Yes. OVS becomes a standard switch when you disable openflow. There may be other things going on. A fresh start may be a good idea. Maen
On Sat, Feb 16, 2019 at 4:18 PM Diarmuid O Briain <diarm...@obriain.com> wrote: > Maen, > > The irony is that when I disable OpenFlow on the switch it immediately > starts switching ;-) I guess Fail-Over is working at least. > > (OF-SW) (Config)# *no openflow enable * > > 199.9.9.11_Host# *ping 199.9.9.22* > > Type escape sequence to abort. > Sending 5, 100-byte ICMP Echos to 199.9.9.22, timeout is 2 seconds: > .!!!! > Success rate is 80 percent (4/5), round-trip min/avg/max = 1/1/4 ms > > 199.9.9.11_Host# *ping 199.9.9.33* > > Type escape sequence to abort. > Sending 5, 100-byte ICMP Echos to 199.9.9.33, timeout is 2 seconds: > .!!!! > Success rate is 80 percent (4/5), round-trip min/avg/max = 1/1/1 ms > > regards, > > Diarmuid > -- > > *Irish by birth, located in Uganda but Munster by the grace of God.* > > > > > On Sat, 16 Feb 2019 at 23:13, Diarmuid O Briain <diarm...@obriain.com> > wrote: > >> Maen, >> >> Cool, thanks, that makes sense. >> >> I redid the test with the Netgear with the addition of the 'eth_src=src' >> in the 'parser.OFPMatch'. There is still no learning happening for some >> reason. It is late here so I will get back to it in the morning to see if I >> can figure out why. Any ideas would be greatly appreciated. >> >> regards, >> >> Diarmuid >> -- >> >> *Irish by birth, located in Uganda but Munster by the grace of God.* >> >> >> >> >> On Sat, 16 Feb 2019 at 22:37, Maen Artimy <arti...@gmail.com> wrote: >> >>> No I didn't see the question. I'm assuming you are talking about this >>> FlowManager: https://github.com/martimy/flowmanager. The two apps may >>> conflict if they are using the same port (8080), otherwise try. >>> >>> ryu-manager ~/flowmanager/flowmanager.py ryu.app.ofctl_rest >>> >>> BTW, you asked >>> What is the origin of the Priority 65535 flow in scenario A ? >>> >>> This entry is created when the controller is run with the >>> --observe-links option. >>> $ ryu-manager --observe-links >>> The controller adds an entry that forwards all LLDP packets to the >>> controller with the highest possible priority. Remove the option and the >>> entry will disappear. >>> Sending LLDP packets to the controller helps the controller "understand" >>> the topology but it is not mandatory. You can use it with the topology >>> display applications. >>> >>> Regards, >>> Maen >>> >>> On Sat, Feb 16, 2019 at 3:01 PM Diarmuid O Briain <diarm...@obriain.com> >>> wrote: >>> >>>> Maen, >>>> >>>> I understand. Why I wonder is the version with the python3-ryu package >>>> as it is? Is that the Ryu group or Ubuntu repository package management? >>>> >>>> BTW did you see the question I asked about Flow Manager not working >>>> with the REST API. Is that normal or can the two be ran together ? >>>> >>>> Regards, >>>> >>>> Diarmuid >>>> >>>> Regards, >>>> >>>> Diarmuid >>>> >>>> On Sat 16 Feb 2019, 21:19 Maen Artimy <arti...@gmail.com wrote: >>>> >>>>> Here is at least one issue that may arise if you match based on >>>>> destination MAC only: >>>>> 1. Say Node A is connected to switch port 1. The controller creates a >>>>> flow entry that forwards packets with destination MAC A to port 1. >>>>> 2. If Node A is moved to switch port 2. The controller never learns >>>>> about the change and all packets with destination MAC A match existing >>>>> entry and go to port 1 (and never reach A). >>>>> >>>>> If the match is based on both source and destination MACs: >>>>> 2. If Node A is moved to switch port 2, the packets from A will be >>>>> sent to the controller because they don't match any entry and the >>>>> controller will create a new entry withe proper output port (The old entry >>>>> will remain though because there is not time-out assigned to it). >>>>> >>>>> I hope this helps. I may have been too brief. >>>>> >>>>> Regards, >>>>> Maen >>>>> >>>>> >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >>>>> Virus-free. >>>>> www.avast.com >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >>>>> <#m_-7593294074257726621_m_-6334976357517993422_m_-1790520956640482131_m_3873663500120092052_m_3224595215056297631_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> >>>>> >>>>> On Sat, Feb 16, 2019 at 1:57 PM Diarmuid O Briain < >>>>> diarm...@obriain.com> wrote: >>>>> >>>>>> Maen, >>>>>> >>>>>> Interesting, I compared the *simple_switch13.py* file you linked me >>>>>> to and the one that comes with the *python3-ryu* package (installed >>>>>> with apt-get install python3-ryu). >>>>>> >>>>>> $ *diff /usr/lib/python3/dist-packages/ryu/app/simple_switch_13.py >>>>>> maen_version_simple_switch_13.py* >>>>>> 105c105 >>>>>> < match = parser.OFPMatch(in_port=in_port, eth_dst=dst) >>>>>> --- >>>>>> > match = parser.OFPMatch(in_port=in_port, eth_dst=dst, >>>>>> eth_src=src) >>>>>> 120d119 >>>>>> < >>>>>> >>>>>> That would seem to bear out what you say. I am a little lost as to >>>>>> what that means though. Why the difference? >>>>>> >>>>>> regards, >>>>>> >>>>>> Diarmuid >>>>>> -- >>>>>> >>>>>> *Irish by birth, located in Uganda but Munster by the grace of God.* >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On Sat, 16 Feb 2019 at 19:04, Maen Artimy <arti...@gmail.com> wrote: >>>>>> >>>>>>> Diarmuid, >>>>>>> There is something odd about the application you are running. The >>>>>>> simple_switch_13.py should match three fields not two. Also it should >>>>>>> not >>>>>>> output the packet to the same input port (see line 105) . So I would >>>>>>> start >>>>>>> by checking if the application is correct. >>>>>>> https://github.com/osrg/ryu/blob/master/ryu/app/simple_switch_13.py >>>>>>> >>>>>>> Regards, >>>>>>> Maen >>>>>>> >>>>>>> On Sat, Feb 16, 2019 at 11:36 AM Diarmuid O Briain < >>>>>>> diarm...@obriain.com> wrote: >>>>>>> >>>>>>>> Hi, >>>>>>>> >>>>>>>> I have spent ages playing with Ryu and mininet and decided to try >>>>>>>> some actual physical switches. I compared using Ryu with an OvS >>>>>>>> switching >>>>>>>> for physical ports in a server with Ryu controlling a Netgear >>>>>>>> M4300-48G in >>>>>>>> OpenFlow v1.3 mode. The response in both cases were different and I >>>>>>>> documented it in the attached. I would appreciate it if someone who >>>>>>>> knows >>>>>>>> more about the Ryu controller than me would look and try explain the >>>>>>>> different outcomes from scenarios on A and B. >>>>>>>> >>>>>>>> much appreciated. >>>>>>>> >>>>>>>> Diarmuid O'Briain >>>>>>>> -- >>>>>>>> >>>>>>>> *Irish by birth, located in Uganda but Munster by the grace of God.* >>>>>>>> >>>>>>>> >>>>>>>> _______________________________________________ >>>>>>>> Ryu-devel mailing list >>>>>>>> Ryu-devel@lists.sourceforge.net >>>>>>>> https://lists.sourceforge.net/lists/listinfo/ryu-devel >>>>>>>> >>>>>>> >>>>> >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >>>>> Virus-free. >>>>> www.avast.com >>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >>>>> <#m_-7593294074257726621_m_-6334976357517993422_m_-1790520956640482131_m_3873663500120092052_m_3224595215056297631_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> >>>>> >>>>
_______________________________________________ Ryu-devel mailing list Ryu-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ryu-devel
