Re: [ovs-discuss] IPv4 address support for port configuration in OVS 2.8.0 version

[Ben Pfaff]

On Tue, Sep 26, 2017 at 06:36:50AM +, Ranjith Kumar D wrote:
> Hello All,
> 
> I had recently downloaded the OVS 2.8.0 version code and was
> performing the code walkthrough for ovs-vsctl command handling for
> adding port to bridge. I see that structures related to port
> configuration takes only IPv6 address. Hence kindly let me know
> whether we will be able to set the interface of a port in a bridge
> with IPv4 remote and local IP.
> 
> 
> ovs-vsctl set interface  type=gre/rte
> option:remote_ip=172.20.1.1 option:local_ip=172.24.1.0

IPv4 addresses work and always have.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] ovn-trace not showing full logical flow

[Ben Pfaff]

On Sun, Sep 17, 2017 at 05:03:19PM +0530, Vikrant Aggarwal wrote:
>  0. ls_in_port_sec_l2 (ovn-northd.c:2979): inport ==
> "4c72cee2-35b7-4bcd-8c77-135a22d16df1" && eth.src == {fa:16:3e:55:3f:be},
> priority 50, uuid b6b8d57a
> next;
>  1. ls_in_port_sec_ip (ovn-northd.c:2113): inport ==
> "4c72cee2-35b7-4bcd-8c77-135a22d16df1" && eth.src == fa:16:3e:55:3f:be &&
> ip4.src == {10.10.10.4}, priority 90, uuid ba02f466
> next;
>  3. ls_in_pre_acl (ovn-northd.c:2397): ip, priority 100, uuid 25d55e7b
> reg0[0] = 1;
> next;
>  5. ls_in_pre_stateful (ovn-northd.c:2515): reg0[0] == 1, priority 100,
> uuid b84a160f
> ct_next;
> *** ct_* actions not implemented

ovn-trace in Open vSwitch 2.7 doesn't support the ct_* actions, which
means that tracing through a distributed firewall tends to end up this
way.

ovn-trace in Open vSwitch 2.8 does support these actions, so you'll see
more success there.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

[ovs-discuss] Open vSwitch Fall 2017 Conference: Call for Participation and Registration

[Ben Pfaff]

[Sorry about the repost]

The Open vSwitch project, a Linux Foundation Collaborative Project,
will host its fourth annual conference focused on Open vSwitch and OVN
on November 16 and 17, 2017, at Club Auto Sport in San Jose,
California.

We are seeking long and short ("lightning") talks on topics related to
Open vSwitch and OVN.  We expect long talks to last 20 minutes with an
additional 5 minutes for questions, and short talks to last 5 minutes.

Topics that may be considered, among others, include:

 * The future of Open vSwitch (e.g., P4 and eBPF).

 * NAT, DPI, and stateful processing with Open vSwitch. 

 * Deploying and using OVN.

 * Testing and scaling OVN.

 * NIC acceleration of Open vSwitch.

 * Using Open vSwitch to realize NFV and service chaining.

 * Porting Open vSwitch to new operating systems, hypervisors,
   or container systems.

 * Integrating Open vSwitch into larger systems.

 * Troubleshooting and debugging Open vSwitch installations.

 * Open vSwitch development and testing practices.

 * Performance measurements or approaches to improving
   performance.

 * End-user or service provider experiences with Open vSwitch.

 * Hardware ports of Open vSwitch (existing, in progress, or
   speculative).

 * The relationship between OpenFlow and Open vSwitch.

 * Using, developing, or administering OpenFlow controllers in
   conjunction with Open vSwitch.

 * Comparisons to other implementations of features found in
   Open vSwitch (e.g. other OpenFlow implementations, other
   software switches, etc.).

 * Increasing the size and diversity of the Open vSwitch user
   and developer base.

 * Tutorials and walkthroughs of use cases.

 * Demos.

Talks will be recorded and made available online.


How to propose a talk
-

This year, we are soliciting proposals for full talks and short
("lightning") talks in a single round.  If you request a full talk, we
will also consider it for a lightning talk in the event that we cannot
accept it as a full talk.

We will also accept proposals for panel discussions.  Please submit
them as full talks and make it clear in the description that it is a
panel.

Please submit proposals to to the following URL by September 29:

https://goo.gl/forms/erjZIQICXVjfUh863

Speakers will be notified of acceptance by October 6.

Speakers should plan to attend the event in person.  Travel to and
accommodations in San Jose are the responsibility of attendees.


How to attend
-

To partially cover expenses, we expect to charge $200 for
registration.  We offer complimentary registration to speakers,
students, academics, and anyone for whom the registration fee is
burdensome.  Please contact us if you need any help obtaining a
complimentary registration.

To register, visit https://www.regonline.com/openvswitch2017.


Related Event: DPDK Summit
--

The Open vSwitch conference, on Nov. 16-17, follows the DPDK North
America Summit on Nov. 14-15, in the same venue.  The CFP for the DPDK
Summit is open until Oct. 4:
http://events.linuxfoundation.org/events/dpdk-north-america-summit/program/cfp


Important URLs
--

CFP: https://goo.gl/forms/erjZIQICXVjfUh863
general: https://events.linuxfoundation.org/events/open-vswitch-fall-event
registration: https://www.regonline.com/openvswitch2017


More information


To reach the organizers, email ovs...@openvswitch.org.  For general
discussion of the conference, please use the ovs-discuss mailing list
at disc...@openvswitch.org.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Compilation error : lib/netdev-dpdk.c:55:28: fatal error: rte_virtio_net.h: No such file or directory

[Ben Pfaff]

I'm confused, then.  How did you successfully link DPDK 2.0 to OVS 2.4.0
without first compiling it?  Why is OVS 2.8 mentioned?

On Thu, Sep 14, 2017 at 06:05:52AM +, Ranjith Kumar D wrote:
> Hi Ben Pfaff,
> 
> I am building with OVS 2.4.0.
> 
> Regards,
> Ranjith 
> 
> -Original Message-
> From: Ben Pfaff [mailto:b...@ovn.org] 
> Sent: Thursday, September 14, 2017 11:33 AM
> To: Ranjith Kumar D <ranjith.ku...@radisys.com>
> Cc: disc...@openvswitch.org
> Subject: Re: [ovs-discuss] Compilation error : lib/netdev-dpdk.c:55:28: fatal 
> error: rte_virtio_net.h: No such file or directory
> 
> On Thu, Sep 14, 2017 at 05:58:07AM +, Ranjith Kumar D wrote:
> > Hello All,
> > 
> > We have GTP packet handling functionality with OVS 2.4.0  and it's 
> > very hard to merge changes to OVS 2.8.0
> > 
> > I have installed DPDK 2.0 and successfully linked with OVS 2.4.0, but 
> > compilation is failing with below error:
> > 
> > lib/netdev-dpdk.c:55:28: fatal error: rte_virtio_net.h: No such file 
> > or directory
> 
> It isn't clear whether you're building OVS 2.4 or 2.8, but OVS 2.8 works with 
> DPDK 17.05.1 (not 2.0).
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Dependenci in debian/control

[Ben Pfaff]

On Wed, Sep 13, 2017 at 10:10:37PM -0700, Ben Warren via discuss wrote:
> Hi Wolfgang,
> 
> > On Sep 13, 2017, at 3:48 AM, Wolfgang Link  wrote:
> > 
> > The new package libopenvswitch depends on libssl-dev.
> > Is this intentionally made or is libssl1.1 also good?
> > 
> This is intentional.  In general 'libfoo-dev' packages contain header files 
> and symlinks to shared libraries, while ‘libfoo' packages contain the actual 
> shared libraries and other binaries.  Also, in general, the 'libfoo-dev' 
> packages have a dependency on the ‘libfoo' libraries, so they will be 
> installed automatically by apt or dpkg.
> 
> In this case, building the ‘libopenvswitch' package involves compiling and 
> linking OpenVswitch code and also some OpenSSL stuff (I believe for TLS), so 
> the header files are necessary.  You’ll get compilation errors without 
> libssl-dev.  You may not notice this if compiling for your host environment, 
> since it may already have the header files in place, but it will quickly 
> become apparent if you try to cross-compile.

I think it might be a mistake, actually.  Shouldn't only
libopenvswitch-dev depend on libssl-dev, not libopenvswitch?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Compilation error : lib/netdev-dpdk.c:55:28: fatal error: rte_virtio_net.h: No such file or directory

[Ben Pfaff]

On Thu, Sep 14, 2017 at 05:58:07AM +, Ranjith Kumar D wrote:
> Hello All,
> 
> We have GTP packet handling functionality with OVS 2.4.0  and it's very hard 
> to merge changes to OVS 2.8.0
> 
> I have installed DPDK 2.0 and successfully linked with OVS 2.4.0, but 
> compilation is failing with below error:
> 
> lib/netdev-dpdk.c:55:28: fatal error: rte_virtio_net.h: No such file or 
> directory

It isn't clear whether you're building OVS 2.4 or 2.8, but OVS 2.8 works
with DPDK 17.05.1 (not 2.0).
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How to submit a bug report

[Ben Pfaff]

I didn't realize that there was different information on the mailing
list and in ovs-issues.  Thanks for adding the missing information.

"git bisect" would probably take 10 tries or less to find the commit, so
I'd still like to encourage you to try that approach.

On Mon, Sep 11, 2017 at 09:27:46PM +, Llorente Santos Jesus wrote:
> Hi Ben, thank you for getting back to me on this.
> 
> I noticed that my previous email with the steps to reproduce was sent to 
> ovs-dev mailing list. It somewhat felt like the appropriate place to send 
> such detailed issue.
> In any case, I have added those same steps to the open issue in the 
> ovs-issues repo. I hope that simplifies the follow up!
> 
> I spent quite bit of time looking at the code but I couldn't really see 
> anything odd. Comparing  2.5.3 with 2.7.0, showed a vast number of commits. I 
> deepened into the tunnelling code for GRE, VXLAN and Geneve but couldn't stop 
> the difference. It's quite odd, in my experience, marking packets does not 
> invalidate in any case receiving them into the socket.
> 
> The issue here is that 2.5.3 is the current LTS version, which in late June 
> received a new update. The release note says "Bug fixes", so maybe there is 
> still something we can fix there...
> 
> Best,
> Jesus
> 
> 
> -Original Message-
> From: Ben Pfaff [mailto:b...@ovn.org] 
> Sent: 12 September 2017 00:03
> To: Llorente Santos Jesus <jesus.llorente.san...@aalto.fi>
> Cc: ovs-discuss@openvswitch.org
> Subject: Re: [ovs-discuss] How to submit a bug report
> 
> On Mon, Sep 11, 2017 at 07:27:56PM +, Llorente Santos Jesus wrote:
> > I have been seeing buggy behaviour with the tunnelling ports of OvS. I 
> > would like to ask some help in how to properly submit a bug report I 
> > created an issue in the git "ovs-issues" 
> > (https://github.com/openvswitch/ovs-issues/issues/134) about 2 weeks ago, 
> > but it seems no one is monitoring that.
> > I sent an email to  b...@openvswitch.org  following these instructions 
> > (http://docs.openvswitch.org/en/latest/internals/bugs/)
> > I have also sent an email to this mailing list, but no one has responded my 
> > message.
> > 
> > Can anyone help me? I would really like to get this annoying bug fixed!
> 
> Your report has definitely been recorded.  I can think of a few reasons why 
> no one has acted on it, besides simply being busy.  One is that it does not 
> give readers a good way to reproduce the problem, even though it talks about 
> a very particular scenario.  Developers are more likely to try to reproduce 
> and fix a bug when it doesn't require a lot of forethought on how to get 
> started.  Another is that it sounds like the problem has been fixed in a 
> later version.  It's not too rewarding to fix bugs that are already fixed.
> 
> Given that it's been fixed, you could use "git bisect" to try to find out 
> what commit fixed it.  If you reported that, it could allow a developer to 
> understand and fix the problem without much additional work.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How to submit a bug report

[Ben Pfaff]

On Mon, Sep 11, 2017 at 07:27:56PM +, Llorente Santos Jesus wrote:
> I have been seeing buggy behaviour with the tunnelling ports of OvS. I would 
> like to ask some help in how to properly submit a bug report
> I created an issue in the git "ovs-issues" 
> (https://github.com/openvswitch/ovs-issues/issues/134) about 2 weeks ago, but 
> it seems no one is monitoring that.
> I sent an email to  b...@openvswitch.org  following these instructions 
> (http://docs.openvswitch.org/en/latest/internals/bugs/)
> I have also sent an email to this mailing list, but no one has responded my 
> message.
> 
> Can anyone help me? I would really like to get this annoying bug fixed!

Your report has definitely been recorded.  I can think of a few reasons
why no one has acted on it, besides simply being busy.  One is that it
does not give readers a good way to reproduce the problem, even though
it talks about a very particular scenario.  Developers are more likely
to try to reproduce and fix a bug when it doesn't require a lot of
forethought on how to get started.  Another is that it sounds like the
problem has been fixed in a later version.  It's not too rewarding to
fix bugs that are already fixed.

Given that it's been fixed, you could use "git bisect" to try to find
out what commit fixed it.  If you reported that, it could allow a
developer to understand and fix the problem without much additional
work.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] ovsdb-idl.run-msg many update msgs about br-ex and controller

[Ben Pfaff]

On Mon, Sep 11, 2017 at 04:27:45PM +0800, 顾兆伦 wrote:
> Hi all, sorry to bother you.
> 
> When I use python/ovs, idl.run() always costs about 50 ms to execute, and 
> it runs 50times, what is too long for me. When I check the 
> msg(msg=self._sesson.recv()), I found this:
> 
[snip almost 2000 lines]
>
>   I want to know why idl.run() recv so many update msgs about br-ex and 
> controller, it’s because of ‘tx_packets’ changed? How to avoid this so I can 
> take less time 

Usually these fields should only update about every 5 seconds.  Perhaps
some bug is updating them too often.

However, if you don't need particular tables or columns, you can simply
not monitor those.  That is more efficient regardless of how often those
tables or columns update.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How to build openvswitch with Ddpdk enabled debian package (openvswitch-switch-dpdk.2.7.2-1.deb) for OVS ver. 2.7.2

[Ben Pfaff]

On Mon, Sep 11, 2017 at 11:25:59PM +1200, Brad Cowie wrote:
> On 10 September 2017 at 03:17, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Sat, Sep 09, 2017 at 02:26:04AM +, Chou, David J wrote:
> > > Hi,
> > >
> > > I tried to build openvswitch with dpdk enabled debian package
> > (openvswitch-switch-dpdk.2.7.2-1.deb)  for OVS ver. 2.7.2.
> >
> > The Debian packaging that comes with OVS doesn't build a DPDK-based
> > switch.
> >
> 
> The folks at Ubuntu do provide OVS packaging that includes DPDK:
> https://git.launchpad.net/~ubuntu-server-dev/ubuntu/+source/openvswitch

Oh.  I'm surprised they haven't upstreamed them.  Wonder why.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [OVS-Discuss]: configure: error: cannot link with dpdk

[Ben Pfaff]

On Fri, Sep 08, 2017 at 01:33:35PM +, Ranjith Kumar D wrote:
> Hi,
> 
> I am trying to install OVS with DPDK and I getting below error while 
> configuring OVS
> 
> configure: error: cannot link with dpdk
> 
> The below are my OVS and DPDK software  version details:
> 
> DPDK version : dpdk-stable-17.02.1
> OVS version :  rte-openvswitch-2.4.0

It sounds like you're building the Intel-initiated fork of OVS based on
DPDK.  This isn't the right mailing list to discuss that fork.  Also, to
the best of my knowledge, that fork has not been under active
development for some time.

I suggest that you use mainline OVS 2.8.0 (or master) instead of the
fork.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How to build openvswitch with Ddpdk enabled debian package (openvswitch-switch-dpdk.2.7.2-1.deb) for OVS ver. 2.7.2

[Ben Pfaff]

On Sat, Sep 09, 2017 at 02:26:04AM +, Chou, David J wrote:
> Hi,
> 
> I tried to build openvswitch with dpdk enabled debian package 
> (openvswitch-switch-dpdk.2.7.2-1.deb)  for OVS ver. 2.7.2.
> 
> I downloaded openvswitch-2.7.2.tar.gz from http://openvswitch.org/download, 
> and dpdk-stable-16.11.2.targ.gz from 
> http://dpdk.org/browse/dpdk-stable/tag/?h=v16.11.2. I could build and install 
> dpdk and openvswitch with dpdk enabled by the "make install" method on Ubuntu 
> 16.04 LTS, and verify they could work together.
> 
> Also, by getting dpdk_16.11.2.orig.tar.gz and dpdk_16.11.2-4.debian.tar from 
> https://packages.debian.org/source/sid/dpdk , I  could build dpdk-16.11.2-4 
> debian package and I verified that this dpdk_16.11.2-4 debian package work.
> 
> Also, I  could build openvswitch without dpdk enabled debian package by 
> following the instruction in debian.rst in openvswith _2.7.2.   But when I 
> tried to build openvswitch with dpdk enabled by  following:
> 
> 1.   Install dpdk_16.11.2-4 debian package I built in the my building 
> system
> 
> 2. export DATAPATH_CONFIGURE_OPTS="--with-dpdk=/usr"
> Then build openvswitch debian package again, I saw some libdpdk built, and 
> the build completed, but openvswitch-switch-dpdk.2.7.2-1.deb wasn't built 
> while other openswitch debian package built.
> What do I miss?  It seems to me that the debian spec (debian sub-directory  
> in openvswith 2.7.2) doesn't have all info necessary to build 
> openvswitch-switch-dpdk.2.7.2-1.deb, am I right? How to fix this?

The Debian packaging that comes with OVS doesn't build a DPDK-based
switch.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Question about megaflow

[Ben Pfaff]

On Thu, Aug 31, 2017 at 11:05:34PM -0400, Heyang Wang wrote:
> I have several questions about implementation of megaflow in OVS:
> 
> 1. I want to make sure that megaflow is not supporting priority, so it will
> not cache the overlapping rules in userspace, am I right?
> 
> 2. if overlapping rule will not be added to megaflow cache, what if a new
> rule is adding to the userspace, which will overlap with the rules already
> in the megaflow cache already, what will OVS do at this time?
> 
> 3. Will you have a good solution to solving overlapping rules problem in
> megaflow cache in next version of OVS?

There seems to be an assumption here that overlapping flows cause some
kind of problem for megaflows.  They don't.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Problem when installing OVS with DPDK

[Ben Pfaff]

On Mon, Aug 21, 2017 at 06:09:57PM +0800, 唐志军 wrote:
> But I also get problems. I am using OVS-2.72 and DPDK-17.05.1

OVS 2.7.2 uses DPDK 16.11.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Regarding CFM/OAM support in Openvswitch

[Ben Pfaff]

[moving ovs-dev to BCC]

I don't know of anyone working on them, but we'd accept implementations
if they were well written.

On Mon, Aug 21, 2017 at 09:43:22AM +, ankaiah.nallamek...@wipro.com wrote:
> Thanks a lot Ben for clarification.
> 
> Any future plans for supporting full pledged 802.1ag support, i.e. all 
> 3(CCM,LBM,LTM) protocols, Please let me know.
> 
> 
> Thanks,
> Anki
> 
> -Original Message-
> From: Ben Pfaff [mailto:b...@ovn.org]
> Sent: Tuesday, August 8, 2017 10:26 PM
> To: Ankaiah Nallamekala (MFG & Tech) <ankaiah.nallamek...@wipro.com>
> Cc: ovs-...@openvswitch.org; ovs-discuss@openvswitch.org
> Subject: Re: [ovs-discuss] Regarding CFM/OAM support in Openvswitch
> 
> ** This mail has been sent from an external source. Treat hyperlinks and 
> attachments in this email with caution**
> 
> On Tue, Aug 08, 2017 at 02:07:25PM +, ankaiah.nallamek...@wipro.com wrote:
> > Ethernet CFM/OAM supports Loopback protocol(LBM) and Link Trace
> > protocol(LTM), Is ovs is supporting these two protocols as well, if so
> > could you please describe how to use these two protocols.
> 
> OVS doesn't support those.
> 
> __
> This email has been scanned by the Symantec Email Security.cloud service.
> For more information please visit http://www.symanteccloud.com 
> __
> The information contained in this electronic message and any attachments to 
> this message are intended for the exclusive use of the addressee(s) and may 
> contain proprietary, confidential or privileged information. If you are not 
> the intended recipient, you should not disseminate, distribute or copy this 
> e-mail. Please notify the sender immediately and destroy all copies of this 
> message and any attachments. WARNING: Computer viruses can be transmitted via 
> email. The recipient should check this email and any attachments for the 
> presence of viruses. The company accepts no liability for any damage caused 
> by any virus transmitted by this email. www.wipro.com
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [ovs-dev] Why 'OVS_VSWITCHD_STOP' do not run?

[Ben Pfaff]

Please don't post to both ovs-dev and ovs-discuss.  I've moved discuss
to BCC.

On Fri, Aug 18, 2017 at 01:07:46PM +0800, Sam wrote:
> I'm running my ovs test case(my-test.at) like this, but when
> 'OVS_VSWITCHD_START'  or other AT_CHECK failed, 'OVS_VSWITCHD_STOP' will
> not run, why and how to fix this to let 'OVS_VSWITCHD_STOP' run anyway.

When AT_CHECK fails, the remainder of the test doesn't run.  If you need
to add cleanup code, you can use the on_exit shell function.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] BFD with 'option : remote_ip = flow'

[Ben Pfaff]

On Fri, Aug 11, 2017 at 08:48:20AM +, Vivek Srivastava V wrote:
> 1.  Is there any way to configure/manage BFD sessions to destinations,
> independent of the tunnel port/interface created?

Not currently.  If you have a good idea for how to extend the OVS BFD
support to be more flexible, we'd accept patches.

> 2.   Does OVS support multi-hop BFD? In roadmap?

No.  I hadn't heard of multi-hop BFD before, so I looked around a bit
and found RFC 5883.  That RFC, though, doesn't really provide a
specification for how to do this.  Is there a detailed specification
somewhere else?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Regarding CFM/OAM support in Openvswitch

[Ben Pfaff]

On Wed, Aug 09, 2017 at 03:12:22AM +, ankaiah.nallamek...@wipro.com wrote:
> Both BFD and CFM (CCM) can be used to monitor connectivity between a
> pair of Ethernet devices., Could you please share the major
> differences between BFD & CFM CCM in openvswitch.

They're both about the same.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [openvswitch 2.8.0 dpdk] testsuite: 1211 1212 1213 1214 1215 1217 1218 1219 1220 1221 1222 1224 1225 1226 2338 failed

[Ben Pfaff]

On Tue, Aug 08, 2017 at 04:26:38PM +, Darrell Ball wrote:
> 
> 
> From:  on behalf of James Page 
> 
> Date: Tuesday, August 8, 2017 at 2:49 AM
> To: "b...@openvswitch.org" 
> Subject: [ovs-discuss] [openvswitch 2.8.0 dpdk] testsuite: 1211 1212 1213 
> 1214 1215 1217 1218 1219 1220 1221 1222 1224 1225 1226 2338 failed
> 
> Hi
> 
> I'm cutting builds from branch-2.8 in preparation for the ovs 2.8.0 release 
> for Ubuntu; we build and test two sets of binaries - a vanilla one and one 
> with dpdk enabled.
> 
> I see test failures on all of the "ofproto-dpif - conntrack" tests with the 
> DPDK build and with the ovn ACL test (see attached logs).  Vanilla build is 
> fine.
> 
> James
> 
> These are generic tests and should not be run with-dpdk set.
> If you run these tests --with-dpdk, some tests will consider the packets 
> coming an actual dpdk interface, which they are not.
> In this case, the packets will be marked with a bad checksum.

All of the tests in the testsuite should always pass, or be skipped,
regardless of configuration, so if some of the tests are inappropriate
for a given configuration then they need AT_SKIP_IF([...]) to ensure
that they get skipped.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [openvswitch 2.8.0 dpdk] testsuite: 1211 1212 1213 1214 1215 1217 1218 1219 1220 1221 1222 1224 1225 1226 2338 failed

[Ben Pfaff]

On Tue, Aug 08, 2017 at 10:52:36PM +, Darrell Ball wrote:
> 
> 
> -Original Message-
> From: Joe Stringer 
> Date: Tuesday, August 8, 2017 at 3:43 PM
> To: Darrell Ball 
> Cc: James Page , "b...@openvswitch.org" 
> 
> Subject: Re: [ovs-discuss] [openvswitch 2.8.0 dpdk] testsuite: 1211 1212 1213 
> 1214 1215 1217 1218 1219 1220 1221 1222 1224 1225 1226 2338 failed
> 
> On 8 August 2017 at 09:26, Darrell Ball  wrote:
> >
> >
> >
> >
> > From:  on behalf of James Page
> > 
> > Date: Tuesday, August 8, 2017 at 2:49 AM
> > To: "b...@openvswitch.org" 
> > Subject: [ovs-discuss] [openvswitch 2.8.0 dpdk] testsuite: 1211 1212 
> 1213
> > 1214 1215 1217 1218 1219 1220 1221 1222 1224 1225 1226 2338 failed
> >
> >
> >
> > Hi
> >
> >
> >
> > I'm cutting builds from branch-2.8 in preparation for the ovs 2.8.0 
> release
> > for Ubuntu; we build and test two sets of binaries - a vanilla one and 
> one
> > with dpdk enabled.
> >
> >
> >
> > I see test failures on all of the "ofproto-dpif - conntrack" tests with 
> the
> > DPDK build and with the ovn ACL test (see attached logs).  Vanilla 
> build is
> > fine.
> >
> >
> >
> > James
> >
> >
> >
> > These are generic tests and should not be run with-dpdk set.
> >
> > If you run these tests --with-dpdk, some tests will consider the packets
> > coming an actual dpdk interface, which they are not.
> >
> > In this case, the packets will be marked with a bad checksum.
> >
> >
> >
> > Are you able to run these tests as we do without “–with-dpdk” ?
> 
> Hmm, this seems surprising to me - I thought that "--with-dpdk" mostly
> just enables another netdevice implementation. Why would this affect
> input/output with netdev-dummy devices?
> 
> For what it's worth, I tried a run of the testsuite with OVS built
> "--with-dpdk" on branch-2.7 and it worked fine:
> 
> https://urldefense.proofpoint.com/v2/url?u=https-3A__travis-2Dci.org_joestringer_openvswitch_jobs_262439494=DwIFaQ=uilaK90D4TOVoH58JNXRgQ=BVhFA09CGX7JQ5Ih-uZnsw=2rYtIAwBngD_iZxhgs9_RxL9aNIlVqYJNRfdSppMEKw=j1JxZ5I8Yj0xapAPLtfpPqwTHiqQEUmUf2ZBdqdJkOo=
>  
> 
> The test failures for the first few are hard-failures (ie ovs uses
> WAIT_UNTIL for something that never succeeds), examples below where
> OVS was waiting to receive packets that never arrive:
> 
> ../../tests/ofproto-dpif.at:9016: ovs-appctl netdev-dummy/receive p2
> 
> 'in_port(2),eth(src=50:54:00:00:00:0a,dst=50:54:00:00:00:09),eth_type(0x0800),ipv4(src=10.1.1.2,dst=10.1.1.1,proto=17,tos=0,ttl=64,frag=no),udp(src=2,dst=1)'
> ../../tests/ofproto-dpif.at:9018: hard failure
> 
> ---
> 
> Some of the later failures are a bit more interesting:
> 
> ../../tests/ofproto-dpif.at:9161: ovs-appctl netdev-dummy/receive p2
> 
> 'in_port(2),eth(src=50:54:00:00:00:0a,dst=50:54:00:00:00:09),eth_type(0x0800),ipv4(src=10.1.1.2,dst=10.1.1.1,proto=17,tos=0,ttl=64,frag=no),udp(src=2,dst=1)'
> ../../tests/ofproto-dpif.at:9164: cat ovs-vswitchd.log | strip_ufid |
> filter_flow_install
> --- - 2017-08-08 09:39:36.051525087 +
> +++ 
> /build/openvswitch-NQWKUM/openvswitch-2.8.0~git20170807.17b6e3ce8/_dpdk/tests/testsuite.dir/at-groups/1214/stdout
> 2017-08-08 09:39:36.046218819 +
> @@ -1,5 +1,4 @@
> 
> -ct_state(+new-est+trk),recirc_id(0x1),in_port(2),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(frag=no),
> actions:drop
> 
> -ct_state(-new+est+trk),recirc_id(0x1),in_port(2),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(proto=17,frag=no),
> actions:1
> 
> +ct_state(-new-est+trk),recirc_id(0x1),in_port(2),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(frag=no),
> actions:drop
>  
> recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(proto=17,frag=no),
> actions:ct(commit),2
>  
> recirc_id(0),in_port(2),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(proto=17,frag=no),
> actions:ct,recirc(0x1)
> 
> ---
> 
> ../../tests/ofproto-dpif.at:9738: ovs-appctl netdev-dummy/receive p1
> 
> '505400095054000a08004528258e40004006ff3d0a0101020a01010100020001396bb55e8cadbf8a501a5ec1'
> ../../tests/ofproto-dpif.at:9740: ovs-appctl revalidator/purge
> ../../tests/ofproto-dpif.at:9744: cat ofctl_monitor.log
> --- /dev/null 2017-04-26 10:10:32.404961898 +
> +++ 
> /build/openvswitch-NQWKUM/openvswitch-2.8.0~git20170807.17b6e3ce8/_dpdk/tests/testsuite.dir/at-groups/1225/stdout
> 2017-08-08 09:40:40.454215126 +
> @@ -0,0 +1,22 @@
> +NXT_PACKET_IN (xid=0x0): table_id=1 cookie=0x1 total_len=54
> 

Re: [ovs-discuss] Lost connectivity when having multiple ports

[Ben Pfaff]

On Mon, Jul 24, 2017 at 01:04:34PM +0200, Pablo Pousada wrote:
> I've been encountering an error on the testbed I'm building, where having
> multiple ports added to a ovs bridge blocks all outwards communication.
> 
> Example: Having the following setup, i have connectivity through the eth0.3
> port:
> 
> root@LEDE:~# ovs-vsctl show
> 41ae4f8f-55db-4cd0-8dd1-3e7b001d8f54
> Bridge "br0"
> Controller "tcp:192.168.1.151"
> Port "br0"
> Interface "br0"
> type: internal
> Port "eth0.3"
> Interface "eth0.3"
> 
> Whenever I add another port to that bridge, inward packages are received,
> but outward packages are never sent. The problem persists with or without
> controller.
> 
> ¿Does anyone have any info on where the problem might be?
> 
> I'm using Open vSwitch version 2.5.0 over LEDE.

This sounds like the second question here:
http://docs.openvswitch.org/en/latest/faq/issues/
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How does OVS ensure only the first packet of a flow is upcalled?

[Ben Pfaff]

On Fri, Jul 14, 2017 at 04:14:29PM +0800, 张克尧 wrote:
> The document of ovs says the datapath is a flow-based software
> switch. A flow consists of many packets. Datapath needs to handle
> every packet. When it does't match any datapath flows, it will do
> upcall. Vswitchd needs to handle upcalls and send netlink messages to
> datapath to install the datapath flows. But before the datapath flows
> are added, another packet of the same flow arrived, will the packet be
> upcalled? That means if I send packets at faster rate, more upcalls
> will be expected?

Yes, more than one packet in a flow can go through an upcall.  It's
unusual, though, because most flows start out with a single packet and
do not continue until they receive a response.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Regarding CFM/OAM support in Openvswitch

[Ben Pfaff]

On Tue, Aug 08, 2017 at 02:07:25PM +, ankaiah.nallamek...@wipro.com wrote:
> Ethernet CFM/OAM supports Loopback protocol(LBM) and Link Trace
> protocol(LTM), Is ovs is supporting these two protocols as well, if so
> could you please describe how to use these two protocols.

OVS doesn't support those.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How OVN works with Openflow fast-failover group table

[Ben Pfaff]

On Tue, Jul 25, 2017 at 11:17:12AM +0800, Harbor Wang wrote:
>I'm learning OVN and wonder how to let OVN works with  Openflow
> fast-failover group table? We have a case that one port done the flow
> should redirect to backup port.

OVN doesn't use fast failover groups, only select groups.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] reply: ovs_assert when do classifier_insert

[Ben Pfaff]

On Thu, Aug 03, 2017 at 03:05:30AM +, Yinpeijun wrote:
> 
> >>On Thu, Aug 03, 2017 at 02:04:34AM +, Yinpeijun wrote:
> >>Recently, there's been a problem when I use ovs, 
> >> the problem is ovs_assert when do classifier_insert. I read the source 
> >> code and annotation, but I don't understand how the problem is been 
> >> triggered.
> 
> >Can you provide a backtrace?  Which assertion is failing?  When did this 
> >start?
> 
>  Thank you for your reply, Ben.
>  
> I use ovs2.0.2(may be too old) and backtrace as follow:

OK.  We'll happily take a bug fix for OVS 2.0, if you come up with one,
but it's too old for me to spend time tracking it down myself.  I hope
that you can figure out how to make it work.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] ovs_assert when do classifier_insert

[Ben Pfaff]

On Thu, Aug 03, 2017 at 02:04:34AM +, Yinpeijun wrote:
>Recently, there's been a problem when I use ovs,
> the problem is ovs_assert when do classifier_insert. I read the source
> code and annotation, but I don't understand how the problem is been
> triggered.

Can you provide a backtrace?  Which assertion is failing?  When did this
start?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] using FQDN for Controller name

[Ben Pfaff]

This is mostly for historical reasons, by the way.  We'd accept a patch
that adds DNS support.  (It's nontrivial though because it would have to
be asynchronous.)

On Mon, Jul 31, 2017 at 01:16:36PM -0700, Shivaram Mysore wrote:
> Thanks Ben for clarification.  Much appreciated.
> 
> On Mon, Jul 31, 2017 at 1:15 PM, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Mon, Jul 31, 2017 at 01:09:07PM -0700, Shivaram Mysore wrote:
> > > Hello,
> > >
> > > I am trying to use FQDN for controller name - example:
> > >
> > > # ovs-vsctl set-controller ovs-br0 ssl:faucet.example.com:6653
> > >
> > > Per OVS logs, I see the below errors:
> > >
> > > Address family not supported by protocol
> > > connection failed (Address family not supported by protocol)
> > >
> > > *Question*: can we not set FQDN as a valid controller name?
> >
> > No.  As documented, OVS requires the controller to be set by IP address.
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] using FQDN for Controller name

[Ben Pfaff]

On Mon, Jul 31, 2017 at 01:09:07PM -0700, Shivaram Mysore wrote:
> Hello,
> 
> I am trying to use FQDN for controller name - example:
> 
> # ovs-vsctl set-controller ovs-br0 ssl:faucet.example.com:6653
> 
> Per OVS logs, I see the below errors:
> 
> Address family not supported by protocol
> connection failed (Address family not supported by protocol)
> 
> *Question*: can we not set FQDN as a valid controller name?

No.  As documented, OVS requires the controller to be set by IP address.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Using libopenvswitch in C++

[Ben Pfaff]

On Sun, Jul 30, 2017 at 08:33:41AM -0700, Ben Warren via discuss wrote:
> Hello Xiao,
> 
> > On Jul 29, 2017, at 11:30 PM, Xiao Liang  wrote:
> > 
> > Hi,
> > 
> > I've encountered some problems building a controller with
> > libopenvswitch in C++. Although they can be solved by some hacks, I
> > want to know if OVS is meant to be used in such case.
> > 
> I don’t think there are very many consumers of libopenvswitch.  It was only 
> added a year and a half ago, and is rarely mentioned in discussions.  The 
> exportability of OVS is definitely geared towards C (and, in my case golang), 
> so it’s more likely the case that people simply weren’t thinking about C++ at 
> the time.
> > 1. In include/openvswitch, some headers are wrapped with 'extern "C"'
> > (e.g. ofpbuf.h), while some are not (e.g. ofp-util.h).
> > 2. The identifier "public" in "struct ofputil_packet_in_private"
> > conflicts with C++ keyword. Not sure if there're others.
> > 3. Private and public declarations (like the
> > ofputil_packet_in_private) could be separated to different files. Also
> > some more APIs (like rconn) could be extracted from lib directory?
> > 
> I expect that the maintainers would accept patches to make the code more C++ 
> friendly, as long as you don’t break the C/golang support.  I can test the 
> latter.

I think that Open vSwitch can do better here.

I sent out a series, would you mind taking a look?  It starts here:
https://mail.openvswitch.org/pipermail/ovs-dev/2017-July/336426.html
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] openflow with openwrt

[Ben Pfaff]

On July 27, 2017 10:36:52 AM PDT, Mohamed Ibrahem 
<mohamed_ibrahem1...@yahoo.com> wrote:
>no, i am using OVS software on my router and give me this error and i
>can not solve it any more
>
>On Thu, 7/27/17, Ben Pfaff <b...@ovn.org> wrote:
>
> Subject: Re: [ovs-discuss] openflow with openwrt
> To: "Mohamed Ibrahem" <mohamed_ibrahem1...@yahoo.com>
> Cc: ovs-discuss@openvswitch.org
> Date: Thursday, July 27, 2017, 6:55 PM
> 
> On Thu, Jul 27, 2017 at 03:32:22PM +,
> Mohamed Ibrahem via discuss wrote:
> > > hello guys,
> > > i have a problem with router tplink
> wr841N , i have installed openwrt to
> >
> > install openflow and get the following message when
> starting the openflow
> > >
> /etc/init.d.openflow/ start    
> > >
> appears, 
> > > / sbin / ofup: line1:
> ofdatapath: not found no need for further
> > > configuration out-of-band control 
> > > / sbin / ofup: line4: ofprotocol: not
> found
> > 
> > can any one
> help me to solve this problem??
> 
> This mailing list is about OVS, but you
> aren't using OVS.
> 
> -Inline Attachment Follows-
> 
> 

The commands and output that you quoted are not part of OVS.___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Multiple Virtual Wireless Ports

[Ben Pfaff]

On Thu, Jul 27, 2017 at 01:33:23PM +, Michael Williams wrote:
> We have OvS running on a wireless router with 4 wired Ethernet
> ports. We can apply rules on the wired ports but when we try to apply
> rules on the wireless port the rules don't work between multiple
> wireless devices. Is there a way within OvS to treat the wireless
> interface like multiple virtual ports so that when a wireless device
> connects we can apply rules to govern behavior between the wireless
> devices like we can with the wired devices?

OVS doesn't distinguish between different kinds of ports, so the
restrictions you're describing don't make sense; OVS doesn't work that
way.  You might be using a vendor's modified version of OVS.  If so,
then you should ask the vendor for assistance.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] openflow with openwrt

[Ben Pfaff]

On Thu, Jul 27, 2017 at 03:32:22PM +, Mohamed Ibrahem via discuss wrote:
> > hello guys,
> > i have a problem with router tplink wr841N , i have installed openwrt to
> > install openflow and get the following message when starting the openflow
> > /etc/init.d.openflow/ start
> > appears, 
> > / sbin / ofup: line1: ofdatapath: not found no need for further
> > configuration out-of-band control 
> > / sbin / ofup: line4: ofprotocol: not found
> 
> can any one help me to solve this problem??

This mailing list is about OVS, but you aren't using OVS.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] openvswitch-2.5.0 ovs-vsctl slow

[Ben Pfaff]

On Tue, Jul 25, 2017 at 07:24:04PM +, MY-OVS DISCUSS via discuss wrote:
> Is there a way to set/remove/clear port's trunks using another command
> other than ovs-vsctl?ovs-vsctl is really slow in executing these
> operations, especially when there are more number of ports.  Currently
> we use ovs-vsctl add port foo1 trunks 123.

You could use a controller.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] A tool to feel-up flow-tables in ofproto

[Ben Pfaff]

On Mon, Jul 17, 2017 at 10:47:12AM +0300, Sara Gittlin wrote:
> On Sun, Jul 16, 2017 at 09:37:05AM +0300, Sara Gittlin wrote:
> > On Wed, Jul 12, 2017 at 11:40:41AM +0300, Sara Gittlin wrote:
> > > Thank you Ben
> > >
> > > >> I need a tool to feel up the flow tables in ofproto.
> > > >> do you know what tool/other should i use ?
> > >
> > > >What does it mean "to feel up" a flow table?
> > >
> > > i want to be able to read ofctl tables and to get a big output ~ 1
> > > - 5 flows.
> > > i need it to order to run some algo. on these flows (e.g. try to
> > > generate megaflows .. )
> > > i know that i can do it with sdn controller -  is there other tools ?
> >
> > >Do you just want to run "ovs-ofctl dump-flows"?
> > >
> >>  Thanks Ben
> >>  this can be an option, unless there is a better way to retrieve all
> > > flows in ofproto
> 
> > So far, you've said that you don't want to use a controller and you want
> > something "better" than ovs-ofctl.  What would be better?  It's
> > difficult to help when you don't give us much information on what you
> > want.
> 
> Ben,  i think i understand now.
> I should use SDN controller , but how to create a big topology so lot
> of flows can be created ?
> mininet ? openstack ? other ?

I'll have to leave that question to others, because it's not my area of
expertise.  I do hear a lot of good things about mininet.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [openvswitch 2.7.90] testsuite: 2341 failed

[Ben Pfaff]

On Mon, Jul 17, 2017 at 02:42:54PM +, Alin Serdean wrote:
> ## -- ##
> ## openvswitch 2.7.90 test suite. ##
> ## -- ##
> 2341. ovn.at:6815: testing ovn -- packet test with HA distributed router 
> gateway port ...

[...]

> ./ovn.at:6975: sort $rcv_text
> --- expout  2017-07-17 17:39:40 +0300
> +++ /c/_2017/july/17/ovs/tests/testsuite.dir/at-groups/2341/stdout  
> 2017-07-17 17:39:40 +0300
> @@ -1 +1,2 @@
>  
> f0010204020102030800451c3f110100c0a80102ac10010300350008
> +020102030806000108000604000102010203ac100101ac100103

This test also hangs, apparently consistently, on my usual test machine.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] A tool to feel-up flow-tables in ofproto

[Ben Pfaff]

On Sun, Jul 16, 2017 at 09:37:05AM +0300, Sara Gittlin wrote:
> On Wed, Jul 12, 2017 at 11:40:41AM +0300, Sara Gittlin wrote:
> > Thank you Ben
> >
> > >> I need a tool to feel up the flow tables in ofproto.
> > >> do you know what tool/other should i use ?
> >
> > >What does it mean "to feel up" a flow table?
> >
> > i want to be able to read ofctl tables and to get a big output ~ 1
> > - 5 flows.
> > i need it to order to run some algo. on these flows (e.g. try to
> > generate megaflows .. )
> > i know that i can do it with sdn controller -  is there other tools ?
> 
> Do you just want to run "ovs-ofctl dump-flows"?
> 
> Thanks Ben
> this can be an option, unless there is a better way to retrieve all
> flows in ofproto

So far, you've said that you don't want to use a controller and you want
something "better" than ovs-ofctl.  What would be better?  It's
difficult to help when you don't give us much information on what you
want.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] building from source / ovs-dpctl-top crash

[Ben Pfaff]

On Mon, Jul 10, 2017 at 04:21:36PM -0400, Blue Lang wrote:
> New to the list. My company is building an (awesome) industrial security
> networking platform and OVS is a big part of the solution, so thank you all
> for your work.
> 
> I'd like to use ovs-dpctl-top as part of our testing suite, but it is
> crashing on launch. I built it and mininet from scratch/TOT to be sure it
> wasn't an old bug. I have two suggestions for the build scripts, and the
> error I'm seeing with -top.
> 
> 
>- Should make install create /usr/local/var/run/openvswitch by default?
>Right now this is a needless manual step.

ovs-ctl will create this before it starts any daemon, and the distro
packages also create it, so it's usually not necessary.

>- Can ovsdb-server create a symlink to db.sock to avoid requiring
> "--remote=punix:/usr/local/var/run/openvswitch/db.sock" in what seems to
>be the default invocation? All of the related tools seem to want to use
>db.sock rather than the instantiation-specific socket.

What symlink do you mean?  I don't know of installations that use a
symlink here.

> Sorry if I misunderstood the operation of either of the above - this is
> just based on my experience today building it for the first time.
> 
> -top dies with the following stack trace the moment mininet instantiates:
> 
> Traceback (most recent call last):
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 1290, in 
> 
> sys.exit(main())
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 1282, in main
> 
> flows_top(args)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 1191, in flows_top
> 
> flows_read(ihdl, flow_db)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 593, in flows_read
> 
> flow_db.flow_line_add(line)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 988, in flow_line_add
> 
> self.flow_event(fields_dict, stats_old_dict, stats_dict)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 1080, in flow_event
> 
> matches = flow_aggregate(fields_dict, stats_new_dict)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 576, in flow_aggregate
> 
> field, stats_dict)
> 
>   File "/usr/local/bin/ovs-dpctl-top", line 268, in element_ipv6_get
> 
> element_show = fmt % (field_type, element["src"], element["dst"])
> 
> KeyError: 'src'
> 
> If this is not a known error I can spend some time debugging it.

I think that ovs-dpctl-top is probably not used very much.  Maybe bugs
have crept in because no one is using it.  If you find the problem and
fix it, please pass it along.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Subject: [openvswitch 2.7.1] testsuite: 7 8 767 1891 1892 1893 1894 1895 1896 1900 1902 failed

[Ben Pfaff]

On Mon, Jul 10, 2017 at 01:14:38PM -0400, Lance Richardson wrote:
> > From: "Stuart Cardall" 
> > To: b...@openvswitch.org
> > Sent: Saturday, 8 July, 2017 5:22:17 PM
> > Subject: [ovs-discuss] Subject: [openvswitch 2.7.1] testsuite: 7 8 767 1891 
> > 1892 1893 1894 1895 1896 1900 1902 failed
> > 
> > 
> > 
> > Hello,
> > 
> > Attached is the test suite log for ovs 2.7.1 in Alpine Linux / musl c:
> > 
> > testsuite: 7 8 767 1891 1892 1893 1894 1895 1896 1900 1902 failed
> > 
> > 
> > Kind Regards,
> > 
> > Stuart Cardall.
> > 
> 
> These will pass if GNU awk is used instead of busybox awk (apk add gawk).
> 
>7: completion.at:343  vsctl-bashcomp - basic verification
>8: completion.at:425  vsctl-bashcomp - argument completion

That's interesting.  I thought I'd try to figure out the root of the
problem, but I don't get failures in my usual environment if I replace
"awk" by "busybox awk" and run these tests, so I wonder whether there's
something else at play.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Could anyone tell me whether there is a buffer between M2 and vswitchd?

[Ben Pfaff]

On Fri, Jul 14, 2017 at 12:56:28AM +0530, Numan Siddique wrote:
> On Fri, Jul 14, 2017 at 12:21 AM, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Thu, Jul 13, 2017 at 04:13:33PM +0800, 杨润垲 wrote:
> > > hi all,
> > > Recently, I analyze the performance of Open vSwitch. By reading the
> > source code, I abstract the packet-receiving process as the model of Fig.1.
> > Queue1 (denoting NIC DMA) is a buffer that stores the incoming packets. M1
> > processes the packets, and looks up the flow table to match an entry. If
> > the match fails, the packet is sent to Queue2 in order to be handled by
> > ovs-vswitchd. If successful, the packet will be sent out by M2, which
> > finishes the left processing.
> > > Please help me regarding this.
> >
> > What's your question?
> >
> 
> Looks like the question is in the attached pdf file - question.pdf
> 
> 
> Could anyone tell me whether there is a buffer between M2 and vswitchd?
> Recently, I analyze the performance of Open vSwitch. By reading the source
> code, I
> abstract the packet-receiving process as the model of Fig.1. Queue1
> (denoting NIC DMA) is
> a buffer that stores the incoming packets. M1 processes the packets, and
> looks up the flow
> table to match an entry. If the match fails, the packet is sent to Queue2
> in order to be handled
> by ovs-vswitchd. If successful, the packet will be sent out by M2, which
> finishes the left
> processing.

There's no buffer.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [ovs-dev] how to get packet information in ovs-dpdk?

[Ben Pfaff]

Now I see that you're reporting a bug in "ovs-dpdk", not in Open
vSwitch.  You should report that bug to an ovs-dpdk mailing list.

On Thu, Jul 13, 2017 at 03:24:40PM +0800, Sam wrote:
> Yes it is.
> 
> I'm also confused, as when I use normal kernel based ovs-2.3.0, `ovs-ofctl
> snoop br0` works good and will not quit.
> 
> 2017-07-12 23:24 GMT+08:00 Ben Pfaff <b...@ovn.org>:
> 
> > On Wed, Jul 12, 2017 at 04:21:04PM +0800, Sam wrote:
> > > hi all,
> > >
> > > I'm running ovs-dpdk(ovs-2.4.9), I found the counter of bond port of br_t
> > > is increasing, but I want to know what's these packets, how could I do
> > this?
> > >
> > > first, I use `ovs-ofctl snoop br_t`, but it quit quickly as below:
> > >
> > > > [root@yf-mos-test-net07 ~]# /usr/local/bin/ovs-ofctl snoop
> > > > /usr/local/var/run/openvswitch/br_t.mgmt
> > > > [root@yf-mos-test-net07 ~]# /usr/local/bin/ovs-ofctl snoop
> > > > /usr/local/var/run/openvswitch/br_t.mgmt
> > > > ovs-ofctl: /usr/local/var/run/openvswitch/ovs-ofctl.pid: already
> > running
> > > > as pid 22082, aborting
> >
> > That is a strange message.  It should only appear if you use --pidfile
> > and --daemon and there's already an ovs-ofctl running that way.  Is the
> > above the complete context from your session?
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Could anyone tell me whether there is a buffer between M2 and vswitchd?

[Ben Pfaff]

On Thu, Jul 13, 2017 at 04:13:33PM +0800, 杨润垲 wrote:
> hi all,
> Recently, I analyze the performance of Open vSwitch. By reading the source 
> code, I abstract the packet-receiving process as the model of Fig.1. Queue1 
> (denoting NIC DMA) is a buffer that stores the incoming packets. M1 processes 
> the packets, and looks up the flow table to match an entry. If the match 
> fails, the packet is sent to Queue2 in order to be handled by ovs-vswitchd. 
> If successful, the packet will be sent out by M2, which finishes the left 
> processing.
> Please help me regarding this.

What's your question?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Cannot match correct ethertype after POP vlan and GOTO table

[Ben Pfaff]

On Tue, Jul 04, 2017 at 03:27:50PM +0800, Dickens Yeh wrote:
> Hi,
> I have an question, and I haven't found any OpenFlow Spec to defined it ,
> please give me a help.
> 
> I am trying to work with 3 vlan tags, but it cannot match correct ethertype
> after pop 1 vlan tag.
> OVS Bridge already set with vlan-limit = 0, and it can watch over 2 vlan
> and match the correct ethertype like ARP.
> 
> My question is:  the packet have 3 vlan tags, pop 1 vlan tag and goto-table
> 1. In table 1, will it match with the new packet (modified) or the original
> packet like in table 0?

The OVS documentation says:

   other_config  :  vlan-limit: optional string, containing an integer, at
   least 0
  Limits the number of VLAN headers that can  be  matched  to  the
  specified  number.  Further VLAN headers will be treated as pay‐
  load, e.g. a packet with more 802.1q headers will match Ethernet
  type 0x8100.

  Value  0  means  unlimited.  The actual number of supported VLAN
  headers is the smallest of vlan-limit, the number of VLANs  sup‐
  ported  by  Open vSwitch userspace (currently 2), and the number
  supported by the datapath.

  If this value is absent, the default is currently 1. This  main‐
  tains backward compatibility with controllers that were designed
  for use with Open vSwitch versions earlier than 2.8, which  only
  supported one VLAN.

It's not clear to me whether you're describing a bug.  Does OVS behavior
differ from what is documented abovve?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Remote connection method for ovn-controller ovs-ofctl invocation.

[Ben Pfaff]

Sure, send a patch to add a configuration option.

On Thu, Jul 13, 2017 at 05:41:11PM +, Rana, JaiSingh wrote:
> Hi Ben, 
> Sorry for the confusion. I meant ovn-controller invocation of function for 
> pushing open-flows   from southdb to vswitchd which is running in nic using 
> tcp connection and I agree with Lance that using external-ids for this 
> purpose might be better idea.
> Also for now, only this assumption i.e. ovn-controller and vswitchd are 
> running on same host is blocking me. As as a proof of concept, i hardcoded 
> tcp connection instead of unix socket file in ovn and now flows are getting 
> updated in vswitchd. 
> Other configuration change that is done is to setup OF controller on 
> integration bridge with ptcp instead of tcp method. 
> 
> Thanks,
> Jai
> 
> 
> On July 13, 2017 10:35:07 PM Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Thu, Jul 13, 2017 at 12:53:11PM -0400, Lance Richardson wrote:
> >> > From: "Ben Pfaff" <b...@ovn.org>
> >> > To: "Lance Richardson" <lrich...@redhat.com>
> >> > Cc: "JaiSingh Rana" <jaisingh.r...@cavium.com>, 
> >> > ovs-discuss@openvswitch.org
> >> > Sent: Thursday, 13 July, 2017 12:47:47 PM
> >> > Subject: Re: [ovs-discuss] Remote connection method for ovn-controller 
> >> > ovs-ofctl invocation.
> >> > 
> >> > On Thu, Jul 13, 2017 at 09:35:01AM -0400, Lance Richardson wrote:
> >> > > > From: "JaiSingh Rana" <jaisingh.r...@cavium.com>
> >> > > > To: ovs-discuss@openvswitch.org
> >> > > > Sent: Thursday, 13 July, 2017 3:49:15 AM
> >> > > > Subject: [ovs-discuss] Remote connection method for ovn-controller
> >> > > > ovs-ofctlinvocation.
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > Hi,
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > Currently in Openvswitch-2.7.1, ovn-controller hard codes connection
> >> > > > method
> >> > > > for ovs-ofctl invokation as unix:file as assumption is there will be 
> >> > > > unix
> >> > > > file created by vswitchd in OVS_RUNDIR for managing OF controller on
> >> > > > bridge
> >> > > > e.g. br-int.mgmt
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > There is an issue in our Openvswitch offload model where vswitchd is
> >> > > > running
> >> > > > on nic and ovn-controller on host. As there is no option for telling
> >> > > > ovn-controller to use tcp:port connection method for ovs-ofctl , 
> >> > > > flows
> >> > > > are
> >> > > > not being pushed to vswitchd.
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > 
> >> > > > For providing the patch, need some input. Should connection method be
> >> > > > provided as an argument to ovn-controller or it can be written to
> >> > > > /etc/sysconfig/openvswitch from where ovn-controller can read and 
> >> > > > use if
> >> > > > option is present otherwise it defaults to unix file method.
> >> > > > 
> >> > > > 
> >> > > 
> >> > > My suggestion would be to use external-ids in the local ovsdb, as is
> >> > > currently done for ovn-remote and ovn-encap. Maybe something like:
> >> > > 
> >> > > ovs-vsctl set open . external-ids:ovn-ofctl=tcp:w.x.y.z:abcd
> >> > 
> >> > Using the name "ovs-ofctl" here is weird.  ovn-controller doesn't use
> >> > ovs-ofctl.
> >> > 
> >> 
> >> Well, I'm terrible with naming things, and should have added a comment to 
> >> that
> >> effect :-) 
> >> 
> >> My rationale for "ovn-ofctl" was shallowly based on the the related file,
> >> ovn/controller/ofctrl.c (well, I omitted the 'r' for some reason,.)
> >> 
> >> What would be a more appropriate name?
> >
> > I really should have directed that to the original poster, since he
> > suggested that ovs-ofctl was involved.
> >
> > Probably, something that used words like "switch", "openflow", "remote",
> > "connection", "ovs", etc., like "ovn-switch-connection".
> >
> > However, this is not going to be the only problem that the OP will
> > encounter, since ovn-controller makes other assumptions that it is
> > running on the ovs-switchd host.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Remote connection method for ovn-controller ovs-ofctl invocation.

[Ben Pfaff]

On Thu, Jul 13, 2017 at 12:53:11PM -0400, Lance Richardson wrote:
> > From: "Ben Pfaff" <b...@ovn.org>
> > To: "Lance Richardson" <lrich...@redhat.com>
> > Cc: "JaiSingh Rana" <jaisingh.r...@cavium.com>, ovs-discuss@openvswitch.org
> > Sent: Thursday, 13 July, 2017 12:47:47 PM
> > Subject: Re: [ovs-discuss] Remote connection method for ovn-controller 
> > ovs-ofctlinvocation.
> > 
> > On Thu, Jul 13, 2017 at 09:35:01AM -0400, Lance Richardson wrote:
> > > > From: "JaiSingh Rana" <jaisingh.r...@cavium.com>
> > > > To: ovs-discuss@openvswitch.org
> > > > Sent: Thursday, 13 July, 2017 3:49:15 AM
> > > > Subject: [ovs-discuss] Remote connection method for ovn-controller
> > > > ovs-ofctl   invocation.
> > > > 
> > > > 
> > > > 
> > > > Hi,
> > > > 
> > > > 
> > > > 
> > > > 
> > > > Currently in Openvswitch-2.7.1, ovn-controller hard codes connection
> > > > method
> > > > for ovs-ofctl invokation as unix:file as assumption is there will be 
> > > > unix
> > > > file created by vswitchd in OVS_RUNDIR for managing OF controller on
> > > > bridge
> > > > e.g. br-int.mgmt
> > > > 
> > > > 
> > > > 
> > > > 
> > > > There is an issue in our Openvswitch offload model where vswitchd is
> > > > running
> > > > on nic and ovn-controller on host. As there is no option for telling
> > > > ovn-controller to use tcp:port connection method for ovs-ofctl , flows
> > > > are
> > > > not being pushed to vswitchd.
> > > > 
> > > > 
> > > > 
> > > > 
> > > > For providing the patch, need some input. Should connection method be
> > > > provided as an argument to ovn-controller or it can be written to
> > > > /etc/sysconfig/openvswitch from where ovn-controller can read and use if
> > > > option is present otherwise it defaults to unix file method.
> > > > 
> > > > 
> > > 
> > > My suggestion would be to use external-ids in the local ovsdb, as is
> > > currently done for ovn-remote and ovn-encap. Maybe something like:
> > > 
> > > ovs-vsctl set open . external-ids:ovn-ofctl=tcp:w.x.y.z:abcd
> > 
> > Using the name "ovs-ofctl" here is weird.  ovn-controller doesn't use
> > ovs-ofctl.
> > 
> 
> Well, I'm terrible with naming things, and should have added a comment to that
> effect :-) 
> 
> My rationale for "ovn-ofctl" was shallowly based on the the related file,
> ovn/controller/ofctrl.c (well, I omitted the 'r' for some reason,.)
> 
> What would be a more appropriate name?

I really should have directed that to the original poster, since he
suggested that ovs-ofctl was involved.

Probably, something that used words like "switch", "openflow", "remote",
"connection", "ovs", etc., like "ovn-switch-connection".

However, this is not going to be the only problem that the OP will
encounter, since ovn-controller makes other assumptions that it is
running on the ovs-switchd host.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Remote connection method for ovn-controller ovs-ofctl invocation.

[Ben Pfaff]

On Thu, Jul 13, 2017 at 09:35:01AM -0400, Lance Richardson wrote:
> > From: "JaiSingh Rana" 
> > To: ovs-discuss@openvswitch.org
> > Sent: Thursday, 13 July, 2017 3:49:15 AM
> > Subject: [ovs-discuss] Remote connection method for ovn-controller 
> > ovs-ofctlinvocation.
> > 
> > 
> > 
> > Hi,
> > 
> > 
> > 
> > 
> > Currently in Openvswitch-2.7.1, ovn-controller hard codes connection method
> > for ovs-ofctl invokation as unix:file as assumption is there will be unix
> > file created by vswitchd in OVS_RUNDIR for managing OF controller on bridge
> > e.g. br-int.mgmt
> > 
> > 
> > 
> > 
> > There is an issue in our Openvswitch offload model where vswitchd is running
> > on nic and ovn-controller on host. As there is no option for telling
> > ovn-controller to use tcp:port connection method for ovs-ofctl , flows are
> > not being pushed to vswitchd.
> > 
> > 
> > 
> > 
> > For providing the patch, need some input. Should connection method be
> > provided as an argument to ovn-controller or it can be written to
> > /etc/sysconfig/openvswitch from where ovn-controller can read and use if
> > option is present otherwise it defaults to unix file method.
> > 
> > 
> 
> My suggestion would be to use external-ids in the local ovsdb, as is
> currently done for ovn-remote and ovn-encap. Maybe something like:
> 
> ovs-vsctl set open . external-ids:ovn-ofctl=tcp:w.x.y.z:abcd

Using the name "ovs-ofctl" here is weird.  ovn-controller doesn't use
ovs-ofctl.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] max_speed and curr_speed isnt sent correctly

[Ben Pfaff]

On Wed, Jun 28, 2017 at 08:52:56AM +0430, Tahereh Yaghoubi wrote:
> I want to collect of_port information from ovs switches. But in its reply,
> value of max_speed is set to 0 and value of curr_speed is set to 10M while
> I can receive ofp_port_stats correctly. Also I can ping all hosts. In my
> mininet topology I determined bandwidth of links and I expect max_speed set
> to bw.
> when packets of of_port is inspected from ovs by using wireshark, ovs sent
> these values.

OVS just reports what the kernel tells it.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [ovs-dev] how to get packet information in ovs-dpdk?

[Ben Pfaff]

On Wed, Jul 12, 2017 at 04:21:04PM +0800, Sam wrote:
> hi all,
> 
> I'm running ovs-dpdk(ovs-2.4.9), I found the counter of bond port of br_t
> is increasing, but I want to know what's these packets, how could I do this?
> 
> first, I use `ovs-ofctl snoop br_t`, but it quit quickly as below:
> 
> > [root@yf-mos-test-net07 ~]# /usr/local/bin/ovs-ofctl snoop
> > /usr/local/var/run/openvswitch/br_t.mgmt
> > [root@yf-mos-test-net07 ~]# /usr/local/bin/ovs-ofctl snoop
> > /usr/local/var/run/openvswitch/br_t.mgmt
> > ovs-ofctl: /usr/local/var/run/openvswitch/ovs-ofctl.pid: already running
> > as pid 22082, aborting

That is a strange message.  It should only appear if you use --pidfile
and --daemon and there's already an ovs-ofctl running that way.  Is the
above the complete context from your session?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] A tool to feel-up flow-tables in ofproto

[Ben Pfaff]

On Wed, Jul 12, 2017 at 11:40:41AM +0300, Sara Gittlin wrote:
> Thank you Ben
> 
> >> I need a tool to feel up the flow tables in ofproto.
> >> do you know what tool/other should i use ?
> 
> >What does it mean "to feel up" a flow table?
> 
> i want to be able to read ofctl tables and to get a big output ~ 1
> - 5 flows.
> i need it to order to run some algo. on these flows (e.g. try to
> generate megaflows .. )
> i know that i can do it with sdn controller -  is there other tools ?

Do you just want to run "ovs-ofctl dump-flows"?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] A tool to feel-up flow-tables in ofproto

[Ben Pfaff]

On Tue, Jul 11, 2017 at 10:58:39AM +0300, Sara Gittlin wrote:
> I need a tool to feel up the flow tables in ofproto.
> do you know what tool/other should i use ?

What does it mean "to feel up" a flow table?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Question regarding group reference count

[Ben Pfaff]

It's probably a bug in OVS.  The group reference count is, from an OVS
perspective, a bookkeeping exercise without real value, so it can easily
get out of sync.

On Thu, Jul 06, 2017 at 09:37:49PM +, Charles Chan wrote:
> Hi everyone,
> 
> I just downloaded the latest OVS release (2.7.1) but unfortunately I am
> still able to reproduce the issue.
> It seems to be an OVS-specific issue since the reference count is reported
> correctly on other OF-DPA based hardware switches.
> Any thought?
> 
> Thanks,
> Charles
> 
> On Mon, Jun 19, 2017 at 4:46 PM Charles Chan  wrote:
> 
> > Hi everyone,
> >
> > I noticed a strange behavior of group reference count in ovs 2.7.0 and
> > therefore I am writing to confirm if this is expected.
> >
> > I created 2 indirect groups and 1 select group. There are 2 buckets in the
> > select group, each of them points to one of the indirect group.
> > sudo ovs-ofctl -O OpenFlow13 dump-groups ovs-br
> > OFPST_GROUP_DESC reply (OF1.3) (xid=0x2):
> > group_id=201,type=indirect,bucket=actions=output:1
> > group_id=202,type=indirect,bucket=actions=output:2
> > group_id=101,type=select,
> > *bucket=actions=group:201,bucket=actions=group:202*
> >
> > I also created a flow that points to the select group.
> > sudo ovs-ofctl -O OpenFlow13 dump-flows ovs-br
> > OFPST_FLOW reply (OF1.3) (xid=0x2):
> >  cookie=0x0, duration=31.409s, table=0, n_packets=0, n_bytes=0,
> > vlan_tci=0x000a/0x1fff actions=group:101
> >  cookie=0x0, duration=275.500s, table=0, n_packets=0, n_bytes=0,
> > priority=0 actions=NORMAL
> >
> > However, the reference count of these 2 indirect groups are both zero.
> > sudo ovs-ofctl -O OpenFlow13 dump-group-stats ovs-br
> > OFPST_GROUP reply (OF1.3) (xid=0x2):
> >  group_id=201,duration=185.249s,*ref_count=0*
> > ,packet_count=0,byte_count=0,bucket0:packet_count=0,byte_count=0
> >  group_id=202,duration=159.981s,*ref_count=0*
> > ,packet_count=0,byte_count=0,bucket0:packet_count=0,byte_count=0
> >
> >  
> > group_id=101,duration=111.349s,ref_count=1,packet_count=0,byte_count=0,bucket0:packet_count=0,byte_count=0,bucket1:packet_count=0,byte_count=0
> >
> > I further looked up OF 1.3 spec and discovered this description in the
> > section of OFPMP_GROUP.
> > uint32_t ref_count; /* Number of *flows or groups that directly forward
> > to this group*. */
> > According to my interpretation of the spec, the reference count of these
> > indirect groups should both be 1 instead of 0 since they are referenced by
> > the select group.
> >
> > This can be easily reproduced by following steps.
> > sudo ovs-vsctl add-br ovs-br
> > sudo ovs-ofctl -O OpenFlow13 add-group ovs-br
> > group_id=201,type=indirect,bucket=actions=output:1
> > sudo ovs-ofctl -O OpenFlow13 add-group ovs-br
> > group_id=202,type=indirect,bucket=actions=output:2
> > sudo ovs-ofctl -O OpenFlow13 add-group ovs-br
> > group_id=101,type=select,bucket=actions=group:201,bucket=actions=group:202
> > sudo ovs-ofctl -O OpenFlow13 add-flow ovs-br vlan_vid=10,actions=group:101
> >
> > I installed OVS from the 2.7.0 tarball. This is the version I am using.
> > sudo ovs-vsctl --version
> > ovs-vsctl (Open vSwitch) 2.7.0
> > DB Schema 7.14.0
> >
> > Thanks,
> > Charles
> > --
> >
> > Charles Chan
> > Member of Technical Staff, ON.Lab
> > Member of Ambassador Steering Team, ONOS/CORD Community
> >
> -- 
> 
> Charles Chan
> Member of Technical Staff, ONF
> Member of Ambassador Steering Team, ONOS/CORD Community

> ___
> discuss mailing list
> disc...@openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] 答复: 答复: about

[Ben Pfaff]

usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==22676==by 0x41BC02: ofproto_destroy__ (ofproto.c:1569)
> ==22676==by 0x4B8345: ovsrcu_call_postponed (ovs-rcu.c:293)
> ==22676==by 0x4B8513: ovsrcu_postpone_thread (ovs-rcu.c:308)
> ==22676==by 0x4B9263: ovsthread_wrapper (ovs-thread.c:340)
> ==22676==by 0x5491DF4: start_thread (in /usr/lib64/libpthread-2.17.so)
> ==22676==by 0x5CA61AC: clone (in /usr/lib64/libc-2.17.so)
> ==22676== 
> ovs-vswitchd(monitor6): ofproto/ofproto-dpif.c:4395: ovs_mutex_lock_at() 
> passed uninitialized ovs_mutex
> ,==22676== 
> ==22676== HEAP SUMMARY:
> ==22676== in use at exit: 4,277,720 bytes in 1,039 blocks
> ==22676==   total heap usage: 34,041 allocs, 33,002 frees, 25,258,154 bytes 
> allocated
> ==22676== 
> ==22676== LEAK SUMMARY:
> ==22676==definitely lost: 89 bytes in 4 blocks
> ==22676==indirectly lost: 0 bytes in 0 blocks
> ==22676==  possibly lost: 4,196,601 bytes in 21 blocks
> ==22676==still reachable: 81,030 bytes in 1,014 blocks
> ==22676== suppressed: 0 bytes in 0 blocks
> ==22676== Rerun with --leak-check=full to see details of leaked memory
> ==22676== 
> ==22676== For counts of detected and suppressed errors, rerun with: -v
> ==22676== Use --track-origins=yes to see where uninitialised values come from
> ==22676== ERROR SUMMARY: 7 errors from 7 contexts (suppressed: 2 from 2)
> Killed
> [root@localhost ~]#
> 
> 
> 
> the reproduction information as follows:
> 
> 1、create a bridge "br1" with the netdev type
> 2、run the command “ovs-vsctl set int br1 lldp:enable=true “ to make the 
> interface br1 enable the function lldp
> 3、run the command ''ovs-vsctl del-br br1" to delete the br1
> =
> -邮件原件-
> 发件人: Ben Pfaff [mailto:b...@ovn.org] 
> 发送时间: 2017年4月25日 1:15
> 收件人: qintao (F)
> 抄送: Justin Pettit; Liuguifeng; ovs-discuss@openvswitch.org; wuhao (S); 
> zhouyong (R); Lukai (Look); Guoyilong
> 主题: Re: [ovs-discuss] 答复: about  "ovs-vswitchd"with the userspace ovs 2.5.0>
> 
> OK, let's figure out the problem.  Can you provide a backtrace?  Or run OVS 
> until valgrind and provide valgrind's report?  Or can you provide 
> reproduction information for us?
> 
> Thanks,
> 
> Ben.
> 
> On Thu, Apr 20, 2017 at 08:21:54AM +, qintao (F) wrote:
> > hi,Pettit
> > I have reprodeuced the same issue with 2.5.2, but the result is still 
> > disappointing ,which is the crash of the process "ovs-vswitchd".
> > 
> > best regards,
> >  Tony tao
> > -邮件原件-
> > 发件人: Justin Pettit [mailto:jpet...@ovn.org]
> > 发送时间: 2017年4月20日 5:05
> > 收件人: qintao (F)
> > 抄送: ovs-discuss@openvswitch.org; wuhao (S); Liuguifeng; Lukai (Look); 
> > Guoyilong
> > 主题: Re: [ovs-discuss] about  > process "ovs-vswitchd"with the userspace ovs 2.5.0>
> > 
> > 
> > > On Apr 18, 2017, at 6:59 PM, qintao (F) <qint...@huawei.com> wrote:
> > > 
> > >  
> > >  
> > > Dear all ,
> > >  we create a bridge “br1” with the type of netdev .And the version of 
> > > the ovs is 2.5.0. Then we run the command “ovs-vsctl set int br1 
> > > lldp:enable=true “ to make the interface br1 enable the function 
> > > lldp.After that ,we delete the bridge br1 ,we found the the process 
> > > “ovs-vswitchd” has been lost.
> > > ”
> > 
> > Thanks for the report.  There are a couple of releases in the 2.5.x series 
> > since 2.5.0.  Are you able to reproduce the same issue with 2.5.2?
> > 
> > --Justin
> > 
> > 
> > ___
> > discuss mailing list
> > disc...@openvswitch.org
> > https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Open vSwitch 2.7.1 Available

[Ben Pfaff]

On Thu, Jul 06, 2017 at 10:57:06AM +0300, Markos Chandras wrote:
> On 07/06/2017 10:30 AM, Justin Pettit wrote:
> > The Open vSwitch team is pleased to announce the release of Open vSwitch 
> > 2.7.1:
> > 
> >http://openvswitch.org/releases/openvswitch-2.7.1.tar.gz
> > 
> > This release contains bug fixes and minor improvements for 2.7.0.
> > 
> > --The Open vSwitch Team 
> > 
> 
> Hello Justin,
> 
> Could you also push a v2.7.1 tag for this release? I can't see one in
> https://github.com/openvswitch/ovs/releases

Pushed.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] set-controller for IPv6 gives invalid argument error

[Ben Pfaff]

On Tue, Jul 04, 2017 at 05:52:06PM +, Ali Volkan Atli wrote:
> Hi 
> 
> I'm trying to connect OVS to a controller using IPv6 as below, but it is not 
> working.
> 
> # sudo ovs-vsctl set-controller s1 tcp:[fe80::41f3:ab56:bbab:a528]
> 
> However, it gives the following error: 
> 
> 2017-07-04T17:47:00Z|00730|rconn|WARN|s1<->tcp:[fe80::41f3:ab56:bbab:a528]: 
> connection failed (Invalid argument)
> 2017-07-04T17:47:08Z|00731|stream_tcp|ERR|tcp:[fe80::41f3:ab56:bbab:a528]: 
> connect: Invalid argument
> 
> When I check ping6, I can ping
> 
> # ping6 -I eth0 fe80::41f3:ab56:bbab:a528
> PING fe80::41f3:ab56:bbab:a528(fe80::41f3:ab56:bbab:a528) from 
> fe80::f22:a6c2:6603:13b eth0: 56 data bytes
> 64 bytes from fe80::41f3:ab56:bbab:a528: icmp_seq=1 ttl=128 time=0.429 ms
> 64 bytes from fe80::41f3:ab56:bbab:a528: icmp_seq=2 ttl=128 time=0.740 ms
> 
> What should I do? Thanks in advance..

Interesting, I'd never even heard of such issues.

I sent a patch that should allow you to use this via:
sudo ovs-vsctl set-controller s1 'tcp:[fe80::41f3:ab56:bbab:a528%eth0]'

(I recommend quoting anything with shell metacharacters like [], by the
way.)
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] how to understand ofctl request msg ?

[Ben Pfaff]

On Fri, Jun 23, 2017 at 01:37:30PM +0800, Margin Hu wrote:
> Hi Guys,
> 
> Now I met an issue about neutron and openvswitch,  how to understand ofctl
> request msg like as following ?　
> 
> 2017-06-23 06:21:12.783 7 DEBUG
> neutron.plugins.ml2.drivers.openvswitch.agent.openflow.native.ofswitch
> [req-381c7dde-6c3d-4e3b-af32-ae20c75f59d9 - - - - -] ofctl request 
> version=0x4,msg_type=0xe,msg_len=0x68,xid=0xd791e92,OFPFlowMod(buffer_id=4294967295,command=0,cookie=9477051674213136541L,cookie_mask=0,flags=0,hard_timeout=0,idle_timeout=0,instructions=[OFPInstructionActions(actions=[OFPActionPopVlan(len=8,type=18),
>  
> OFPActionOutput(len=16,max_len=0,port=4294967290,type=0)],len=32,type=4)],match=OFPMatch(oxm_fields={'in_port':
> 2, 'vlan_vid': 4098}),out_group=0,out_port=0,priority=4,table_id=0) result
> None _send_msg 
> /usr/lib/python2.7/site-packages/neutron/plugins/ml2/drivers/openvswitch/agent/openflow/native/ofswitch.py:93
> 
>  command=0  mean "add flow rule" , I want to know which bridge this request
> is sent to.

The log message doesn't say.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] failed to "ovs-vsctl add-port"

[Ben Pfaff]

On Thu, Jun 22, 2017 at 01:48:23PM +0800, Margin Hu wrote:
> [root@cloud ]# ovs-vsctl del-port provision bond0
> [root@cloud]# ovs-vsctl add-port provision bond0
> ovs-vsctl: Error detected while setting up 'bond0'.  See ovs-vswitchd log
> for details.
> 
> ovs-vswitchd.log
> 
> 1581163:2017-06-22T05:42:07.644Z|283577|dpif|WARN|system@ovs-system: failed
> to add bond0 as port: File exists
> 
> how to solve the issue?   which file  did "File exists" mean?

This may mean that bond0 is already in use in some incompatible way; for
example, that it is part of a Linux bridge created with "brctl addbr".
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Receiving a bridge error from devstack

[Ben Pfaff]

Usually this means that some error prevented ovs-vswitchd from actually
creating what the database configured.  It can also mean that
ovs-vswitchd is not running.  Check the ovs-vswitchd log for more
information.

On Thu, Jun 22, 2017 at 09:58:32AM +, Ali Volkan Atli wrote:
> 
> Hi all
> 
> I met an issue about devstack with openvswitch using ./stack.sh script. In 
> the log below, it says br-ex and br-int is not a bridge but when I check 
> ovs-vsctl, the bridges are exist. How can this happen?
> 
> Thanks in advance
> 
> - Volkan
> 
> ./stack.sh log
> 
> 2017-06-22 09:20:26.795 | + ./stack.sh:main:1233 :   
> /usr/local/bin/nova-manage --config-file /etc/nova/nova.conf network create 
> private 10.7.7.0/24 1 256
> 2017-06-22 09:20:29.913 | ERROR: Network commands are not supported when 
> using the Neutron API.  Use python-neutronclient instead.
> 2017-06-22 09:20:30.013 | ++ ./stack.sh:main:1233 :   
> err_trap
> 2017-06-22 09:20:30.017 | ++ ./stack.sh:err_trap:504  :   
> local r=2
> 2017-06-22 09:20:30.021 | stack.sh failed: full log in 
> /opt/stack/logs/stack.sh.log.2017-06-22-120812
> 2017-06-22 09:20:30.024 | Error on exit
> 2017-06-22 09:20:30.685 | ovs-ofctl: br-ex is not a bridge or a socket
> 2017-06-22 09:20:30.722 | ovs-ofctl: br-int is not a bridge or a socket
> 2017-06-22 09:20:30.735 | ovs-ofctl: br-ex is not a bridge or a socket
> 2017-06-22 09:20:30.743 | ovs-ofctl: br-int is not a bridge or a socket
> 2017-06-22 09:20:30.762 | ovs-ofctl: br-ex is not a bridge or a socket
> 2017-06-22 09:20:30.778 | ovs-ofctl: br-int is not a bridge or a socket
> 2017-06-22 09:20:30.804 | ovs-ofctl: br-ex is not a bridge or a socket
> 2017-06-22 09:20:30.814 | ovs-ofctl: br-int is not a bridge or a socket
> 
> 
> stack@of-sw:~/workspace/devstack$ sudo ovs-vsctl show
> 10a3e1a0-80c5-4d16-9a02-92f98f22761e
> Bridge br-int
> Port br-int
> Interface br-int
> type: internal
> Bridge br-ex
> Port br-ex
> Interface br-ex
> type: internal
> 
> ___
> discuss mailing list
> disc...@openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Modify 'only' cookie

[Ben Pfaff]

On Wed, Jun 21, 2017 at 05:11:48AM +, Deepak S wrote:
> Hello,
> 
> 
> Is there a way to update a flow by changing only the cookie field?
> 
> 
> What works:
> 
> ovs-ofctl mod-flows br0 cookie=1/-1,cookie=2,actions=normal
> 
> What I would like to have:
> 
> ovs-ofctl mod-flows br0 cookie=1/-1,cookie=2
> 
> This is required for a particular restart scenario where stale flows are 
> deleted based on cookie.

OpenFlow doesn't have a way to modify the cookie fields without
modifying the actions.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [Question] how can I set debug log level

[Ben Pfaff]

On June 20, 2017 10:36:17 AM GMT+08:00, Margin Hu <jhu_...@qq.com> wrote:
>yes , how to do ?
>
>
>On 6/20 10:28, Ben Pfaff wrote:
>> Do you want to disable the logging or the controller?
>>
>> On Tue, Jun 20, 2017 at 10:23:39AM +0800, Margin Hu wrote:
>>> thanks , you're right , I can capture the changement  as following:
>>>
>>> 340483:2017-06-20T02:17:20.825Z|62030|vconn|DBG|unix: received:
>>> OFPT_FLOW_MOD (xid=0x2): ADD actions=NORMAL
>>> 340486:2017-06-20T02:17:20.825Z|62033|connmgr|INFO|provision<->unix:
>1
>>> flow_mods in the last 0 s (1 adds)
>>> 341089:2017-06-20T02:18:29.545Z|62376|vconn|DBG|unix: received:
>>> NXT_FLOW_MOD_TABLE_ID (xid=0x2): enable
>>> 341092:2017-06-20T02:18:29.545Z|62379|vconn|DBG|unix: received:
>>> OFPT_FLOW_MOD (xid=0x4): ADD table:1 actions=NORMAL
>>> 341095:2017-06-20T02:18:29.545Z|62382|connmgr|INFO|provision<->unix:
>1
>>> flow_mods in the last 0 s (1 adds)
>>> 341754:2017-06-20T02:19:35.742Z|62746|vconn|DBG|unix: received:
>>> NXT_FLOW_MOD_TABLE_ID (xid=0x2): enable
>>> 341757:2017-06-20T02:19:35.742Z|62749|vconn|DBG|unix: received:
>>> OFPT_FLOW_MOD (xid=0x4): DEL table:1 actions=drop
>>>
>>> but the ovs-vswitchd.log is filled with useless info  such as:
>>>
>>> 2017-06-20T02:21:31.658Z|63348|vconn|DBG|tcp:127.0.0.1:6633:
>received:
>>> OFPST_PORT_DESC request (OF1.3) (xid=0x61e993ec): port=ANY
>>> 2017-06-20T02:21:31.658Z|63349|vconn|DBG|tcp:127.0.0.1:6633: sent
>(Success):
>>> OFPST_PORT_DESC reply (OF1.3) (xid=0x61e993ec):
>>>   1(bond0): addr:24:8a:07:55:41:e8
>>>   config: 0
>>>   state:  0
>>>   speed: 0 Mbps now, 0 Mbps max
>>>   2(phy-provision): addr:26:40:cc:43:82:5d
>>>   config: 0
>>>   state:  0
>>>   speed: 0 Mbps now, 0 Mbps max
>>>   LOCAL(provision): addr:24:8a:07:55:41:e8
>>>   config: 0
>>>   state:  0
>>>   speed: 0 Mbps now, 0 Mbps max
>>>
>>> "tcp:127.0.0.1:6633" seems is a controller, I don't know what
>goodness after
>>> use it , do you know ?
>>>
>>> how can I disable it ?
>>>
>>>
>>>
>>> On 6/20 9:52, Ben Pfaff wrote:
>>>> On Tue, Jun 20, 2017 at 08:49:23AM +0800, Margin Hu wrote:
>>>>> My openvswitch  version is
>openvswitch-2.6.1-4.1.git20161206.el7.x86_64 ,
>>>>>
>>>>> can you appoint out set which module for my target ?
>>>>>
>>>>> [root@cloud-sz-master-b12-02 ~]# ovs-appctl vlog/list
>>>> I suggest "ovs-appctl vlog/set vconn".
>>> -- 
>>> -
>>> Margin Hu
>>> Love open source software
>>> mail:jhu_...@qq.com
>>> github: http://www.github.com/todaygood
>>> -
>>>
>>>
>>>
>
>-- 
>-
>Margin Hu
>Love open source software
>mail:jhu_...@qq.com
>github: http://www.github.com/todaygood
>-

Which one?___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [Question] how can I set debug log level

[Ben Pfaff]

Do you want to disable the logging or the controller?

On Tue, Jun 20, 2017 at 10:23:39AM +0800, Margin Hu wrote:
> thanks , you're right , I can capture the changement  as following:
> 
> 340483:2017-06-20T02:17:20.825Z|62030|vconn|DBG|unix: received:
> OFPT_FLOW_MOD (xid=0x2): ADD actions=NORMAL
> 340486:2017-06-20T02:17:20.825Z|62033|connmgr|INFO|provision<->unix: 1
> flow_mods in the last 0 s (1 adds)
> 341089:2017-06-20T02:18:29.545Z|62376|vconn|DBG|unix: received:
> NXT_FLOW_MOD_TABLE_ID (xid=0x2): enable
> 341092:2017-06-20T02:18:29.545Z|62379|vconn|DBG|unix: received:
> OFPT_FLOW_MOD (xid=0x4): ADD table:1 actions=NORMAL
> 341095:2017-06-20T02:18:29.545Z|62382|connmgr|INFO|provision<->unix: 1
> flow_mods in the last 0 s (1 adds)
> 341754:2017-06-20T02:19:35.742Z|62746|vconn|DBG|unix: received:
> NXT_FLOW_MOD_TABLE_ID (xid=0x2): enable
> 341757:2017-06-20T02:19:35.742Z|62749|vconn|DBG|unix: received:
> OFPT_FLOW_MOD (xid=0x4): DEL table:1 actions=drop
> 
> but the ovs-vswitchd.log is filled with useless info  such as:
> 
> 2017-06-20T02:21:31.658Z|63348|vconn|DBG|tcp:127.0.0.1:6633: received:
> OFPST_PORT_DESC request (OF1.3) (xid=0x61e993ec): port=ANY
> 2017-06-20T02:21:31.658Z|63349|vconn|DBG|tcp:127.0.0.1:6633: sent (Success):
> OFPST_PORT_DESC reply (OF1.3) (xid=0x61e993ec):
>  1(bond0): addr:24:8a:07:55:41:e8
>  config: 0
>  state:  0
>  speed: 0 Mbps now, 0 Mbps max
>  2(phy-provision): addr:26:40:cc:43:82:5d
>  config: 0
>  state:  0
>  speed: 0 Mbps now, 0 Mbps max
>  LOCAL(provision): addr:24:8a:07:55:41:e8
>  config: 0
>  state:  0
>  speed: 0 Mbps now, 0 Mbps max
> 
> "tcp:127.0.0.1:6633" seems is a controller, I don't know what goodness after
> use it , do you know ?
> 
> how can I disable it ?
> 
> 
> 
> On 6/20 9:52, Ben Pfaff wrote:
> >On Tue, Jun 20, 2017 at 08:49:23AM +0800, Margin Hu wrote:
> >>My openvswitch  version is openvswitch-2.6.1-4.1.git20161206.el7.x86_64 ,
> >>
> >>can you appoint out set which module for my target ?
> >>
> >>[root@cloud-sz-master-b12-02 ~]# ovs-appctl vlog/list
> >I suggest "ovs-appctl vlog/set vconn".
> 
> -- 
> -
> Margin Hu
> Love open source software
> mail:jhu_...@qq.com
> github: http://www.github.com/todaygood
> -
> 
> 
> 
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [Question] how can I set debug log level

[Ben Pfaff]

On Tue, Jun 20, 2017 at 08:49:23AM +0800, Margin Hu wrote:
> My openvswitch  version is openvswitch-2.6.1-4.1.git20161206.el7.x86_64 ,
> 
> can you appoint out set which module for my target ?
> 
> [root@cloud-sz-master-b12-02 ~]# ovs-appctl vlog/list

I suggest "ovs-appctl vlog/set vconn".
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] [Question] how can I set debug log level

[Ben Pfaff]

On Tue, Jun 20, 2017 at 06:22:00AM +0800, Margin Hu wrote:
> I found  flow rule entry in flow table is dropped by unknow cause,  you can
> review detail issue at
> 
> https://bugs.launchpad.net/bugs/1697243
> 
> now,  how can I  set debug log level  ?

You can use ovs-appctl to set the log levels for a running process.
Please see its manpage for details.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Can you match on packet length?

[Ben Pfaff]

On Mon, Jun 19, 2017 at 03:11:23PM -0700, Joe Stringer wrote:
> On 19 June 2017 at 14:08, Keith Holleman via discuss
>  wrote:
> >
> > Went to add a match rule based on packet length and couldn't find the right
> > term or syntax.  Then I searched the ovs-ofctl man page and couldn't find
> > anything.  Even more surprising was that I couldn't find an example of this
> > previously being asked on any OVS discussion groups.   Couldn't find it in
> > the FAQ either.  I know in stating so many places looked I will inevitably
> > get proven wrong but I did try.
> >
> > So.can you add a rule via ovs-ofctl to match on a packet's length?
> > Probably tricky because it would have to have greater-than and less-than
> > semantics and that doesn't play well with the fixed-length matches of
> > OpenFlow.  Are there any extensions that enable this?  Or any pointers to
> > previous discussions about it?
> 
> I don't think there's any such support. The matching in the datapath
> is based on unique bit patterns and masks, so if you wanted to render
> range semantics into that then you would practically end up generating
> several matches, one for each bitpattern within the range. I could
> imagine that if the packet length were matchable then worst case we
> could end up generating a different datapath flow for each possible
> packet length (or at least one per power-of-two bit up to the argument
> for lt/gt).

A arbitrary integer range over an N-bit field requires no more than 2N
bitwise matches.

The OpenFlow working group declined to define this kind of match on the
basis, among others, that the length of a packet is rather difficult to
define.  Also, no one seemed to have a really good use case for it.  The
main one raised was some kind of vague "security" justification.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] LLDP handing in OVS

[Ben Pfaff]

On Fri, Jun 09, 2017 at 03:12:45PM +0200, sothy shan wrote:
> I am playing in writing funny SDN controller and play with LLDP message.
> To test my SDN controller, I am using mininet with ovs (2.0.2) .
> 
> I send LLDP message. The message I sent is given below except Ethernet
> headers.
> 
> Chassisi TLV: 01 09 04 00 00 00 00 00 00 00 03  (from data path id) In this
> case, I didnt take 48 bit, took 64 ok??
> Port TLV: 02 07 03 62 03 03 c7 99 e1   (mac addres of port )
> ttl: 03 02 00 78
> 
> My question if I send only three data, OVS will repond ? it needed to some
> data?.

For the most part, OVS doesn't implement LLDP--that's the responsibility
of a controller.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] If ovs doesn't masking process in flow_tbl_lookup, What are the disadvantages ?

[Ben Pfaff]

With microflow cache only, there's one cache miss per microflow.
With an effective megaflow cache, the miss rate is much lower.

The paper gives examples for your other question.

On Fri, Jun 09, 2017 at 01:47:14PM +0900, Heung Sik Choi wrote:
> Thanks to reply.
> 
> Regarding your second question, which flow rules are you asking about?
> 
> I just want to know that when ovs had only an in-kernel microflow cache,
> How many in-kernel cache miss there are, and  also how much improving
>  there are when using megaflow.
> 
> And I have a last question. short-lived flows were main problem when there
> was micro flow cache. I don't know the circumstance where short lived
> connection happen often. can you let me know about it?
> 
> 
> Thanks!!
> 
> 
> 2017-06-09 13:29 GMT+09:00 Ben Pfaff <b...@ovn.org>:
> 
> > When OVS had only an in-kernel microflow cache, there were at least two
> > reasons for performance problems with many short-lived flows.  The first
> > was the cost of sending packets to userspace.  The second was the cost
> > of translating the packets through the entire OpenFlow pipeline.  The
> > megaflow cache solves both problems: it eliminates both the
> > kernel-to-user-to-kernel transition and (if the megaflow cache is
> > effective) the additional OpenFlow translations.
> >
> > Regarding your second question, which flow rules are you asking about?
> >
> > On Fri, Jun 09, 2017 at 12:22:09PM +0900, Heung Sik Choi wrote:
> > > What does "if only the microflow cache works" mean?
> > >
> > > sorry to confuse you. I'm not good at English.
> > >
> > > In the paper, the authors say that at start of ovs implementation, there
> > > has been microflow cache(EMC), but megaflow wasn't implemented.
> > >
> > > At that time, they say  there was a problem caused by short lived
> > > connections.
> > >
> > > I want to know the problem detail. I guess that the problem made ovs get
> > > flow rules from userlevel, and that made performance degradation.
> > >
> > > also I want to know that when not using megaflow, are there big
> > difference
> > > in aspect of number of flow rules.If there are the big difference, can
> > you
> > > tell me the difference in numbers?
> > >
> > >
> > >
> > > 2017-06-09 0:43 GMT+09:00 Ben Pfaff <b...@ovn.org>:
> > >
> > > > On Thu, Jun 08, 2017 at 04:33:54PM +0900, Heung Sik Choi wrote:
> > > > > 1. If only microflow cache works and there are many short lived
> > > > > connections, does it make many tuples in table, and does it suffers
> > > > serious
> > > > > performance degradation by the many tuples(very many tuple makes
> > context
> > > > > switching to Userlevel)?
> > > >
> > > > What does "if only the microflow cache works" mean?
> > > >
> > > > > 2. I know that OVS uses hash for table index. if only microflow
> > cache and
> > > > > there are small amount of entries in the table, is it possible to
> > find
> > > > > entry in O(1) time?
> > > >
> > > > Yes.
> > > >
> > > > > 3. In paper, microflow and megaflow use the word 'cache'. However,
> > when I
> > > > > open the OVS code, it looks like they are in main memory area.
> > doesn't
> > > > the
> > > > > 'cache' mean hardware cache?
> > > >
> > > > No.
> > > >
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] If ovs doesn't masking process in flow_tbl_lookup, What are the disadvantages ?

[Ben Pfaff]

When OVS had only an in-kernel microflow cache, there were at least two
reasons for performance problems with many short-lived flows.  The first
was the cost of sending packets to userspace.  The second was the cost
of translating the packets through the entire OpenFlow pipeline.  The
megaflow cache solves both problems: it eliminates both the
kernel-to-user-to-kernel transition and (if the megaflow cache is
effective) the additional OpenFlow translations.

Regarding your second question, which flow rules are you asking about?

On Fri, Jun 09, 2017 at 12:22:09PM +0900, Heung Sik Choi wrote:
> What does "if only the microflow cache works" mean?
> 
> sorry to confuse you. I'm not good at English.
> 
> In the paper, the authors say that at start of ovs implementation, there
> has been microflow cache(EMC), but megaflow wasn't implemented.
> 
> At that time, they say  there was a problem caused by short lived
> connections.
> 
> I want to know the problem detail. I guess that the problem made ovs get
> flow rules from userlevel, and that made performance degradation.
> 
> also I want to know that when not using megaflow, are there big difference
> in aspect of number of flow rules.If there are the big difference, can you
> tell me the difference in numbers?
> 
> 
> 
> 2017-06-09 0:43 GMT+09:00 Ben Pfaff <b...@ovn.org>:
> 
> > On Thu, Jun 08, 2017 at 04:33:54PM +0900, Heung Sik Choi wrote:
> > > 1. If only microflow cache works and there are many short lived
> > > connections, does it make many tuples in table, and does it suffers
> > serious
> > > performance degradation by the many tuples(very many tuple makes context
> > > switching to Userlevel)?
> >
> > What does "if only the microflow cache works" mean?
> >
> > > 2. I know that OVS uses hash for table index. if only microflow cache and
> > > there are small amount of entries in the table, is it possible to find
> > > entry in O(1) time?
> >
> > Yes.
> >
> > > 3. In paper, microflow and megaflow use the word 'cache'. However, when I
> > > open the OVS code, it looks like they are in main memory area. doesn't
> > the
> > > 'cache' mean hardware cache?
> >
> > No.
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] If ovs doesn't masking process in flow_tbl_lookup, What are the disadvantages ?

[Ben Pfaff]

On Thu, Jun 08, 2017 at 04:33:54PM +0900, Heung Sik Choi wrote:
> 1. If only microflow cache works and there are many short lived
> connections, does it make many tuples in table, and does it suffers serious
> performance degradation by the many tuples(very many tuple makes context
> switching to Userlevel)?

What does "if only the microflow cache works" mean?

> 2. I know that OVS uses hash for table index. if only microflow cache and
> there are small amount of entries in the table, is it possible to find
> entry in O(1) time?

Yes.

> 3. In paper, microflow and megaflow use the word 'cache'. However, when I
> open the OVS code, it looks like they are in main memory area. doesn't the
> 'cache' mean hardware cache?

No.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How can I change OpenVSwitch

[Ben Pfaff]

Maybe someone else can help.  I'm lost.

On Wed, Jun 07, 2017 at 08:25:09PM +0200, Simone Aglianò wrote:
> Ok I have to choose reply all
> My question is the initial question regarding the architecture that I have
> attached, I don t want to turn my switch in a Open Flow switch but I would
> use it like that with Open V Switch I mean I want take the advantage of the
> forwarding capacity of the Catalyst cisco switch but in the meantime I
> would use the Open Flow features.
> 
> So If I send a openflow command from the Floodlight, does my architecture
> work as if I had an openflow switch?
> 
> Thank you.
> 
> On 7 Jun 2017 20:17, "Ben Pfaff" <b...@ovn.org> wrote:
> 
> > On Wed, Jun 07, 2017 at 06:48:41PM +0200, Simone Aglianň wrote:
> > > What do you mean with " don t drop the mail list"?
> >
> > This discussion is on a mailing list, but you keep dropping the CC.
> >
> > > Do you mean If I connect the machine where OpenVSwitch is running with
> > the
> > > Catalyst Cisco Switch that does not support, it will work?
> >
> > Is your actual question, "How do I use Open vSwitch to turn a Cisco
> > switch into an OpenFlow switch?"  If that is your question, then the
> > answer is you cannot.  Only Cisco can turn a Cisco switch into an
> > OpenFlow switch, so you should contact their sales or support teams with
> > questions about how to do that.
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How can I change OpenVSwitch

[Ben Pfaff]

Please don't drop the mailing list.

Open vSwitch is an OpenFlow switch, so of course it works as if you have
an OpenFlow switch: you do!

On Wed, Jun 07, 2017 at 05:42:37PM +0200, Simone Aglianò wrote:
> So if I run Floodlight and OpenVSwitch in the same laptop and I just
> connect the switch with an Ethernet port everything works as I had an
> OpenFlow Switch?
> 
> 
> On Wed, Jun 7, 2017 at 5:37 PM, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Wed, Jun 07, 2017 at 02:50:25PM +0200, Simone Aglianň wrote:
> > > I am working on a project with some challenges and I would like to know
> > > which part of Open vSwitch can I change to reach my goal to communicate
> > > with a Catalyst Cisco Switch that does not support OpenFlow.
> >
> > Open vSwitch supports Ethernet ports, as do many Cisco switches, so with
> > an appropriate cable, you may be able to connect them.  This doesn't
> > require any changes to Open vSwitch.
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] How can I change OpenVSwitch

[Ben Pfaff]

On Wed, Jun 07, 2017 at 02:50:25PM +0200, Simone Aglianò wrote:
> I am working on a project with some challenges and I would like to know
> which part of Open vSwitch can I change to reach my goal to communicate
> with a Catalyst Cisco Switch that does not support OpenFlow.

Open vSwitch supports Ethernet ports, as do many Cisco switches, so with
an appropriate cable, you may be able to connect them.  This doesn't
require any changes to Open vSwitch.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] meter action for OF1.5

[Ben Pfaff]

I forgot to thank you for working on this.  Thanks!

ovs_instruction_name_from_type() takes an ovs_instruction_type, so its
result should be invariant of the OpenFlow version.

ofpact_is_apply_actions() is only used in one place, to encode
instructions, so it should be straightforward to update it.

But please keep in mind the big picture.  OVS tries to keep as much
compatibility as it can among the various OpenFlow versions it supports.
In this case, I guess this means that if a meter action is added via
OpenFlow 1.5, and then dumped in OpenFlow 1.4 or earlier, then it should
be translated to a meter instruction if it is in an appropriate position
for that, or to an extension action that we would define if it is not.

On Tue, Jun 06, 2017 at 03:25:41PM +, Ali Volkan Atli wrote:
> ovs_instruction_name_from_type and ofpact_is_apply_actions
> 
> Meter is instruction for OF1.3 but it is an action-set for OF1.5, so I need 
> to distinguish
> 
> - Volkan
> 
>  Ben Pfaff wrote 
> 
> On Tue, Jun 06, 2017 at 12:37:56PM +, Ali Volkan Atli wrote:
> > I'm trying to contribute meter action for OF1.5 by using the existing
> > patch "https://patchwork.ozlabs.org/patch/731701/; in which meter is
> > instruction. Currently it seems working using ovs-ofctl, but I am not
> > sure if I can pass the version info as a parameter to the related
> > function, I need a little help from people who know OvS architecture.
> 
> "pass the version info as a parameter to the related function"?  What
> function?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] meter action for OF1.5

[Ben Pfaff]

On Tue, Jun 06, 2017 at 12:37:56PM +, Ali Volkan Atli wrote:
> I'm trying to contribute meter action for OF1.5 by using the existing
> patch "https://patchwork.ozlabs.org/patch/731701/; in which meter is
> instruction. Currently it seems working using ovs-ofctl, but I am not
> sure if I can pass the version info as a parameter to the related
> function, I need a little help from people who know OvS architecture.

"pass the version info as a parameter to the related function"?  What
function?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Questions about IPFIX

[Ben Pfaff]

On Fri, Jun 02, 2017 at 02:42:49PM -0700, Greg Rose wrote:
> On 06/02/2017 09:08 AM, Ben Pfaff wrote:
> >ipfix_cache_entry_init() initializes the start and end to be the same:
> >
> > xgettimeofday();
> > entry->flow_end_timestamp_usec = now.tv_usec + 100LL * 
> > now.tv_sec;
> > entry->flow_start_timestamp_usec = entry->flow_end_timestamp_usec;
> >
> >and then I don't see anything that ever updates the "end" time.  I guess
> >that's the issue.
> 
> From looking at the code it appears to me that the ipfix_cache_update() 
> function isn't doing
> anything to update the time stamps.  It updates the stats but the stats 
> function doesn't update
> time stamps.  From what I can tell though that seems the place where it 
> should be updated.  I can
> try whipping up a bug fix patch for this if ipfix_cache_update() is the right 
> place to do the time
> stamp update.

Seems reasonable.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Questions about IPFIX

[Ben Pfaff]

ipfix_cache_entry_init() initializes the start and end to be the same:

xgettimeofday();
entry->flow_end_timestamp_usec = now.tv_usec + 100LL * now.tv_sec;
entry->flow_start_timestamp_usec = entry->flow_end_timestamp_usec;

and then I don't see anything that ever updates the "end" time.  I guess
that's the issue.

On Fri, Jun 02, 2017 at 01:57:01PM +0200, Felix Konstantin Maurer wrote:
> I just took a look at the code, and I think it should work, as the the
> timestamps are properly updated in ipfix_cache_aggregate_entries.
> However, I don't know the code base well or C for that matter.
> Can you point me to where the time values are not properly set?
> 
> Regards
> Felix
> 
> On 05/31/2017 05:27 PM, Ben Pfaff wrote:
> > On Wed, May 31, 2017 at 03:05:45PM +0200, Felix Konstantin Maurer wrote:
> >> I'm currently trying to export flow statistics from an Openvswitch
> >> bridge with IPFIX. It works, however in almost all packets, the
> >> flowStartDeltaMicroseconds and the flowEndDeltaMicroseconds have the
> >> same value. Is that to be expected? Do I misunderstand something?
> >>
> >> I expected them to be different and to allow me to calculate over which
> >> time period the octetDeltaCount number of bytes of the flow had been sent.
> > 
> > The IPFIX code hard-codes the start and end times to the same value.
> > It's too bad--it would be nice to have this be more accurate.  I don't
> > know how difficult that would be.
> > 
> 
> -- 
> RWTH Aachen University
> Communication and Distributed Systems
> http://www.comsys.rwth-aachen.de
> 


___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] LISP Tunneling

[Ben Pfaff]

On Wed, May 17, 2017 at 02:52:39PM +0200, Ashish Kurian wrote:
> Dear OVS folks,
> 
> I have some doubts regarding LISP tunneling. I have a setup where I am
> getting incoming LISP tunneled packets into my OVS. What I want to do is to
> check the inner IP destination address and based on that I need to forward
> the packets. Let us say that there are only two possibilities for inner IP
> addresses : 10.0.0.1 and 10.0.0.2.
> 
> If the inner IP address is 10.0.0.1, then I want the packet to be forwarded
> to an interface (say eth1 and port number 1) without doing any change to
> the tunneled packet. If the inner IP address is 10.0.0.2, then I want the
> packet to be forwarded to the another interface (say eth2 and port number
> 2) with only the inner contents of the tunnel packets.
> 
> I am thinking of the following flow entries to do the mentioned rules, but
> correct me if I am wrong.
> 
> *
>  table=0,dl_type=0x0800,nw_dst=10.0.0.2,actions=mod_dl_dst=10:0:0:2,output:2*
> 
>-  Will this flow check for the inner destination IP of the
>tunneled packet and put only the metadata in port eth2?

I don't know what it means to "put only the metadata" in a port.  What
does it mean?

>   *table=0,dl_type=0x0800,action=NORMAL*
> 
> 
>- Will this flow take care of all other flows?

Yes.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] socket handle leak when pkill/kill -9 ovsdb-server/ovs-vswitchd

[Ben Pfaff]

On Wed, May 31, 2017 at 03:26:04AM +, fukaige wrote:
> I am using ovs-2.5.2 and find there may be socket handle leak when kill 
> ovsdb-server/ovs-vswitchd thread.
> When I use “pkill -9 ovs-vswitchd” kill ovs-vswitchd thread and restart it, 
> the socket file “ovs-vswitchd.12237.ctl” isn’t closed.

There's no way for a process to delete files when it's killed with
signal 9.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Questions about IPFIX

[Ben Pfaff]

On Wed, May 31, 2017 at 03:05:45PM +0200, Felix Konstantin Maurer wrote:
> I'm currently trying to export flow statistics from an Openvswitch
> bridge with IPFIX. It works, however in almost all packets, the
> flowStartDeltaMicroseconds and the flowEndDeltaMicroseconds have the
> same value. Is that to be expected? Do I misunderstand something?
> 
> I expected them to be different and to allow me to calculate over which
> time period the octetDeltaCount number of bytes of the flow had been sent.

The IPFIX code hard-codes the start and end times to the same value.
It's too bad--it would be nice to have this be more accurate.  I don't
know how difficult that would be.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] GOTO_TABLE jumping back to lower numbered table

[Ben Pfaff]

On Tue, May 30, 2017 at 09:51:58AM -0700, Saurabh Shrivastava (सौरभ श्रीवास्तव) 
wrote:
> GOTO_TABLE only allows jumping to a higher number table. Is it to prevent
> loops? Is there a way to bypass this restriction?
> 
> BTW man page for GOTO_TABLE doesnt state this restriction.

OpenFlow restricts goto_table this way.  You can refer to the OpenFlow
specification for full details of OpenFlow.

You can use "resubmit", instead, to jump backward.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] OpenFlow stack protocol written in Java fro the switch

[Ben Pfaff]

Since you're not talking about Open vSwitch, you might want to find
someplace other than ovs-discuss to ask your question.  I imagine that
there is a Floodlight mailing list around somewhere.

On Tue, May 30, 2017 at 06:30:20PM +0200, Simone Aglianò wrote:
> I m not talking about Open vSwitch but I would create a software that is
> able to interface with the controller and with a switch that doesn t
> support Open Flow, so I need an Open Flow interface to communicate with the
> controller and with the other interface(like ssh or telnet) I will
> comunicate with the switch
> 
> On Mon, May 29, 2017 at 8:50 PM, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Mon, May 29, 2017 at 04:47:26PM +0200, Simone Aglianň wrote:
> > > I would like to find the openflow stack of the switch implemented in Java
> > > in order to use it like an interface with the Floodlight controller.
> > > I was looking on the code of Floodlight controller regarding the
> > > *OFSwitchHandshakeHandler.java* file, but I have checked that it doesn t
> > > support the state machine of the switch and it generates an exception.
> > > Then, before start over the implementation of that by myself I would ask
> > if
> > > there is that implementation online.
> >
> > The Open vSwitch handshake isn't different from that of any other
> > OpenFlow switch.  I'm pretty surprised to hear that Floodlight breaks
> > with it.
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Limit Number of Flows

[Ben Pfaff]

On Mon, May 29, 2017 at 09:42:55AM +, Asha  P wrote:
>   I am using OpenvSwitch as one of the VNF ( Virtual Network
>   Function). I wanted to know if there any way where we can limit the
>   number of flows in switch, So that it can be used in my usecase. For
>   Example, I must be able to add only 5 flows in table 0 and the sixth
>   flow should be added to table 1.

See the documentation of the Flow_Table table in
ovs-vswitchd.conf.db(5).
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Debian install dep

[Ben Pfaff]

Sure, although it's better to email the patch to the dev list.

On Mon, May 29, 2017 at 09:10:23PM +0100, Mircea Ulinic wrote:
> Thanks, Ben.
> For future reference, if I notice anything else that can be
> improved/adjusted etc., can I just send a pull request directly to
> https://github.com/openvswitch/ovs?
> 
> On 29 May 2017 at 19:41, Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Mon, May 29, 2017 at 12:39:15AM +0100, Mircea Ulinic wrote:
> > > Hi,
> > >
> > > I just installed ovs on a fresh Debian, following the instructions from
> > > http://docs.openvswitch.org/en/latest/intro/install/debian/
> > >
> > > After the build, it complained:
> > >
> > > admin@ip-172-31-5-175:~$ sudo dpkg -i openvswitch-switch_2.7.0-1_
> > amd64.deb
> > > (Reading database ... 169467 files and directories currently installed.)
> > > Preparing to unpack openvswitch-switch_2.7.0-1_amd64.deb ...
> > > Unpacking openvswitch-switch (2.7.0-1) over (2.7.0-1) ...
> > > dpkg: dependency problems prevent configuration of openvswitch-switch:
> > >  openvswitch-switch depends on uuid-runtime; however:
> > >   Package uuid-runtime is not installed.
> > >
> > >
> > > Installing uuid-runtime solved the problem.
> > > I thought this info may help.
> > >
> > > Cheers & many thanks for maintaining the project!
> >
> > OK, I sent a patch adding a note to the documentation:
> > https://mail.openvswitch.org/pipermail/ovs-dev/2017-May/
> > 333077.html
> >
> 
> 
> 
> -- 
> -Mircea
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] OpenFlow stack protocol written in Java fro the switch

[Ben Pfaff]

On Mon, May 29, 2017 at 04:47:26PM +0200, Simone Aglianò wrote:
> I would like to find the openflow stack of the switch implemented in Java
> in order to use it like an interface with the Floodlight controller.
> I was looking on the code of Floodlight controller regarding the
> *OFSwitchHandshakeHandler.java* file, but I have checked that it doesn t
> support the state machine of the switch and it generates an exception.
> Then, before start over the implementation of that by myself I would ask if
> there is that implementation online.

The Open vSwitch handshake isn't different from that of any other
OpenFlow switch.  I'm pretty surprised to hear that Floodlight breaks
with it.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Debian install dep

[Ben Pfaff]

On Mon, May 29, 2017 at 12:39:15AM +0100, Mircea Ulinic wrote:
> Hi,
> 
> I just installed ovs on a fresh Debian, following the instructions from
> http://docs.openvswitch.org/en/latest/intro/install/debian/
> 
> After the build, it complained:
> 
> admin@ip-172-31-5-175:~$ sudo dpkg -i openvswitch-switch_2.7.0-1_amd64.deb
> (Reading database ... 169467 files and directories currently installed.)
> Preparing to unpack openvswitch-switch_2.7.0-1_amd64.deb ...
> Unpacking openvswitch-switch (2.7.0-1) over (2.7.0-1) ...
> dpkg: dependency problems prevent configuration of openvswitch-switch:
>  openvswitch-switch depends on uuid-runtime; however:
>   Package uuid-runtime is not installed.
> 
> 
> Installing uuid-runtime solved the problem.
> I thought this info may help.
> 
> Cheers & many thanks for maintaining the project!

OK, I sent a patch adding a note to the documentation:
https://mail.openvswitch.org/pipermail/ovs-dev/2017-May/333077.html
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] SYN packet mirroring

[Ben Pfaff]

On Thu, May 25, 2017 at 10:26:29AM -0700, Justin Pettit wrote:
> 
> > On May 25, 2017, at 2:10 AM, Avi Cohen (A)  wrote:
> > 
> > Hi All,
> > I need to capture all received SYN packets from all interfaces and to 
> > mirror/output to a specific interface in addition to the operational 
> > interface that these packets should be forwarded.
> > Can I do it with a single dpctl  add-flow cli command ?  and not modify the 
> > 'operational' flows that are used to normally connect TCP clients to TCP 
> > servers ?
> 
> No, if you run ovs-vswitchd, it will be confused when flows are added with 
> ovs-dpctl, and delete them.  Also, I don't think that would work, since the 
> kernel module will only apply actions from a single flow, so you'll either 
> send the SYN packet to your collector or forward it appropriately, but not 
> both.
> 
> You should be able to construct what you want pretty easily with ovs-ofctl 
> flows, though.

Avi might be talking about "dpctl" from the OpenFlow reference
implementation, which (confusingly) uses OpenFlow.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Nat functionality using Openflow, - Flow not allowed

[Ben Pfaff]

On Wed, May 24, 2017 at 01:23:00PM +0530, rahul b wrote:
> Hi i am trying to emulate nat functionality with a flow like this.
> 
> ovs-ofctl add-flow s1 -Oopenflow13
> priority=100,dl_type=0x806,in_port=1,action=mod_nw_src=10.10.10.3,output:2
> 
> this results in an error
> 
> ovs-ofctl: none of the usable flow formate (OpenFlow10,NXM) is among the
> allowed flow formats (OXM-OpenFlow13)

This is because mod_nw_src works with IPv4.

You can use something like this instead:
ovs-ofctl add-flow br0 'arp,action=set_field:10.10.10.3->arp_spa'
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] DUMP-FLOWS: ofctl dump-flows vs dpctl dump-flows

[Ben Pfaff]

On Sun, May 21, 2017 at 12:11:19PM +, Avi Cohen (A) wrote:
> Megaflows  is a  kind of  'flows/rules compression' with don't care bits. 
> (correct me if I'm wrong)
> But this is in a contradiction to what is written in the manpages: " dpctl 
> shows only exact-match flows of
>   packets that  traverse the datapath lately"

If a manpage says that, then it is out of date and should be updated.  I
cannot see what manpage contains that statement.  Can you point it out,
so that we can fix it?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] IGMP snooping control for static groups

[Ben Pfaff]

On Thu, May 18, 2017 at 08:44:22PM +, Ajit Warrier wrote:
> Some legacy multicast applications do not support IGMP join/leaves and
> hence they do not get registered on the switch's MDB. Is there a command to
> add such multicast groups into the MDB as static entries that do not time
> out ?

I don't think that OVS has that feature.  I imagine that we'd accept a
patch to add it.  If you're planning to work on it, it might be worth
discussing the design before jumping in.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] DUMP-FLOWS: ofctl dump-flows vs dpctl dump-flows

[Ben Pfaff]

On Thu, May 18, 2017 at 01:36:54PM +, Avi Cohen (A) wrote:
> I'm seeing in my dpctl dump-flows   about 10,000  flows,  while in ofctl 
> dump-flows I see only ~ 1400 flows 
> - Is this make sense ?

If you have a lot of traffic, and a complicated OpenFlow flow table,
then you might see two datapath flows (in ovs-dpctl) for every TCP
connection.

> - How can I associate an ofcctl flow to a dpctl flow ?  (I can see ufid only 
> in dpctl dump)

You can use "ovs-appctl ofproto/trace".
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] In-band for SDN

[Ben Pfaff]

On Tue, May 16, 2017 at 02:07:20PM +0800, 赵鑫 wrote:
> Hello,
> I am a newcomer to SDN, and I have read your OVS documents on how to 
> implement In-band control.
> 
> 
> But I am still confused on how the control path is set up. Is it
> available to customize the control path for the In-band mode ? Could
> you please give me a brief explication ?

OVS in-band control can't be customized.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Preferable usecase for storing ovs port extra information?

[Ben Pfaff]

On Tue, May 16, 2017 at 12:12:51PM +0300, Nir Magnezi wrote:
> Hi All,
> 
> Context:
> In OpenStack Neutron, we have a script named ovs_cleanup that basically
> deletes all ovs ports in the boot process.
> For some use cases (as mentioned here
> https://review.openstack.org/#/c/460524/ ]), we would like to be able to
> skip the cleanup for specific ports.
> 
> From what I currently know, there are two possible ways do to that:
> 1. Use 'other_config', by appending: -- set Interface myport
> other_config:skip_cleanup=true"
> 
> 2. Use 'external_ids. by appending ': -- set Interface myport
> external_ids:skip_cleanup=true"
> 
> In the patch I already mentioned, I was told that: using external_ids is
> safer than other_config as it won't be interpreted by the switch (with the
> reference http://docs.openvswitch.org/en/latest/topics/integration/ )
> 
> Will usage of 'external_ids' be a better practice to achieve that goal?
> I would really appreciate your feedback on this.

Yes, this reasoning is correct.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

[ovs-discuss] trying to track down a database paper

[Ben Pfaff]

I'm trying to track down a paper that I read several years ago and I'm
having no luck finding it.  Maybe someone here can help.

The paper was about how SQL-style transactions, that is, ones that can
be opened and include a series of query and update operations and then
committed, can be layered on top of systems that only support executing
atomic read-modify-write type operations, along with an explanation of
why this can be good for performance and simplify system design.  I
think it was new around the time that I was initially building OVSDB, so
probably 2010 or so, and I think that it was in a journal or magazine,
e.g. something like Operating System Review, rather than a conference.

I'm interested in it because this is basically how OVSDB works in
practice and I'd like to be able to cite the paper as an extended
explanation.

Anyone have any hints?

Thanks,

Ben.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] duplicate instruction

[Ben Pfaff]

On Sat, May 13, 2017 at 09:24:57PM +0200, matecs wrote:
> i'm trying to set up a flow with the attached capture but ovs refuses it
> with duplicate instruction.
> what do i wrong? how can i apply more actions to a flow through openflow?
> (ps: if i leave only one of the instructions in the packet, ovs accepts it
> so tlvs should be ok...)

This flow has 3 apply-actions instructions.  OpenFlow only allows one
instruction of a given type.  Instead of using multiple apply-actions
instructions, put all of the actions in one apply-actions instruction.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] 回复: How to make the actions of vxlan in port?

[Ben Pfaff]

OVS doesn't support actions for adding or removing VXLAN headers, but
you can get much the same effect by using VXLAN ports.

The code you cite doesn't have anything to do with what you're asking
about.

On Fri, May 12, 2017 at 11:31:08AM +0800, y...@panath.cn wrote:
> Does not support this feature?, I can have such a kernel code:
> 
> do_execute_actions
> ===>execute_masked_set_action
>===>
> 
> 
> 
> 
> y...@panath.cn
>  
> 发件人： y...@panath.cn
> 发送时间： 2017-05-12 10:33
> 收件人： ovs-discuss
> 主题： How to make the actions of vxlan in port?
> Hi everyone,
> I have a doubt about tunnel protocol to add and subtract action implementation
> 
> If you want to enter the port of packets increase or stripping vxlan action, 
> how to implement, this port is just a ordinary port?
> 
> Whether can be like increase or stripping vlan implementation, such as:
> 
> push vlan：
> ovs-ofctl add-flow br0 
> priority=1,in_port=1,dl_vlan=0x,actions=mod_vlan_vid:3,output:2
> 
> pop vlan：
> ovs-ofctl add-flow br0 in_port=1,actions=strip_vlan,output:2
> Best wishes!
> 
> 
> 
> y...@panath.cn



___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Failed to add ovs bridge

[Ben Pfaff]

I'm not aware of such a problem.

If you can explain how to reproduce the problem, we'll fix it.

On Fri, May 12, 2017 at 03:56:51AM +, fukaige wrote:
> Nothing seems unusual in setup. Just hit this by accident. I am checking that 
> if there is any chance failing to delete the netdev.
> 
> > -Original Message-----
> > From: Ben Pfaff [mailto:b...@ovn.org]
> > Sent: Friday, May 12, 2017 10:50 AM
> > To: fukaige
> > Cc: ovs-discuss@openvswitch.org; j...@ovn.org
> > Subject: Re: Failed to add ovs bridge
> > 
> > Are you aware of anything unusual in your setup?  We have not had any
> > similar reports.
> > 
> > On Fri, May 12, 2017 at 01:34:46AM +, fukaige wrote:
> > > I am not using STP/RSTP.I saw the bug fix you mentioned, seems it is
> > irrelevant to my problem.
> > > May be there is some race condition lead to deleting netdev in
> > > netdev_shash. But, I cannot figure it out right now.
> > >
> > > The occurrence probability is very low. I just hit this for three times 
> > > in two
> > month.
> > >
> > > > -Original Message-
> > > > From: Ben Pfaff [mailto:b...@ovn.org]
> > > > Sent: Thursday, May 11, 2017 9:32 PM
> > > > To: fukaige
> > > > Cc: ovs-discuss@openvswitch.org; j...@ovn.org
> > > > Subject: Re: Failed to add ovs bridge
> > > >
> > > > Are you using STP or RSTP?  There's a bug fix related to them on
> > branch-2.5.
> > > >
> > > > On Thu, May 11, 2017 at 11:11:02AM +, fukaige wrote:
> > > > > Hi all,
> > > > >
> > > > > Occasionally, I get error when creating a bridge using “ovs-vsctl
> > > > > add-br
> > > > br-eth”
> > > > >
> > > > >
> > > > > ovs-vsctl: Error detected while setting up 'br-eth'.  See
> > > > > ovs-vswitchd log for
> > > > details.
> > > > >
> > > > >
> > > > > Ovs-vswitched log is below:
> > > > >
> > > > >
> > 2017-05-11T03:45:25.293Z|00026|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports recirculation
> > > > >
> > 2017-05-11T03:45:25.293Z|00027|ofproto_dpif|INFO|system@ovs-system:
> > > > > MPLS label stack length probed as 1
> > > > >
> > 2017-05-11T03:45:25.293Z|00028|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports unique flow ids
> > > > >
> > 2017-05-11T03:45:25.293Z|00029|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports ct_state
> > > > >
> > 2017-05-11T03:45:25.293Z|00030|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports ct_zone
> > > > >
> > 2017-05-11T03:45:25.293Z|00031|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports ct_mark
> > > > >
> > 2017-05-11T03:45:25.293Z|00032|ofproto_dpif|INFO|system@ovs-system:
> > > > > Datapath supports ct_label
> > > > > 2017-05-11T03:45:25.364Z|1|ofproto_dpif_upcall(handler226)|INF
> > > > > O|re ceived packet on unassociated datapath port 0
> > > > > 2017-05-11T03:45:25.368Z|00033|netdev_linux|WARN|ethtool
> > command
> > > > > ETHTOOL_GFLAGS on network device br-eth failed: No such device
> > > > > 2017-05-11T03:45:25.368Z|00034|dpif|WARN|system@ovs-system:
> > failed
> > > > to
> > > > > add br-eth as port: No such device
> > > > > 2017-05-11T03:45:25.368Z|00035|bridge|INFO|bridge br-eth: using
> > > > > datapath ID 2a51cf9f2841
> > > > > 2017-05-11T03:45:25.368Z|00036|connmgr|INFO|br-eth: added service
> > > > controller "punix:/var/run/openvswitch/br-eth.mgmt"
> > > > >
> > > > > Then I delete the br-eth, then try to add it. But, still get same 
> > > > > error as
> > above.
> > > > However, bridge which name is different from br-eth can be created
> > > > successfully.
> > > > >
> > > > > Some clues:
> > > > >
> > > > > 1.   As I kown, the port br-eth’s type is internel, and there is 
> > > > > no
> > way to
> > > > get into netdev_linux_ethtool_set_flag(). But, the log shows that
> > > > request.type is wrong.
> > > > > request.type get wrong value OVS_VPORT_TYPE_NETDEV instead of
> > > > OVS

Re: [ovs-discuss] Flow Entry Statistics Trigger support

[Ben Pfaff]

I'm not aware of any.

On Fri, May 12, 2017 at 06:44:30AM +, Rohith Basavaraja wrote:
> Hi Ben,
> 
> Thanks for sharing the patch details. Looked at the patch diff and
> looks like these patches contains changes to add Support for Extensible Flow 
> Entry Statistics (EXT-334)
> and doesn’t contains changes to add support for Flow Entry Statistics Trigger 
> (EXT-335).
> 
> Is there any ongoing work to add support for Flow Entry Statistics Trigger 
> (EXT-335)?
> or this work is not started yet?
> 
> Thanks
> Rohith
> 
> 
> 
> 
> 
> -- Forwarded message --
> From: Ben Pfaff <b...@ovn.org<mailto:b...@ovn.org>>
> Date: Thu, May 11, 2017 at 7:12 PM
> Subject: Re: [ovs-discuss] Flow Entry Statistics Trigger support
> To: Rohith Basavaraja 
> <rohith.basavar...@gmail.com<mailto:rohith.basavar...@gmail.com>>
> Cc: ovs-discuss@openvswitch.org<mailto:ovs-discuss@openvswitch.org>
> 
> 
> On Thu, May 11, 2017 at 12:08:25PM +0530, Rohith Basavaraja wrote:
> > Wanted to check whether anyone working on the Flow Entry Statistics Trigger
> > support in OVS or this work not started yet?.If anyone started working on
> > this what is targeted OVS release this feature will be supported.
> 
> There's some work toward this feature:
> https://patchwork.ozlabs.org/patch/761074/
> https://patchwork.ozlabs.org/patch/761075/
> 
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Adding Ovs capabilities to physical server

[Ben Pfaff]

Sure, you can run OVS on a physical server.

On Fri, May 12, 2017 at 01:38:51PM +, DELARUELLE Marc-Laurent wrote:
> Hi,
> If NSX runs as OvS, you may control the flows between VM, and between VM and 
> physical servers.
> But you can't control the flows between physical servers.
> 
> Consider 
> - You create a VLAN and a 512 IP subnet befind a firewall
> - You want to create some DMZ using micro segmentation on this subnet
> - You need to deploy  VM and physical servers 
> 
> Using Microsegmentation, you may isolate VM from differents DMZ in the same 
> VLAN. And you may control which physical servers may be reached from which  
> VM according DMZ membership. 
> But you can't control the flow between 2 physical servers on this subnet.
> Using PVLAN, ACL on physical switches or ACL on the physical servers is not 
> an option in my company.
> 
> This is why I wonder if a solution may be considered by ovs for physical 
> servers running Oracle for instance.  Oracle can't be virtualized for licence 
> reasons.
> I imagine a micro vswitch with physical nics on one side and virtual nics, 
> seen from the host, but not using VM or KVM.
> And this micro vswitch could be managed to connect to VxLan and accepting a 
> DFW centrally managed.
> 
> Another use case could be also Network Attach Storage. 
> Regards
> MLD
> 
> 
> -Message d'origine-
> De : Ben Pfaff [mailto:b...@ovn.org] 
> Envoyé : vendredi 12 mai 2017 15:25
> À : DELARUELLE Marc-Laurent <marc-laurent.delarue...@renault.com>
> Cc : ovs-discuss@openvswitch.org
> Objet : Re: [ovs-discuss] Adding Ovs capabilities to physical server
> 
> On Fri, May 12, 2017 at 11:38:19AM +, DELARUELLE Marc-Laurent wrote:
> > I'm currently looking at NSX in my company. NSX is very nice, but it is 
> > limited to VM world, as Ovs.
> > I may suggest to make a tiny Ovs for Linux or Windows physical server. It 
> > may be like a driver, adding VxLan native connectivity and DFW capabilities 
> > to physical servers.
> > 
> > Perhaps this project already exists ?
> 
> What does this need that OVS doesn't already have?
> -- Disclaimer  
> Ce message ainsi que les eventuelles pieces jointes constituent une 
> correspondance privee et confidentielle a l'attention exclusive du 
> destinataire designe ci-dessus. Si vous n'etes pas le destinataire du present 
> message ou une personne susceptible de pouvoir le lui delivrer, il vous est 
> signifie que toute divulgation, distribution ou copie de cette transmission 
> est strictement interdite. Si vous avez recu ce message par erreur, nous vous 
> remercions d'en informer l'expediteur par telephone ou de lui retourner le 
> present message, puis d'effacer immediatement ce message de votre systeme.
> 
> *** This e-mail and any attachments is a confidential correspondence intended 
> only for use of the individual or entity named above. If you are not the 
> intended recipient or the agent responsible for delivering the message to the 
> intended recipient, you are hereby notified that any disclosure, distribution 
> or copying of this communication is strictly prohibited. If you have received 
> this communication in error, please notify the sender by phone or by replying 
> this message, and then delete this message from your system.
> 
> ___
> discuss mailing list
> disc...@openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] load:OXM_OF_ETH_DST[25..32]->NXM_NX_REG0[0..7] problem

[Ben Pfaff]

On Fri, May 12, 2017 at 07:12:32PM +, Ali Volkan Atli wrote:
> table=0,priority=0,dl_type=0x0800,actions=load:OXM_OF_ETH_DST[25..32]->NXM_NX_REG0[0..7],goto_table=1

This should be "move", not "load".
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Adding Ovs capabilities to physical server

[Ben Pfaff]

On Fri, May 12, 2017 at 11:38:19AM +, DELARUELLE Marc-Laurent wrote:
> I'm currently looking at NSX in my company. NSX is very nice, but it is 
> limited to VM world, as Ovs.
> I may suggest to make a tiny Ovs for Linux or Windows physical server. It may 
> be like a driver, adding VxLan native connectivity and DFW capabilities to 
> physical servers.
> 
> Perhaps this project already exists ?

What does this need that OVS doesn't already have?
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Failed to add ovs bridge

[Ben Pfaff]

Are you aware of anything unusual in your setup?  We have not had any
similar reports.

On Fri, May 12, 2017 at 01:34:46AM +, fukaige wrote:
> I am not using STP/RSTP.I saw the bug fix you mentioned, seems it is 
> irrelevant to my problem.
> May be there is some race condition lead to deleting netdev in netdev_shash. 
> But, I cannot figure
> it out right now.
> 
> The occurrence probability is very low. I just hit this for three times in 
> two month.
> 
> > -Original Message-
> > From: Ben Pfaff [mailto:b...@ovn.org]
> > Sent: Thursday, May 11, 2017 9:32 PM
> > To: fukaige
> > Cc: ovs-discuss@openvswitch.org; j...@ovn.org
> > Subject: Re: Failed to add ovs bridge
> > 
> > Are you using STP or RSTP?  There's a bug fix related to them on branch-2.5.
> > 
> > On Thu, May 11, 2017 at 11:11:02AM +, fukaige wrote:
> > > Hi all,
> > >
> > > Occasionally, I get error when creating a bridge using “ovs-vsctl add-br
> > br-eth”
> > >
> > >
> > > ovs-vsctl: Error detected while setting up 'br-eth'.  See ovs-vswitchd 
> > > log for
> > details.
> > >
> > >
> > > Ovs-vswitched log is below:
> > >
> > > 2017-05-11T03:45:25.293Z|00026|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports recirculation
> > > 2017-05-11T03:45:25.293Z|00027|ofproto_dpif|INFO|system@ovs-system:
> > > MPLS label stack length probed as 1
> > > 2017-05-11T03:45:25.293Z|00028|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports unique flow ids
> > > 2017-05-11T03:45:25.293Z|00029|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports ct_state
> > > 2017-05-11T03:45:25.293Z|00030|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports ct_zone
> > > 2017-05-11T03:45:25.293Z|00031|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports ct_mark
> > > 2017-05-11T03:45:25.293Z|00032|ofproto_dpif|INFO|system@ovs-system:
> > > Datapath supports ct_label
> > > 2017-05-11T03:45:25.364Z|1|ofproto_dpif_upcall(handler226)|INFO|re
> > > ceived packet on unassociated datapath port 0
> > > 2017-05-11T03:45:25.368Z|00033|netdev_linux|WARN|ethtool command
> > > ETHTOOL_GFLAGS on network device br-eth failed: No such device
> > > 2017-05-11T03:45:25.368Z|00034|dpif|WARN|system@ovs-system: failed
> > to
> > > add br-eth as port: No such device
> > > 2017-05-11T03:45:25.368Z|00035|bridge|INFO|bridge br-eth: using
> > > datapath ID 2a51cf9f2841
> > > 2017-05-11T03:45:25.368Z|00036|connmgr|INFO|br-eth: added service
> > controller "punix:/var/run/openvswitch/br-eth.mgmt"
> > >
> > > Then I delete the br-eth, then try to add it. But, still get same error 
> > > as above.
> > However, bridge which name is different from br-eth can be created
> > successfully.
> > >
> > > Some clues:
> > >
> > > 1.   As I kown, the port br-eth’s type is internel, and there is no 
> > > way to
> > get into netdev_linux_ethtool_set_flag(). But, the log shows that 
> > request.type
> > is wrong.
> > > request.type get wrong value OVS_VPORT_TYPE_NETDEV instead of
> > OVS_VPORT_TYPE_INTERNAL.
> > >
> > > static int
> > > dpif_netlink_port_add__(struct dpif_netlink *dpif, struct netdev *netdev,
> > > odp_port_t *port_nop)
> > > OVS_REQ_WRLOCK(dpif->upcall_lock)
> > > {
> > >  ……
> > >
> > > if (request.type == OVS_VPORT_TYPE_NETDEV) { #ifdef _WIN32
> > > /* XXX : Map appropiate Windows handle */ #else
> > > netdev_linux_ethtool_set_flag(netdev, ETH_FLAG_LRO, "LRO",
> > > false); #endif }
> > >
> > > ……
> > > }
> > >
> > >
> > > 2.   Debug ovs-vswitchd with gdb. I find that there is a netdev with
> > same name was not deleted(lib/netdev.c:netdev_open).
> > > netdev_open (name=0x6000d6b0 "br-int", type=0x52ca80 "internal",
> > netdevp=0xfc20fab8, netdevp@entry=0xfc20fb28)
> > > at lib/netdev.c:354
> > > 354  {
> > > (gdb) n
> > > 358  netdev_initialize();
> > > (gdb)
> > > 360  ovs_mutex_lock(_class_mutex);
> > > (gdb)
> > > 361  ovs_mutex_lock(_mutex);
> > > (gdb)
> > > 360  ovs_mutex_lock(_class_mutex);
> > > (gdb)
> > > 361  ovs_mutex_lock(_mutex);
> > > (gdb)
> > > 362  netdev = shash_find_data(_shash, name);
> > > (gdb)
> > > 363  if (!netdev) {
> > > (gdb) print netdev->name
> > > $1 = 0x47852e0 "br-int"
> > > (gdb) print netdev->refcnt
> > > There is no member named refcnt.
> > > (gdb) n
> > > 405  netdev->ref_cnt++;
> > > (gdb) print netdev->ref_cnt
> > > $2 = 2
> > > (gdb) n
> > > 406  *netdevp = netdev;
> > > (gdb) print netdev->ref_cnt
> > > $3 = 3
> > >
> > > There must be something wrong when deleting bridge. But, I cannot find out
> > a way to reproduce it and why it was not deleted correctly. Is
> > > any can offer some suggestions to reproduce the error or solve it?
> > >
> > > Note:
> > > ovs version: 2.5.2
> > > kernel version: 4.1
> > >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] Anyway to move/load data between ct_mark/ct_label and field/register ?

[Ben Pfaff]

What's your high-level goal?  Most fields will be the same on every trip
through the connection tracker for a given (half-)connection, so there
would be limited value in that.

On Thu, May 11, 2017 at 02:37:13PM +, Zang MingJie wrote:
> It can only set to immediately value, I want to copy some field into
> ct_label
> 
> On Thu, May 11, 2017, 21:30 Ben Pfaff <b...@ovn.org> wrote:
> 
> > On Thu, May 11, 2017 at 11:11:58AM +, Zang MingJie wrote:
> > > I wonder if there is any way to move data from or to ct_label like nicira
> > > move or load extension ?
> > >
> > > I want to store some value of packet field or register to ct_lable on the
> > > first packet, and the load ct_label to some fields of the following
> > packet.
> >
> > ovs-ofctl(8) says, in the documentation for the ct action:
> >
> >  exec([action][,action...])
> > Perform  actions  within the context of
> > connection
> > tracking. This is  a  restricted  set  of
> > actions
> > which  are  in the same format as their
> > specifica‐
> > tions as part of a flow. Only actions which
> > modify
> > the ct_mark or ct_label fields are accepted
> > within
> > the exec action, and these fields may only be
> > mod‐
> > ified with this option. For example:
> >
> > set_field:value[/mask]->ct_mark
> >Store a 32-bit metadata value with the
> > con‐
> >nection.  Subsequent lookups for
> > packets in
> >this  connection  will populate the
> > ct_mark
> >flow field when the packet is sent  to
> > the
> >connection  tracker  with  the table
> > speci‐
> >fied.
> >
> > set_field:value[/mask]->ct_label
> >Store a 128-bit  metadata  value  with
> > the
> >connection.  Subsequent lookups for
> > packets
> >in  this  connection  will   populate
> >  the
> >ct_label flow field when the packet is
> > sent
> >to the connection tracker  with  the
> > table
> >specified.
> >
> > The  commit  parameter  must  be  specified to
> > use
> > exec(...).
> >
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] a struct to keep flow scores

[Ben Pfaff]

OK.  I'm not certain why you need new high-level data structures for
this.  The code already propagates information from the datapath into
the OpenFlow flows and I think that you can just follow the existing
pattern to convey your new data.

On Thu, May 11, 2017 at 01:50:32PM +, Avi Cohen (A) wrote:
> Thanks Ben,
> We need to test another eviction criteria - not LRU, 
> but learn the flows other scores   such as flow effective time / flow usage , 
> and to do eviction based on these scores
> Best Regards
> avi
> 
> > -Original Message-
> > From: Ben Pfaff [mailto:b...@ovn.org]
> > Sent: Thursday, 11 May, 2017 4:42 PM
> > To: Avi Cohen (A)
> > Cc: ovs-discuss@openvswitch.org
> > Subject: Re: [ovs-discuss] a struct to keep flow scores
> > 
> > OVS already has mechanisms for this kind of thing.  For datapath flows, OVS
> > already evicts flows based on these criteria.  For OpenFlow flows, OVS 
> > already
> > supports approximate-LRU eviction.  I'm not sure what you're going to
> > additionally get out of adding more to these mechanisms.  Do you have a 
> > high-
> > level goal?
> > 
> > On Thu, May 11, 2017 at 06:42:26AM +, Avi Cohen (A) wrote:
> > > Thanks  Ben,
> > > The general idea is to store flow info (pkt, bytes, last_time_used)
> > > which is updated every dumping phase and learn about flow behavior To
> > calculate  flow-usage and and flow-span scores in order to evict  flows 
> > based
> > this criteria.
> > >
> > > I thought of 2 implementation methods:
> > > 1st  impl. Methiod
> > > Store   the flow statistics  (pkt, bytes, last_time_used)  which is 
> > > updated every
> > dumping phase in a  flow ring-buffer
> > > So we need to store info  of flows that are in openflow tables that
> > > were installed  at least one time  at dp-cache and currently are or 
> > > aren't in
> > dp-cache The question is where to allocate this ring buffer ?
> > >
> > > 2nd impl. Method:
> > > Maintain a global sliding window contains all flows statistics, then
> > > at eviction phase - calculate the scores based on the sliding window
> > information The question is where to allocate this sliding window ?
> > > Best Regards
> > > avi
> > >
> > >
> > > > -Original Message-
> > > > From: Ben Pfaff [mailto:b...@ovn.org]
> > > > Sent: Thursday, 11 May, 2017 6:49 AM
> > > > To: Avi Cohen (A)
> > > > Cc: ovs-discuss@openvswitch.org
> > > > Subject: Re: [ovs-discuss] a struct to keep flow scores
> > > >
> > > > On Wed, May 10, 2017 at 07:12:27AM +, Avi Cohen (A) wrote:
> > > > > I need to keep some score information per flow - which exist in
> > > > > the open-flow
> > > > tables at user space , not necessarily in the data-path cache.
> > > > > These scores are learned during flow dumping phase - but should be
> > > > > saved
> > > > also when flow is evicted from data-path.
> > > > > The udpif_key struct is not suitable for this since it exist only
> > > > > when flow exist
> > > > in the data-path.
> > > > > What is the suitable structure to keep this info ?
> > > >
> > > > What kind of information is this?  How to deal with it depends on what 
> > > > it is.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss