Re: [pox-dev] Question about the capacity of pox controller

Sure, though it depends on how fast the machine running POX is, whether you're using CPython or PyPy, etc. How fast the switch is may also be an issue. In general, if you're doing work where performance is critical, you probably shouldn't be using Python. On Thu, Aug 22, 2019 at 10:02 AM

Re: [pox-dev] problem in discovery.py function

AASIF MNIT wrote: > I am trying a way to modify lldp packet such that it can discover topology > in hybrid Sdn(having legacy switches). what fields should I add or delete > from lldp frame formate? > > On Mon, 8 Jul 2019 at 00:52, Murphy McCauley > wrote: > >>

Re: [pox-dev] problem in lldp.py

The constructor arguments are basically the same for all classes in the packet library. raw is used if you are parsing a raw packet that's held in a bytes/str object. prev is mostly used internally to build chains of packet headers. That is, since most packet headers are preceded by other packet

Re: [pox-dev] problem in discovery.py function

So this is the constructor for LLDPSender. The docstring explains what the two parameters are. ttl probably isn't that important. send_cycle_time has to do with how fast discovery is sending packets. You don't usually construct LLDPSenders yourself, though. Normally, the discovery component

Re: [pox-dev] Flow rule installation using POX

; > > Thanks, > > ******** > Himanshu Sukheja > -- > *From:* Murphy McCauley > *Sent:* 14 March 2019 07:08 > *To:* Sukheja, Himanshu > *Cc:* pox-dev@lists.noxrepo.org > *S

Re: [pox-dev] Flow rule installation using POX

You are just creating a message, but that's not enough -- you have to send it to a switch. This section of the POX manual may be of help: https://noxrepo.github.io/pox-doc/html/#communicating-with-datapaths-switches You also might take a look at some of the sample POX components, particularly

Re: [pox-dev] Redirect "ofp_packet_in" packet among multiple controllers in POX?

It sounds like you're on to a solution, but here are some other thoughts... 1) Depending on what kind of switch you're using and exactly how your controllers are supposed to work... it may be possible to just have the switch connect to multiple controllers (and have the controllers behave

Re: [pox-dev] writing flow rules in intermediate nodes

First off, in many cases, it doesn't make much sense to think of the switches as having IP addresses. Hosts have IP addresses. There are exceptions, but... are you sure you're in one of those exceptional cases? If not, this is just going to be confusing. That aside, you mention wanting to get

Re: [pox-dev] Handling of ARP requests at POX

On Nov 27, 2015, at 6:14 AM, Ramana Reddy wrote: > > Hi All, > > I have few concerns while handling arp requests at pox. > > 1. How to handle retransmission of arp reply in case if arp reply is dropped. > Is there any API which allows to retransmit the packet. I'm not

Re: [pox-dev] question about output ports

On Nov 25, 2015, at 12:02 PM, Kostas Chartsias wrote: > > Fisrtly i know this is most likely a mininet question, but is there an easy > way to define which port number is assigned to switch ports other than dpctl > dump-flows?? > for example mininet links output for a

Re: [pox-dev] Reassembling the fragmented packets at pox

If you mean that all the fragments of a packet are being sent to POX via packet_in messages, then... sure? Just follow the IP RFC for doing reassembly. (That is, I don't think POX has any specific helpers for doing this, but it's not a particularly difficult algorithm if you're not trying to

Re: [pox-dev] L2Multi + OpenFlow Discovery

On Nov 24, 2015, at 6:52 PM, Shen Ting Ang wrote: > Hi all, > > Following up from my previous question about POX and fat-tree topologies, > I've tried using forwarding.l2multi and openflow.discovery. However, I now > run into a slightly different problem: there's only

Re: [pox-dev] VLAN and nicira

I think you want match.vlan_tci = 10 On Nov 11, 2015, at 4:59 AM, Amer wrote: > Dear All > > May you please provide me with an example of using VLAN under nicira flow_mod > and match rule. > > I received an error when I used it as the following: > > msg =

Re: [pox-dev] MPLS match and labels popping

On Nov 8, 2015, at 8:04 AM, AMER wrote: > Dear All > > I would like to ask you about the following match commands and actions > >msg1 = nx.nx_flow_mod() >msg1.match = nx.nx_match() >msg1.match.eth_type = pkt.ethernet.MPLS_TYPE >

Re: [pox-dev] Nox installation error

To have any chance of getting help, you probably need to post which version of NOX, and the output of configure/make. Also, the chance of someone else who can help seeing it is probably higher if you post on nox-dev instead of pox-dev. -- Murphy On Oct 22, 2015, at 2:39 PM, Amer

Re: [pox-dev] How to send the whole packet from OVS to pox controller

There are two major possibilities. 1) Install a low priority table entry that sends to the controller and specifies the full payload size (I think this last part is the default, actually). 2) Change the default size for packets which miss the table entirely. The best way to do this is simply

Re: [pox-dev] Protocol between multiple POX controllers

I usually use the messenger component for this. -- Murphy On Oct 1, 2015, at 3:42 AM, Ramana Reddy wrote: > Hi All, > > Is there any mechanism to transfer data between multiple pox controllers. > If not, what typical changes we need in the pox code to implement this. >

Re: [pox-dev] pox with flowvisor

My first guess is that something is wrong with your slice configuration, but I'm the wrong person to ask about that (maybe try openflow-discuss?). If I were trying to debug this, I might... Capture all the OpenFlow traffic between the controllers and FlowVisor. Write a simple controller for

Re: [pox-dev] pox with flowvisor

If the POX instances are running on different FlowVisor slices, I wouldn't think the POX for one slice would be able to delete entries on the other. -- Murphy On Sep 23, 2015, at 8:29 PM, Vishlesh Patel wrote: > Hi all, > > I am using pox with flowvisor. I have 3

Re: [pox-dev] ARP_Responder

Can you give more information? What is the complete POX commandline you're using? Have you made any modifications? What exact behavior are you seeing (who is sending what ARP requests, who is sending what ARP replies, and who is receiving the requests and replies)? What exact behavior are

Re: [pox-dev] Fwd: Content based HTTP load balancer

If I were doing it, I would seriously consider whether there was a reason to implement something myself using OpenFlow rather than using an existing HTTP proxy/load balancer (or modifying an existing one). Good luck. -- Murphy On Sep 3, 2015, at 9:23 AM, Ramana Reddy

Re: [pox-dev] Riplpox with large topology

phy > Yes, it is working correctly with the default number of host (i.e., 16 hosts). > > Best regards, > Amer > > Sent from my iPhone > > On Sep 3, 2015, at 4:00 AM, Murphy McCauley <murphy.mccau...@gmail.com> wrote: > >> Did it work correctly with a smaller number o

Re: [pox-dev] Error while setting up PoX controller with GENI-based fat-tree topology

openflow.discovery --eat-early-packets openflow.spanning_tree --no-flood --hold-down by following this link but still hosts are unable to communicate. On Sun, Aug 16, 2015 at 2:51 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: Pleases see the I'm seeing many packet_in messages

Re: [pox-dev] Error while setting up PoX controller with GENI-based fat-tree topology

Pleases see the I'm seeing many packet_in messages and forwarding isn't working; what gives? and Does POX support topologies with loops? questions in the FAQ. -- Murphy On Aug 15, 2015, at 2:11 AM, Sadia Bashir 11msccssbas...@seecs.edu.pk wrote: Hello, I have set up fat-tree topology

Re: [pox-dev] Multiple POX controller setup

None of the example forwarding components that ship with POX contain any special features to facilitate distribution. -- Murphy On Jul 24, 2015, at 2:25 AM, Sulabh Deshmukh sulabh02.deshm...@gmail.com wrote: Hello Experts, I recently started working with the POX and currently

Re: [pox-dev] Multipath using MPLS

I think you have two requirements here. 1) You need to be able to push and pop MLPS labels on a packet. 2) You need to be able to create table entries which match on the MPLS label. If I remember correctly, both of these are available in POX through the nicira.py file. The first is done with

Re: [pox-dev] How do i interface the POX Controller with the OVS switches

Sounds like you need to configure Open vSwitch to create OpenFlow 1.0 connections to your controller. This involves something roughly like... ovs-vsctl set-controller your-switch-name tcp:controller-ip:6333 ovs-vsctl set bridge your-switch-name protocols=OpenFlow10 .. but you should probably

Re: [pox-dev] Priority doesnt change

: Dear Murphy, This Guy is working with another project. I tried to set different priorities to flows because I need to work with different flows, but it doesn’t work. All of the priorities are set to default. Could you help me? Thanks, Anees From: Murphy McCauley

Re: [pox-dev] Raise event for ofp_packet_out

If you wanted to try to do it for arbitrary unmodified components, I don't see too much alternative to Connection.send(). Unfortunately, there's no requirement that it gets passed message objects instead of raw bytes, and if it's passed raw bytes, there's no requirement that they be complete

Re: [pox-dev] Priority doesnt change

Does this thread from the archive answer your question? https://www.mail-archive.com/pox-dev@lists.noxrepo.org/msg01838.html -- Murphy On Jul 9, 2015, at 9:15 PM, Anees Mohsin Hadi Al-Najjar a.alnajj...@uq.edu.au wrote: Dear Colleagues, In my controller, I applied different priorities

Re: [pox-dev] Packet_In Packet_out

Can you post a minimal working example that demonstrates the problem? On Wednesday, July 1, 2015, Anees Mohsin Hadi Al-Najjar a.alnajj...@uq.edu.au wrote: Dear Guys, I,ve faced a problem when I sent an ARP request as a Packet out of the controller. Instead to send the packet out

Re: [pox-dev] ICMP type packet

The proto.pong component replies to pings. Actually, it replies to *all* ICMP messages. Here's an untested quick hack at very slightly modifying it so that it's a bit more discerning.# Copyright 2012 James McCauley # # Licensed under the Apache License, Version 2.0 (the License); # you may not use

Re: [pox-dev] Set Priority on Flow Rules

In plain OpenFlow 1.0, this isn't possible because l2_learning installs exact matches, which have maximum priority. There are at least two ways to fix this: 1) Don't install exact matches for normal learning. This is pretty reasonable, actually. l2_pairs, for example, installs rules that

Re: [pox-dev] ARP in L2_learning

This would probably be easier if you posted the details of the four packets. -- Murphy On Jun 23, 2015, at 6:54 PM, Talal Alharbi tal...@hotmail.com wrote: Dear All, Running a test on L2_learning module and trying to send an ARP packet from a host using Scapy. After sending an ARP

Re: [pox-dev] L2_Learning POX

Usually if you want to do something when switches disconnect, you do it by registering a handler for the ConnectionDown event on pox.core.openflow, as described in the OpenFlow Events section of the manual. You haven't given enough information to know for sure, but if what you want is to do

Re: [pox-dev] ports in pox

Looking at page 6 and 19 of the OpenFlow spec, you can see that it's done via the local port. On POX, this means creating an output action to the OFPP_LOCAL virtual port. -- Murphy On Jun 3, 2015, at 3:32 AM, AMER amer7...@hotmail.com wrote: Hello, Dear All I would like to build a

Re: [pox-dev] Help Regarding Flow Statistics Module

The statistics that you gather are specific to the table entries that you install. If you have table entries for each address, you should get back flow stats for each entry and therefore each address... -- Murphy On Jun 3, 2015, at 2:19 AM, Tom Mody bug29...@gmail.com wrote: Hi all, I am

Re: [pox-dev] l2_learning flow installation

On May 27, 2015, at 2:29 PM, Adam Pavlidis adampavli...@gmail.com wrote: Are you using l2_learning and only l2_learning? I am using l2_learning, log.level with --DEBUG flag, and a custom module that only handles Flow Removed events (only to print them) Are you sure? Try monitoring

Re: [pox-dev] A flow removal question

The information you want may not be available on the FlowRemoved event object itself, but as described in the OpenFlow Events section of the manual, the actual OpenFlow protocol message that caused the event is available on the event's .ofp attribute. In this case, that means it's an

Re: [pox-dev] help with priorities

I think you're probably using l2_learning, which installs exact match table entries. In OpenFlow, exact matches always have maximum priority -- you can't override it. Here are some options for getting around the problem: 1. Make the exceptional table entries you install *also* be at maximum

Re: [pox-dev] How to use popen to send parallel iperf?

I think this is probably better aimed at the mininet-discuss list. Good luck. -- Murphy On May 14, 2015, at 3:00 AM, 博威张 zbw921...@gmail.com wrote: Hi, The popen function is here : Node.popen(self,args,kwargs) I have read examples/popenpoll.py and examples/popen.py, both of which only

Re: [pox-dev] Host_Tracker

Well it only sends them under certain conditions (IIRC, it sends them when a host has been quiet for a while and h_t is trying to see if the host is still there or not), and it's not a normal ICMP ping -- it's an ARP request. -- Murphy On May 12, 2015, at 11:06 PM, Talal Alharbi

Re: [pox-dev] help with openflow actions

The problem is probably that you're rewriting the IP addresses but not the Ethernet addresses. In such a case, for example, .4 sees the rewritten packet, but it still has .2's Ethernet address. .4 never even looks at the IP because the packet appears to be for someone else at the Ethernet

Re: [pox-dev] cannot convert argument to integer in pack self.max_len

In Python, logical lines of code can span multiple physical lines. When printing tracebacks, it only prints the final physical line of the logical line of code in question. So while you're seeing line 1560 and only seeing the self.max_len, the actual problem may well be on the line before

Re: [pox-dev] How the class of Switch (EventMixin) in l2_multi works?

On May 8, 2015, at 9:49 PM, 博威张 zbw921...@gmail.com wrote: Hi, I have following confusions about the class of Switch (EventMixin) in l2_multi : 1.No instance of the class of Switch (EventMixin) has been found in component of l2_multi, so how does it work? Sure there are instances of

Re: [pox-dev] sFlow and POX

POX presently has no special support for sFlow. But, IIRC, for the most part, sFlow just samples packets and encapsulates them in UDP. You could certainly extend POX to listen on a UDP socket to get the sFlow packets. BUT... if you're going to have a large number of sFlow packets (e.g., your

Re: [pox-dev] Modules Depend on Host_Tracker

I don't think POX ships with anything that depends on it, though gephi_topo can optionally use it to display hosts in Gephi. -- Murphy On Apr 27, 2015, at 12:39 AM, Talal Alharbi tal...@hotmail.com wrote: Dear All, I'm just carious is there any modules depend on Host_tracker ? Like

Re: [pox-dev] Bad OpenFlow version (4) on connection with ovs switches

Is the version of POX you're using more than two years old? I think more recent versions are unlikely to have this happen. If your POX is old, you could update it (in which case, you'd probably want to run my fork of ripl/riplpox), or you could make the relevant change in your local POX (I

Re: [pox-dev] Bad OpenFlow version (4) on connection with ovs switches

for this problem? Thanks a lot for your help/suggestions. Regards, David On Thu, Apr 16, 2015 at 12:47 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: Is the version of POX you're using more than two years old? I think more recent versions are unlikely to have this happen. If your

Re: [pox-dev] Cbench issue with pox controller

It works fine for me using POX eel on both Linux and Mac OS, using both CPython and PyPy, using both --unthreaded-sh and not. Why are you trying to run cbench? What is it that you think cbench is going to tell you? -- Murphy On Apr 10, 2015, at 2:29 AM, Anil Kumar Konasale Krishna

Re: [pox-dev] Reading TLS header.

I think maybe the answer here is a combination of Lucas Brasilino's answer to your other question from earlier today and RFC 5246. -- Murphy On Apr 3, 2015, at 10:37 AM, chaitanya tanwar chaitanya.tan...@gmail.com wrote: HI all, How can i read the TLS header of a packet? Is there any

Re: [pox-dev] frequent packet_in in riplpox

Have you looked at the packet to see what it is it? It might be IPv6 neighbor discovery or something... -- Murphy On Mar 17, 2015, at 8:37 AM, Amer amer7...@hotmail.com wrote: Hello, Dear Murphy I run mininet in fattree, 4 topology and riplpox in proactive mode I have added a counter

Re: [pox-dev] ovsdb

On Mar 9, 2015, at 1:55 AM, Felician Nemeth neme...@tmit.bme.hu wrote: On Thu, Mar 05 2015, Murphy McCauley wrote: On Mar 4, 2015, at 3:51 AM, Felician Nemeth neme...@tmit.bme.hu wrote: Hi Murphy, Finally, I've had a little time to have a look at your ovsdb module. It's absolutely

Re: [pox-dev] POX controller giving error on giving SSL options

You're running POX carp, but SSL support was added in POX dart. See the manual about upgrading. -- Murphy On Feb 10, 2015, at 2:29 AM, Ruchika Verma ruve...@gmail.com wrote: hi, I am receiving following error while starting POX controller with SSL options. The .PEM files are created in

Re: [pox-dev] Getting counter values in POX controller

Haha. What a coincidence. :) On Feb 28, 2015, at 4:10 PM, Sulabh Bista sul...@gmail.com wrote: Hey, A generic method would be: Requesting statistics from a switch (see: https://openflow.stanford.edu/display/ONL/POX+Wiki#POXWiki-ofp_stats_request-Requestingstatisticsfromswitches)

Re: [pox-dev] dec_ttl not working

Have you tried looking at the table on the switch and verifying that you see the entry as you expect (e.g., including the decrement action)? The easiest way to do this is probably using ovs-ofctl. -- Murphy On Feb 28, 2015, at 12:28 PM, AMER amer7...@hotmail.com wrote: Hello, Dear I

Re: [pox-dev] Create ICMP packets in POX

ping. Please share if you have any idea. Thanks, Sandesh Shrestha On Fri, Feb 27, 2015 at 8:39 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: Jumping in with a quick comment below... On Feb 27, 2015, at 11:15 AM, Sandesh Shrestha sande...@ualberta.ca wrote: Hey Lucas

Re: [pox-dev] Getting counter values in POX controller

In OpenFlow, these values are known as statistics. Have you read the relevant sections of the OpenFlow specification and the POX manual? If not, I'd suggest reading the following sections in the POX manual: Statistics Events ofp_stats_request - Requesting statistics from switches Statistics

Re: [pox-dev] Create ICMP packets in POX

Jumping in with a quick comment below... On Feb 27, 2015, at 11:15 AM, Sandesh Shrestha sande...@ualberta.ca wrote: Hey Lucas, Thanks for the help. I was able to send icmp packets to mininet host using the code below. I used arbitrary source mac address and source ip. The destination ip

Re: [pox-dev] error in riplpox

First: What version of riplpox are you using? Second: I don't think any version of riplpox was ever used with and test with POX carp. The original riplpox was, I think, from the dart era or maybe even the angler era. My updated version of riplpox was for dart. -- Murphy On Feb 25, 2015, at

Re: [pox-dev] Install a flow in for loop

I expect the problem is that you're doing packet.IP_TYPE, but packet is a string, and not the part of the packet library with the IP_TYPE constant. As mentioned in the manual, POX convention is to import the packet library as: import pox.lib.packet as pkt .. after which you can do pkt.IP_TYPE

Re: [pox-dev] error usinf riplpox

I don't think riplpox is currently maintained. You could try using my fork, which is probably more up-to-date than the official one (github.com/MurphyMc/riplpox), or you could try using old versions of POX and Mininet. Good luck. -- Murphy On Feb 21, 2015, at 4:33 AM, AMER

Re: [pox-dev] Loop problem and Routing

for future QoS routing. Does the discovery.py component do what I explained without QoS values? Thanks, Murat On Wed, Feb 18, 2015 at 12:16 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: l2_multi is an example that does shortest paths between endpoints (it uses Floyd-Warshall

Re: [pox-dev] Loop problem and Routing

l2_multi is an example that does shortest paths between endpoints (it uses Floyd-Warshall instead of Dijkstra's, but that's just a detail). And the spanning_tree component doesn't disable entire ports, IIRC -- just flooding on those ports. So if you have loops and use l2_multi and

Re: [pox-dev] Error in openflow

That depends on the switch you're using. Unless you're using the POX datapath, you should probably ask on the switch's site/mailing-list/other-support-mechanism. For example, if you're using Open vSwitch, it has its own mailing lists. -- Murphy On Feb 2, 2015, at 1:50 AM, AMER

Re: [pox-dev] Error in openflow

Are you saying that you're altering the OpenFlow protocol so that it can specify VLAN tags larger than two bytes? And modifying a switch so that it can match and probably insert and remove these larger VLAN tags? (If so, I'd have thought that using VXLAN would be an easier way or that using

Re: [pox-dev] Understanding POX

The former is discussed in the Working with POX: The POX Core object section of the manual -- especially the first subsection of that: Registering Components. The second is addressed in the The Event System: pox.lib.revent section of the manual -- particularly in the first few sections (e.g.,

Re: [pox-dev] Spanning Tree Algorithm

I don't have enough information to understand or evaluate the information you've given about port status. Where does it come from? Are you querying the switch, or is it from one of the datastructures in spanning_tree.py, or something else? When are you printing it? And what options are you

Re: [pox-dev] Format of flow entries to use fast failover group type in OF 1.3

Since POX doesn't currently support OpenFlow 1.3, I'm not sure that this applies... -- Murphy On Jan 11, 2015, at 4:09 PM, Padma Jayasankar padmaj...@gmail.com wrote: Hi, I want to use fast fail over group type of OpenFlow 1.3. I need sample flow entries for doing this. According to my

Re: [pox-dev] MAC address encaps

There are multiple types of MAC-in-MAC encapsulation. I think recent versions of OVS are starting to have support for the kind done with Provider Backbone Bridging, but I haven't looked into it. It might be done with special push and pop actions like MPLS, which should be easy enough to add

Re: [pox-dev] Order of handlers for one event

I think you sent out an email saying you'd already found the answer on the POX wiki, but I happened to get the emails in reverse order (I only just got this one, but got the other one 20 minutes ago). For anyone wondering the same thing, I think the relevant answer is in the POX manual wiki in

Re: [pox-dev] Spanning Tree Algorithm

I believe you're mistaken and the ports to which hosts are attached have flood enabled. What makes you think this isn't the case? -- Murphy On Jan 9, 2015, at 7:21 PM, TALAL AL-HARBI tal...@hotmail.com wrote: If you print out ports status, you would see they are false. Sent from my

Re: [pox-dev] LLDP flow installing in discovery module

In the case of l2_multi, there isn't a point. But l2_multi isn't the only component that uses discovery. Ones that don't clear the table don't have this issue at all. And some that do clear the table (e.g., topo_proactive) install the table entry themselves. As for what caused this... I

Re: [pox-dev] ECMP in OF

You might also be interested in riplpox. Here's my fork: https://github.com/MurphyMc/riplpox Also, the Nicira bundle action is useful for this sort of thing. POX doesn't currently support it, unfortunately (I started it, but haven't had time to finish it). -- Murphy On Dec 31, 2014, at 2:01

Re: [pox-dev] discovery module LLDP destination MAC address

Yes, and it has been discussed before. I've added the answer to the POX FAQ. -- Murphy On Dec 26, 2014, at 8:16 AM, tim huang pds...@gmail.com wrote: Hi, I'm working on discovery module, and found that the LLDP destination MAC address is not the standard. Is there any particular reason

Re: [pox-dev] discovery module LLDP destination MAC address

No trouble. It should have been documented anyway! -- Murphy On Dec 26, 2014, at 11:03 AM, tim huang pds...@gmail.com wrote: Sorry to bother you. I should search archive firstly. Thanks for your patient reply. On Fri, Dec 26, 2014 at 1:01 PM Murphy McCauley murphy.mccau...@gmail.com

Re: [pox-dev] fast fail over type in group table

I think you'd use the OVS bundle action with the NX_BD_ALG_ACTIVE_BACKUP algorithm. POX doesn't currently have support for the bundle action, unfortunately. Implementing it shouldn't be that hard; I just haven't had the time. -- Murphy On Dec 16, 2014, at 8:30 PM, Padma Jayasankar

Re: [pox-dev] Over-writing table entries

addresses too. Best Regards. On Fri, Dec 19, 2014 at 8:39 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: Did you read the third item in the FAQ? On Dec 18, 2014, at 8:14 AM, chaitanya tanwar chaitanya.tan...@gmail.com wrote: My overall aim is to divert the traffic I want two

Re: [pox-dev] Over-writing table entries

Regards. On Thu, Dec 18, 2014 at 1:21 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: Please see the second, third, and final entries in the POX FAQ. I also think the second entry you're trying to install has problems (too many actions). -- Murphy On Dec 17, 2014, at 9:54 AM

Re: [pox-dev] Bideirectional Forwarding Detection(BFD) packets

I'm not that familiar with it, but I think BFD is configured entirely through ovsdb. In that case, you could do it out of band, or you could use the experimental POX OVSDB interface to do it. It's currently in my fork (github/MurphyMc/pox experimental, I think), but it will probably be pulled

Re: [pox-dev] Over-writing table entries

Please see the second, third, and final entries in the POX FAQ. I also think the second entry you're trying to install has problems (too many actions). -- Murphy On Dec 17, 2014, at 9:54 AM, chaitanya tanwar chaitanya.tan...@gmail.com wrote: Hi all, I am using the following code to

Re: [pox-dev] Issue with TimedRotatingFile Log Output

Just pushed what's hopefully a fix to the eel branch. -- Murphy On Dec 16, 2014, at 1:47 PM, A Sydney asydney...@gmail.com wrote: Hi POX folks, The backupCount parameter of the TimedRotatingFile log output does not function as expected. For example, from the link

Re: [pox-dev] Suggestions! Whats wrong with 2nd code?

getting. Can you identify something from them? diversion is my 2nd code which we are talking about. about attaching it. Thanks. On Mon, Dec 15, 2014 at 12:31 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: I'm not actually sure how the controller works at all in your example

Re: [pox-dev] Suggestions! Whats wrong with 2nd code?

Does your POX log have any warnings or errors in it? Those can be really informative. -- Murphy On Dec 14, 2014, at 9:56 PM, chaitanya tanwar chaitanya.tan...@gmail.com wrote: I have created a network using the following code, and pox as remote controller with l2.learning module.

Re: [pox-dev] Suggestions! Whats wrong with 2nd code?

to SDN. How can i check to POX logs? In mininet its just showing that POX is up and POX controller is connected to all the switches. Thanks. On Mon, Dec 15, 2014 at 11:53 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: Does your POX log have any warnings or errors in it? Those can

Re: [pox-dev] Suggestions! Whats wrong with 2nd code?

: But there is some problem due to my 2nd code becuase when I ran the pox without my 2nd code. All host were able to ping each other. Its when i installed a rule to the switch they stop pinging. On Mon, Dec 15, 2014 at 12:00 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: I don't know that. Seems like

Re: [pox-dev] Suggestions! Whats wrong with 2nd code?

? On Mon, Dec 15, 2014 at 12:14 PM, Murphy McCauley murphy.mccau...@gmail.com wrote: Yup. I'm sure there's something wrong with it. But trying to program POX without being able to read the log is like trying to do a puzzle in the dark. Figure out how to read the log. -- Murphy

Re: [pox-dev] how pox update counter in active flow table

No part of the POX framework tracks counter values. You can ask the switches to send results. Switches send the counters. You can see the counters the switches send. That's it. For more information on OpenFlow's counters, see the OpenFlow specification. -- Murphy On Dec 12, 2014, at

Re: [pox-dev] forwarding the traffic

the learning behavior and I just want to forward some of the traffic for one of the hosts to other host and the rest of the network should work with l2_learning. Regards, On 11/29/14 8:02 PM, Murphy McCauley wrote: This doesn't sound like learning behavior. Do you just want to set up the table

Re: [pox-dev] forwarding the traffic

This doesn't sound like learning behavior. Do you just want to set up the table by hand? Do that by just crafting a flow_mod for each direction of communication. Match on a destination MAC address, and include actions to change the dl_dst (possibly) and send out a specific port (definitely).

Re: [pox-dev] inserting specific rules to specific switch

You might find the OpenFlow in POX section of the manual helpful. Especially the first two subsections and the OpenFlow Messages/ofp_flow_mod subsubsection. -- Murphy On Nov 24, 2014, at 2:58 PM, Ali Reza Zamani az...@cs.rutgers.edu wrote: Hi all, I have a question regarding defining the

[pox-dev] New branch: POX eel

Just a note that POX has a new branch -- eel -- where I'll be pushing my own new development and possibly merging some of the stuff in my private branches and my public fork. In fact, it's already gotten some. Sadly, I don't think dart has been pushed on as much as I'd like before calling it

Re: [pox-dev] pox console with ipython

? ^ SyntaxError: invalid syntax ./pox.py py --completion also doesn't tab complete stuff. Cheers, Alison 2014-11-14 21:59 GMT-05:00 Murphy McCauley murphy.mccau...@gmail.com: At one point, I think I made some tweaks to make sure that it worked with ipython if you just run ipython pox.py

Re: [pox-dev] pox console with ipython

At one point, I think I made some tweaks to make sure that it worked with ipython if you just run ipython pox.py. Have you tried that? Maybe it's gotten broken -- I'll admit to not testing that regularly (ever :) ). Also if you run py --completion, you can often get some tab-completion with

Re: [pox-dev] traffic rate of a flow

Sorry, I guess I missed this until just now. But there's some discussion of the same topic in another thread which has been active today. -- Murphy On Nov 5, 2014, at 6:16 AM, farshad tajedin farshad.taje...@gmail.com wrote: hi how i can find traffic rate of a flow in pox? -- Best

Re: [pox-dev] Raising LinkEvents when openflow.discovery handles a port status removed

You're not off base. This has actually *almost* reached the mainline in the form of a fast_discovery component which subclasses/tweaks some stuff from discovery, but just never quite has. I wanted the code a bit cleaner than it was, and I think there were some subtleties I wanted to get right

Re: [pox-dev] Creating and Raising Custom Events

, and raise your two new event types on your component instead of on the OpenFlowNexus or Connection. Other components can listen there. Thanks for all your help so far, Adam Pavlidis On Tue, Nov 4, 2014 at 12:42 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: I'm a bit surprised

Re: [pox-dev] msg.match.tp_dst is None

You should probably take a look at the POX FAQ. The second and third questions probably directly relate. -- Murphy On Nov 4, 2014, at 11:30 PM, farshad tajedin farshad.taje...@gmail.com wrote: hi all i have a problem with add_flow entry in my switches. this is my code:

Re: [pox-dev] msg.match.tp_dst is None

? On Wed, Nov 5, 2014 at 11:03 AM, Murphy McCauley murphy.mccau...@gmail.com wrote: You should probably take a look at the POX FAQ. The second and third questions probably directly relate. -- Murphy On Nov 4, 2014, at 11:30 PM, farshad tajedin farshad.taje...@gmail.com wrote: hi all i

Re: [pox-dev] Creating and Raising Custom Events

I'm a bit surprised that you're getting no result at all as opposed to, say, exceptions in the POX log. Have you tried turning logging to DEBUG level and reading through to make sure there are no hints as to what might be going wrong? It's possible that raiseEventNoErrors() is masking the

  1   2   3   4   5   >