Re: [nox-dev] NOX-6 and OpenFlow 1.0: Do not connect...
I assume for by OF1.0, you are referring to the reference implementation at openflowswitch.org, correct? I'm fairly certain the guys at Stanford are using that codebase to work with NOX, so it should work. How are you starting NOX and the switches? You may want to try starting both with the "-v" flag to enable more verbose logging. --Justin On Feb 10, 2010, at 5:13 AM, Macapuna, Carlos A. B. wrote: > Hi Martin, > > Even doing: > git clone git://noxrepo.org/nox > cd nox > "git checkout -b openflow-1.0 origin/openflow-1.0" > ... > in the installation of NOX, I can not connect to OpenFlow 1.0. This works for > the Open vSwitch with support OF1.0, but not with the OF1.0. > > In NOX: > 00064|openflow|DBG:Passive tcp interface received connection > 00065|openflow|DBG:stream: negotiated OpenFlow version 0x01 (we support > versions 0x01 to 0x01 inclusive, peer no later than version 0x01) > 00066|nox|DBG:Success sending in 'sending switch config' > 00067|nox|DBG:Success sending in 'receiving features reply' > 00068|nox|WARN:stream: closing connection due to timeout after 5 seconds in > receiving features reply stat > > In virtual switch (OpenFlow VMS) with OF-1.0 user space: > hda:/cdrom# dpctl ofdatapath ofprotocol runme > hda:/cdrom# sh runme > > hda:/cdrom# Feb 10 04:53:49|00063|rconn|INFO|tcp:10.0.2.2:6633: connecting... > Feb 10 04:53:49|00064|rconn|INFO|tcp:10.0.2.2:6633: connected > Feb 10 04:53:54|00065|rconn|INFO|tcp:10.0.2.2:6633: connection closed by peer > Feb 10 04:53:54|00066|rconn|INFO|tcp:10.0.2.2:6633: waiting 15 seconds before > reconnect > Feb 10 04:53:55|00067|rconn|INFO|unix:/tmp/ofsw: connecting... > Feb 10 04:53:55|00068|vconn_unix|ERR|/tmp/vconn-unix.1174.8: connection to > /tmp/ofsw failed: No such file or directory > Feb 10 04:53:55|00069|rconn|WARN|unix:/tmp/ofsw: connection failed (No such > file or directory) > Feb 10 04:53:55|00070|rconn|INFO|unix:/tmp/ofsw: waiting 15 seconds before > reconnect > > > > Regards, > > -- > Carlos Macapuna > > ___ > nox-dev mailing list > nox-dev@noxrepo.org > http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Install multiple flows in one message
No, there is not. I'm not sure I understand the concern, though. There's maybe a few dozen bytes of overhead per message. Of the criticisms I've heard about OpenFlow, that particular bit of overhead has not been one of them. Have you seen it be a problem in practice? --Justin On Dec 10, 2009, at 8:07 PM, Theophilus Benson wrote: > Is it possible to install multiple distinct flow entries in one openflow > message. Right now, I am currently sending a message for each flow entry > that I want to install, however, I want to combine multiple flow entries in > one message in an attempt to reduce the network footprint. > > > Theo > > ___ > nox-dev mailing list > nox-dev@noxrepo.org > http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Specifying prefix and suffix matches
OpenFlow doesn't support suffix matching like you've described. Your best bet is to push your prefix match and then allow NOX to handle the suffix matching when packets miss in the switch. You can always push an exact match entry for the misses on-demand so that the performance isn't too bad for your suffix matching. --Justin On Dec 6, 2009, at 11:27 AM, "Arun K. Kumar" wrote: Hi, I am trying to install flow table entries on switches which outputs a packet on a specified port based on a matching with the dest ip. I need to specify two types of matches. the first set does a prefix match (eg. 10.2.1.0/24 should match an incoming packet's dest ip with the mask 255.255.255.0). If none of these match, the next set of entries do a suffix match (eg. 0.0.0.2/8 should match with mask 0.0.0.255). I want to install these entries on a switch when it comes up (datapath-join event). So I was wondering how I would mention these masks for the nw_dst field in the match structure. openflow.h mentions something about wildcard bit count and cidr.hh seems to provide some masking functions - do I use these, if so how? Thanks for your help. Regards, Arun ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Preparing for Nox 0.6.0
On Nov 4, 2009, at 11:22 AM, Martin Casado wrote: The goal is certainly for Nox 0.6 to support all 1.0 features. However, that will happen after wire protocol compatibility (shooting for this Friday ... though that might slip). I'd just like to point out that the final feature set for OpenFlow 1.0 has not yet been defined. We will make NOX compatible with the current 1.0 description as defined in the OpenFlow spec as it stands now. The final specification of OpenFlow 1.0 will likely be decided in early December. We should have a short turnaround when it's finalized, though. --Justin ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Modify Ehernet (src and/or dst)
On Sep 29, 2009, at 6:27 AM, Macapuna, Carlos A. B. wrote: What is the syntax to modify headers of a flow, Ethernet for example ( if dl_src==50:54:00:00:00:03, then dl_src=50:54:00:00:00:10)? The exact syntax will depend on whether you're doing it from Python or C++. In C++, you can find an example in "netapps/routing/ sprouting.py". In Python, you can find an example in "netapps/routing/ samplerouting.py". Search for "OFPAT_SET_DL_SRC" in either case. Out of curiosity, because I needed yesterday, has some way to add an action to rewrite some headers, without the use of NOX? To do it without NOX, you may want to look at the "dpctl" command that comes with the OpenFlow distribution. In particular, look for the "mod_dl_dst" action in the dpctl man page. --Justin ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
[nox-dev] FW: [openflow-discuss] Continuous disconnect/reconnect messages on HP-switch
Are you referring to the datapath id? If so, only the lower 48 bits are supposed to be used. This is from section 5.3.1 of the OpenFlow 0.8.9 spec describing the ofp_switch_features structure: uint64_t datapath_id; /* Datapath unique ID. Only the lower 48-bits are meaningful. */ If I'm misinterpreting your message, too, feel free ignore me. :) --Justin > From: Srinivasan Seetharaman > Date: Tue, 7 Jul 2009 06:44:02 -0700 > To: Nikhil Handigol > Cc: , , > > Subject: Re: [openflow-discuss] [nox-dev] Continuous disconnect/reconnect > messages on HP-switch > >> We are running a small experiment on a single switch topology with 2 hosts >> connected to the sole HP-switch. The nox controller is continuously getting >> a series of datapath join/leave messages from the HP-switch. Does anyone >> know the cause? > > I believe this is a consequence of the issue that NOX uses only least > significant 48 bits and individual HP virtual switches vary only in > the most significant bit. > ___ > openflow-discuss mailing list > openflow-disc...@lists.stanford.edu > https://mailman.stanford.edu/mailman/listinfo/openflow-discuss ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] secchan error: /usr/local/var/run/secchan.1281.ctl (No such file or directory)
Yes, we should clean some of this behavior up. Can you try running "make distclean" and then re-run "./configure" with the addition of the "--with-rundir=$HOME" argument? That will put all these runtime file in your home directory. Sorry about the problems. --Justin On Jun 18, 2009, at 12:14 AM, kk yap wrote: Hi, Apparently that option does not work either. qemu:~# /cdrom/secchan --pidfile=/root/testing.pid nl:0 tcp: 10.0.2.2:2525 Could not initialize control socket /usr/local/var/run/secchan.1091.ctl (No such file or directory) secchan: Could not listen for unixctl connections (No such file or directory) It seems to require the /usr/local/var/run/ directory and insist on it. FYI. Regards KK 2009/6/17 kk yap : Hi, For the record of those who search the mailing list, the option is -P, --pidfile[=FILE]create pidfile (default: /usr/local/var/run/secchan.pid) Regards KK 2009/6/17 Martin Casado : That's were secchan appears to be writing to. My guess is that if you were to build and install it would create that directory, but since you're just copying the binaries over, it doesn't exist. You might want to check for command line options to change that directory if you prefer. Hi, It does not. But why should it? New requirements? Regards KK 2009/6/17 Martin Casado : Stupid guess, but double check that /usr/local/var/run/ exists. Hi, I get this error when running qemu:~# /cdrom/secchan nl:0 tcp:10.0.2.2:2525 Could not initialize control socket /usr/local/var/run/secchan.1281.ctl (No such file or directory) This error is completely new to me. Can someone explain what this means? Regards KK ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Inputs on how to emulate new more complicated networks using NOX
You may also want to look at the man pages for more details. The one you would be most likely interested in is "man/man5/vms.conf.5". You may also want to look at "man/man1/start-test-vm.1". --Justin On Apr 2, 2009, at 8:13 AM, Keith Amidon wrote: The number of VMs and their connectivity are specified in the vms.conf file. As an example, here is a file with four VMs total, two acting as switches (vigil1 & vigil2) and two acting as regular machines (end1 and end2), one connected to each switch. To add more VMs, just add new names to the VMS line. The "SWITCHES" provide the connectivity for the VMs connected to them using the *_NETS lines as a regular (non-openflow) switch would. --- Keith SWITCHES="s1 s2 s3 s4 s5" s1_SLIRP=yes s2_SLIRP=yes VMS="vigil1 vigil2 end1 end2" end1_NETS="s4" vigil1_NETS="s1 s3 s4" end2_NETS="s5" vigil2_NETS="s2 s3 s5" On Thu, 2009-04-02 at 06:34 -0400, Surabhi Diwan wrote: Hi All, I did the virtual test environment setup using NOX with 1 controller, 1 switch and 2 virtual QEMU machines as suggested in the wiki. I want to build upwards from this configuration and have more nodes and switches under the purview of the NOX controller. Could you please suggest what would be the best way to emulate more complicated (close to real life networks) using the NOX controller. Thanks a lot! Regards, Surabhi ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Problem in Setting Up a Virtual Testing Environment
Back when Intrepid first came out, I didn't have any luck with their VDE package. Ben developed a patched version of VDE that works. He hasn't applied the patch to the latest version of VDE, so it trails the official Ubuntu version a bit. I haven't had any problems using it, though. Khan, if you'd like me to send you the patched 32-bit x86 package files, let me know. --Justin On Mar 13, 2009, at 10:18 AM, Peter Balland wrote: Hi Khan, I found the VDE packages on Ubuntu 8.04 to be horribly buggy. Can you try running with updated VDE packages (perhaps the ones from intrepid)? Peter khan.ferd...@upf.edu wrote: Hi, I tried to run nox and openflow on the virtual environment. I did the following steps very carefully. At first I did the openflow installation for Ubuntu. All were done successfully. I saw the virtual ethernet pairs also. I killed the avahi-daemon by "avahi- daemon -k". Then I did the NOX installation by "mkdir nox" at first, then I downloaded the noxcore from the git repository inside the nox directory as the nox website described. So I have "/nox/ noxcore" path. I used the "cd build/", "../configure --with- python=yes" at the first time, and all went fine. Then I followed the VM steps. But when I ran the "start-test-vm", it gave the same output as the last one: "*** buffer overflow detected ***: slirpvde terminated". I checked all packages again and they are fine. So I did not understand the problem. Next, I removed the build directory and followed the INSTALL file's instuctions under the noxcore directory, which does not require the build directory (quick steps I think). Here I did "./configure -- with-python=yes --with-openflow=/home/khan/openflow", but when I tried to do the "make" it gave the error in output which is attached below. Then I do "make distclean" and again create the build directory but now I do "cd build/", "../configure --with-python=yes --with- openflow=/home/khan/openflow", but when I try to "make" the same error appears. So whenever I tried to configure with openflow, it could not make. Without "--with-openflow=/home/khan/openflow", all steps were going fine but "start-test-vm" gave buffer overflow. There is another issue: I put the line "export PATH=$PATH:$HOME/nox/ noxcore/src/utilities/" inside the .bashrc and I could see the path correctly by "echo $PATH" but when I run "start-test-vm", it does not work. So I put this file in /bin directory and it works fine. Here I have described all the problems, so please help on this topic. /Khan Output: - k...@netcom:~/nox/noxcore$ make make all-recursive make[1]: Entering directory `/home/khan/nox/noxcore' Making all in src make[2]: Entering directory `/home/khan/nox/noxcore/src' make all-recursive make[3]: Entering directory `/home/khan/nox/noxcore/src' Making all in include make[4]: Entering directory `/home/khan/nox/noxcore/src/include' make[4]: Nothing to be done for `all'. make[4]: Leaving directory `/home/khan/nox/noxcore/src/include' Making all in lib make[4]: Entering directory `/home/khan/nox/noxcore/src/lib' make all-am make[5]: Entering directory `/home/khan/nox/noxcore/src/lib' /bin/bash ../../libtool --tag=CXX --mode=compile g++ - DHAVE_CONFIG_H -I. -I../.. -DPKGDATADIR=\"/usr/local/share/nox\" - DPKGLIBDIR=\"/usr/local/bin/lib\" -DPKGLOCALSTATEDIR=\"/usr/local/ var/nox\" -DPKGSYSCONFDIR=\"/usr/local/etc/nox\" -include ../../ config.h -I ../../src/include -I../../src/include/openflow -I/usr/ include -D_GNU_SOURCE=1 -D_REENTRANT -D__STDC_LIMIT_MACROS=1 - D__STDC_FORMAT_MACROS=1 -D__STDC_CONSTANT_MACROS=1 - D_GLIBCXX_CONCEPT_CHECKS=1 -D_GLIBCXX_DEBUG=1 - D_GLIBCXX_DEBUG_PEDANTIC=1 -I/usr/include/python2.5 -I../../src/ include/openflow -I../../src/nox -fno-omit-frame-pointer -Wall -Wno- sign-compare -Winit-self -Wformat-nonliteral -Wformat-security -g - O2 -MT async_file.lo -MD -MP -MF .deps/async_file.Tpo -c -o async_file.lo async_file.cc libtool: compile: Failed to create `.libs' make[5]: *** [async_file.lo] Error 1 make[5]: Leaving directory `/home/khan/nox/noxcore/src/lib' make[4]: *** [all] Error 2 make[4]: Leaving directory `/home/khan/nox/noxcore/src/lib' make[3]: *** [all-recursive] Error 1 make[3]: Leaving directory `/home/khan/nox/noxcore/src' make[2]: *** [all] Error 2 make[2]: Leaving directory `/home/khan/nox/noxcore/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/khan/nox/noxcore' make: *** [all] Error 2 k...@netcom:~/nox/noxcore$ echo $PATH /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/ games:/home/khan/nox/noxcore/src/utilities/ Please solve the problem. Thank you. /Khan ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox
Re: [nox-dev] pyswitch application output
my question is about last two lines (46,47). Why we are getting this arp reply at the end with out any request and both hosts are already seding ICMP packets to each other (by mentioning correct destination mac). Can anyone explain please. Hi, Siraj. I don't think there are any problems in the log you sent, but the behavior might look a little strange due to OpenFlow not distinguishing between ARP requests and replies for matching. The destination MAC address in line 30 is unknown, so NOX just floods it (and does not push down an OpenFlow entry). In line 38, NOX pushes down a rule telling the switch to forward ARP traffic from 10.0.0.1 to 10.0.0.2. This was due to a ARP reply, but OpenFlow does not distinguish between ARP requests and replies. Therefore, my guess is that there was an ARP request going from 10.0.0.1 to 10.0.0.2, that matched that rule. However, since a rule was never pushed down for ARP traffic going from 10.0.0.2 to 10.0.0.1, you see ARP reply in NOX but not the original request. After line 47, you shouldn't see any more ARP traffic between these two hosts at NOX until the flows expire. Does that make sense? --Justin ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] FLOW_MOD message ignored
Unfortunately, the description in the OpenFlow spec doesn't match the reference implementation. The OpenFlow v0.8.9 release notes do match the implementation, though: http://openflowswitch.org/wk/index.php/OpenFlow_0.8.9_release_notes#Modify_Actions_in_Existing_Flow_Entries The difference is that the spec states that if an OFPFC_MODIFY action doesn't match any entries, then it must be added. The implementation only modifies entries that it can find, it will never add them. According to the spec, KK's code should have worked regardless of which method he used. My personal opinion is that the reference implementation's behavior makes more sense, but this is a discussion for one of the OpenFlow dev lists. --Justin On Feb 15, 2009, at 10:54 PM, Brandon Heller wrote: Hi KK, Please see section 4.6 of the spec for the differences. Available at: http://www.openflowswitch.org/documents/openflow-spec-v0.8.9.pdf In short, modify keeps the associated flow counters, while add overwrites them. -b On Sun, Feb 15, 2009 at 10:23 PM, kk yap wrote: Hi, I managed to do the flow mod by changing the command the ADD instead of MODIFY. Can someone explain what is the difference? Thanks. Regards KK 2009/2/14 Justin Pettit : > Can you send us a either a pcap file or a printout of the bytes on the wire? > > --Justin > > > On Feb 14, 2009, at 5:31 AM, kk yap wrote: > >> Hi, >> >> I have sent a ofp_flow_mod message, which I verified to be correctly >> formed using a wireshark to a software reference switch. However, I >> do not see the flow entry when I do a dpctl dump-flows. The entry is >> made to be Permanent (both hard and idle timeout). Can anyone think >> of a reason why this is so? >> >> Thanks in advance. >> >> Regards >> KK >> >> ___ >> nox-dev mailing list >> nox-dev@noxrepo.org >> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org > > ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] send_openflow_packet command needs in port?
For the most part, it doesn't matter what input port you use. The only two exceptions in OpenFlow are: - If the output port is OFPP_FLOOD or OFPP_ALL, then it will be sent out every port but the input port. - If the input port is equal to the output port, the packet will be discarded. If you need that ability, then the output port must be OFPP_IN_PORT. The safest thing is probably to just set it to OFPP_NONE. --Justin On Feb 15, 2009, at 10:25 PM, kk yap wrote: Hi, I was trying to send an entire packet to some port on an openflow switch. I realized that the send_openflow_packet takes the in port as a parameter. Why is that even needed if I am sending an packet not stored in the buffer of the switch? In my case, where I am sending a packet from elsewhere, what should I put as the in port number? Any advice is appreciated. Thanks in advance. Regards KK ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] FLOW_MOD message ignored
Can you send us a either a pcap file or a printout of the bytes on the wire? --Justin On Feb 14, 2009, at 5:31 AM, kk yap wrote: Hi, I have sent a ofp_flow_mod message, which I verified to be correctly formed using a wireshark to a software reference switch. However, I do not see the flow entry when I do a dpctl dump-flows. The entry is made to be Permanent (both hard and idle timeout). Can anyone think of a reason why this is so? Thanks in advance. Regards KK ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] 0.4.0 up at noxrepo
I just wanted to add that this release uses the official 0.8.9 release of OpenFlow available here: http://www.openflowswitch.org/wp/downloads/ --Justin On Dec 11, 2008, at 5:54 PM, Martin Casado wrote: Hi all, We've updated noxrepo with nox 0.4.0. It is available both from git and as a downloadable source tarfile. The manual has not been updated, however very little of the core functionality has changed. Let us know if you have any problems accessing the source or building it. thanks, Martin ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Openflow Error
My guess is that your controller is running v0.8.2, but your switch is running the version from noxrepo.org (v0.9.0~b1). You can check this by running both the switch and controller with the "--version" flag. If they are running v0.8.2, you should see something like: [EMAIL PROTECTED] openflow-0.8.2] controller/controller --version controller/controller v0.8.2 compiled Nov 24 2008 14:51:59 [EMAIL PROTECTED] openflow-0.8.2] switch/switch --version switch/switch v0.8.2 compiled Nov 24 2008 14:52:01 Obviously, they should both have "v0.8.2" as their version if you're using the "official" OpenFlow. You can check the version number of the sources in "include/ openflow.h". For v0.8.2, you'll see this: #define OFP_VERSION 0x85 For v0.9.0~b1, you'll see this: #define OFP_VERSION 0x91 To get them both to communicate together, you'll need both sides to be running the same version. We're hoping most of these version problems go away with the OpenFlow Consortium's release of v0.8.9 release either this week or next. This should be a fairly stable release in terms of the wire protocol. --Justin On Nov 24, 2008, at 3:26 AM, Adil Razzaq wrote: Hi I was trying to connect a "user space openflow switch" with the "openflow controller(sample controller provided with the openflow switch)" running on two separate machines on Debian Linux. But i was getting the following error message on the controller Received openflow version 91 != expected 85 Can you please guide me why this error occurred and how to resolve it? I ran following command on the controller controller ptcp: & I used the following command on the switch switch tcp:ip_of_controller -i eth1,eth2 Note: I am using openflow version 0.8.2 for both controller and switch. Best Regards Adil Razzaq ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] openflow switch issue regarding kernel space installation
Hi, Siraj. Did you type "--with-l26=/path/to/linux-2.6/" in directly or did you give it the correct path to your kernel headers? On Debian, you'd probably want to run it like this: ./configure --with-l26=/lib/modules/`uname -r`/build The OpenFlow wiki has full instructions for Debian: http://openflowswitch.org/wk/index.php/Debian_Install By the way, this mailing list is really focused on NOX, which is a controller for OpenFlow. There are a lot of OpenFlow developers on this mailing list, but questions specific to OpenFlow should probably go to them. It doesn't appear that there's a general mailing list at openflowswitch.org, so I'll ping them about that... Also, you'll need a newer version of OpenFlow than is available on openflowswitch.org. I'd recommend the following: http://noxrepo.org/releases/openflow-0.9.0~b1.tar.gz In the very near future, the OpenFlow consortium will be releasing v0.8.9, which will be compatible with NOX. Until then, you'll need to use the tarball shown above. Good luck. Let me know if you continue to have issues. --Justin On Nov 23, 2008, at 2:14 PM, siraj rathore wrote: Dear Sir I am PhD student at Royal Institute of technology Sweden. I am interested to learn about openflow switches and to do research in this field. I am facing problem regarding Kernel based open flow installation. i am using Debian unstable kernel 2.6.26. To build open flow switch in kernel space i executed following commands according to instructions given on the nox and open flow websites. ( i am using openflow version 0.8.2) ./configure --with-l26=/path/to/linux-2.6/ make In result following binaries are created openflow-v0.8.2/secchan/secchan openflow-v0.8.2/utilities/dpctl openflow-v0.8.2/controller/controller But openflow-v0.8.2/datapath/linux-2.6/openflow_mod.ko "openflow_mod.ko" is not present there. Therefore i could not load kernel space switch. Can you guide me to solve this issue. Regards Siraj ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] OpenFlow switch connection to NOX
Ah, right. Can you try running seccan with "--no-stp"? --Justin On Sep 23, 2008, at 2:44 PM, kk yap wrote: > Hi, > > I am using the public repo, with hash > d7886e16ee8a735e7e327ea2007455f0bd903657. > And so, there is no update in that repo. > > Regards > KK > > 2008/9/23 Justin Pettit <[EMAIL PROTECTED]>: >> We had a couple of problems related to STP and ports changing state >> that >> were fixed this morning. Would you be able to pull again? If it >> still >> doesn't work, can you send the commit number for both OpenFlow and >> NOX? You >> can see it by running "git log" and looking at the first line, which >> contains a SHA-1 hash. (Since we're in active development, the >> version is >> generally useful information in these bug reports.) >> >> Thanks! >> >> --Justin >> >> >> On Sep 23, 2008, at 2:07 PM, kk yap wrote: >> >>> Hi, >>> >>> In the VM, using >>> /cdrom/secchan nl:0 tcp:10.0.2.2:2525 >>> >>> Btw, I tried both the latest vde and Ben's fixed version. >>> >>> Regards >>> KK >>> >>> >>> 2008/9/23 Martin Casado <[EMAIL PROTECTED]>: >>>> >>>> That shouldn't be the case: >>>> >>>> ./secchan nl:0 tcp:1.2.3.4:2525 >>>> >>>> should sit and poll for NOX without exiting. Can you post the >>>> command >>>> line >>>> you are using? >>>> >>>> On Sep 23, 2008, at 1:57 PM, kk yap wrote: >>>> >>>>> Hi, >>>>> >>>>> I have the latest version of NOX with the VM environment. I >>>>> find that >>>>> I cannot start the OpenFlow switch before starting NOX. Is this >>>>> behavior intentional? I would prefer to cut my NOX and restart it >>>>> anytime without re-running the switch. >>>>> >>>>> Regards >>>>> KK >>>>> >>>>> ___ >>>>> nox-dev mailing list >>>>> nox-dev@noxrepo.org >>>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >>>> >>>> >>> >>> ___ >>> nox-dev mailing list >>> nox-dev@noxrepo.org >>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >> >> ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] OpenFlow switch connection to NOX
We had a couple of problems related to STP and ports changing state that were fixed this morning. Would you be able to pull again? If it still doesn't work, can you send the commit number for both OpenFlow and NOX? You can see it by running "git log" and looking at the first line, which contains a SHA-1 hash. (Since we're in active development, the version is generally useful information in these bug reports.) Thanks! --Justin On Sep 23, 2008, at 2:07 PM, kk yap wrote: > Hi, > > In the VM, using > /cdrom/secchan nl:0 tcp:10.0.2.2:2525 > > Btw, I tried both the latest vde and Ben's fixed version. > > Regards > KK > > > 2008/9/23 Martin Casado <[EMAIL PROTECTED]>: >> That shouldn't be the case: >> >> ./secchan nl:0 tcp:1.2.3.4:2525 >> >> should sit and poll for NOX without exiting. Can you post the >> command line >> you are using? >> >> On Sep 23, 2008, at 1:57 PM, kk yap wrote: >> >>> Hi, >>> >>> I have the latest version of NOX with the VM environment. I find >>> that >>> I cannot start the OpenFlow switch before starting NOX. Is this >>> behavior intentional? I would prefer to cut my NOX and restart it >>> anytime without re-running the switch. >>> >>> Regards >>> KK >>> >>> ___ >>> nox-dev mailing list >>> nox-dev@noxrepo.org >>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >> >> > > ___ > nox-dev mailing list > nox-dev@noxrepo.org > http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] IPv6 support in NOX
NOX uses the OpenFlow protocol (http://openflowswitch.org) to control switches. At this time, there is no support for IPv6 in OpenFlow. Sorry. --Justin On Sep 11, 2008, at 11:40 PM, Magesh mannar wrote: > Dear all > > Recently I came across the Nox Network control platform. It seems to > be interested. However, I wondering whether the current > implementation supports IPv6 addresses. > > Does Nox support IPv6 address. > > Kindly reply please > > Many thanks > > Magesh > > > > > ___ > nox-dev mailing list > nox-dev@noxrepo.org > http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] [EMAIL PROTECTED] download error
Hmm, not sure. We typically go to Ben for these sorts of issues. But attached is the latest openflow.h. --Justin openflow.h Description: Binary data On Sep 5, 2008, at 10:52 AM, Brandon Heller wrote: heller-laptop:OpenFlow brandonh$ git clone git://noxrepo.org/openflow Initialized empty Git repository in /Users/brandonh/Stanford/ Research/OpenFlow/o remote: Generating pack... remote: Done counting 3756 objects. remote: error: remote: unable to find 5e70c38c78be569f075ec4d8f6e224f0d2e3df53re remote: fatal: remote: unable to get type of object 5e70c38c78be569f075ec4d8f6er remote: aborting due to possible repository corruption on the remote side. fatal: early EOF fatal: index-pack failed I get the same error when trying to pull or clone. Any ideas? If not, could someone send me the most recent openflow.h file? Thanks, -Brandon ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Port mod python example?
Here's the definition for the flags from "openflow.h": /* Flags to indicate behavior of the physical port */ enum ofp_port_flags { OFPPFL_NO_FLOOD = 1 << 0, /* Do not include this port when flooding */ }; So by default, the flags should be off. If you do not want to include them when flooding, then they should be enabled. I think things are working properly, but maybe I'm misunderstanding something here... --Justin On Aug 9, 2008, at 1:22 PM, Glen Gibb wrote: > Hi Justin, > > It seems that the OpenFlow reference implementation is incorrectly > reporting the flags field for ofp_phy_port in the features reply (or > I'm mis-interpreting what this should be). > > Looking at a packet capture for the features reply (see attached > packet) it always reports 0 in the flags -- I would expect this to > be 1 since by default all ports participate in flooding. > > Glen > > > > Glen Gibb wrote: >> I'm looking to set the flood flag so that when I have loops in my >> topology I don't end up looping packets forever. >> >> I've actually written most of the code -- I'm just having trouble >> right now copying the name into the name field from Python. >> Actually, do I need to send the name back to the OF switch or will >> it work without setting a name? >> >> Here's the one snippet of code I haven't yet gotten working. In >> openflow.i I have: >> >> %extend ofp_phy_port { >> void set_hw_addr(uint8_t *addr) { >> memcpy($self->hw_addr, addr, ethernetaddr::LEN); >> } >> void set_name(uint8_t *name) { >> memcpy($self->name, name, OFP_MAX_PORT_NAME_LEN); >> } >> }; >> >> (I'm basically trying to copy what you had in that file for >> ofp_match.) >> >> The set_hw_addr works perfectly, the set_name fails because I >> haven't yet managed to convert my string to a uint8_t array. Is >> there an easy way to pass in a python string to set_name and copy >> from the Python string to the name field? >> >> Glen >> >> >> >> >> Justin Pettit wrote: >>> Glen, what are you looking to modify? Currently, you are only >>> able to set the port flags over OpenFlow and the only flag is one >>> to indicate whether the port should be included in "flood" >>> actions. We're looking to include actions such as disabling >>> ports, but that's not in there yet. >>> >>> --Justin >>> >>> >>> On Aug 9, 2008, at 12:36 PM, Martin Casado wrote: >>> >>>> I'm not sure port mod is supported in the OF reference >>>> implementation. Justin/Ben? >>>> >>>> On Aug 8, 2008, at 4:42 PM, Glen Gibb wrote: >>>> >>>>> Hi all, >>>>> >>>>> Is there an example available that sends port mod messages to a >>>>> switch? >>>>> I'm currently looking for examples but can't find one. >>>>> >>>>> Thanks, >>>>> Glen >>>>> >>>>> ___ >>>>> nox-dev mailing list >>>>> nox-dev@noxrepo.org >>>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >>>> >>>> >>>> ___ >>>> nox-dev mailing list >>>> nox-dev@noxrepo.org >>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >>> >>> >> >> > > ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Port mod python example?
The name doesn't matter. Just make sure you have the Ethernet address set properly, since OpenFlow performs a sanity-check on that. --Justin On Aug 9, 2008, at 12:53 PM, Glen Gibb wrote: > I'm looking to set the flood flag so that when I have loops in my > topology I don't end up looping packets forever. > > I've actually written most of the code -- I'm just having trouble > right now copying the name into the name field from Python. > Actually, do I need to send the name back to the OF switch or will > it work without setting a name? > > Here's the one snippet of code I haven't yet gotten working. In > openflow.i I have: > > %extend ofp_phy_port { > void set_hw_addr(uint8_t *addr) { > memcpy($self->hw_addr, addr, ethernetaddr::LEN); > } > void set_name(uint8_t *name) { > memcpy($self->name, name, OFP_MAX_PORT_NAME_LEN); > } > }; > > (I'm basically trying to copy what you had in that file for > ofp_match.) > > The set_hw_addr works perfectly, the set_name fails because I > haven't yet managed to convert my string to a uint8_t array. Is > there an easy way to pass in a python string to set_name and copy > from the Python string to the name field? > > Glen > > > > > Justin Pettit wrote: >> Glen, what are you looking to modify? Currently, you are only able >> to set the port flags over OpenFlow and the only flag is one to >> indicate whether the port should be included in "flood" actions. >> We're looking to include actions such as disabling ports, but >> that's not in there yet. >> >> --Justin >> >> >> On Aug 9, 2008, at 12:36 PM, Martin Casado wrote: >> >>> I'm not sure port mod is supported in the OF reference >>> implementation. Justin/Ben? >>> >>> On Aug 8, 2008, at 4:42 PM, Glen Gibb wrote: >>> >>>> Hi all, >>>> >>>> Is there an example available that sends port mod messages to a >>>> switch? >>>> I'm currently looking for examples but can't find one. >>>> >>>> Thanks, >>>> Glen >>>> >>>> ___ >>>> nox-dev mailing list >>>> nox-dev@noxrepo.org >>>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >>> >>> >>> ___ >>> nox-dev mailing list >>> nox-dev@noxrepo.org >>> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org >> >> > ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org
Re: [nox-dev] Port mod python example?
Glen, what are you looking to modify? Currently, you are only able to set the port flags over OpenFlow and the only flag is one to indicate whether the port should be included in "flood" actions. We're looking to include actions such as disabling ports, but that's not in there yet. --Justin On Aug 9, 2008, at 12:36 PM, Martin Casado wrote: > I'm not sure port mod is supported in the OF reference > implementation. Justin/Ben? > > On Aug 8, 2008, at 4:42 PM, Glen Gibb wrote: > >> Hi all, >> >> Is there an example available that sends port mod messages to a >> switch? >> I'm currently looking for examples but can't find one. >> >> Thanks, >> Glen >> >> ___ >> nox-dev mailing list >> nox-dev@noxrepo.org >> http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org > > > ___ > nox-dev mailing list > nox-dev@noxrepo.org > http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org ___ nox-dev mailing list nox-dev@noxrepo.org http://noxrepo.org/mailman/listinfo/nox-dev_noxrepo.org