauriceMedhat - you report that node3 is reporting a high number for rx-missed. What is running on node3? Is it just pktgen? If that host isn't running ONVM it seems like it must be an issue with your DPDK or pktgen configuration.

Does the ONVM manager on node2 report that it is receiving and sending out the NIC port at the correct rates?

Node3 has testpmd running on it to check the port stats, node1 has the pktgen to send to node3. Yes ONVM manager seems to be working fine on node2 as you can see in the picture. My questions are can we increase the RX on node3 to receive more than 255 packets and why does the packets have one IP address (Node3 IP which is 192.168.3.3) for the source and destination? onvm mgr