On Wed, May 25, 2016 at 12:12:13PM +0500, Abdul Basit wrote:
> For an experiment, instead of throttling bandwidth in mininet using 'bw'
> parameter. My objective is to create virtual interfaces with different
> speeds. As per my understanding, OVS creates virtual interfaces using veth
> pair on beh
On Wed, May 25, 2016 at 05:42:57PM +0200, Zhongliang Zhao wrote:
> I am using OpenWrt on TL-1043ND. I installed Open vSwitch on OpenWrt using
> the Open vSwitch package from the OpenWrt repository.
> "ps w | grep ovs" returns "ovsdb-server" and "ovs-vswitchd" are running
> fine.
>
> My goal is to
On Wed, May 25, 2016 at 09:27:31AM -0500, Ryan Moats wrote:
> As I understand it, Table 0 identifies the logical port and logical
> flow. I'm worried that this means we'll end up with separate bucket
> rules for each ingress port of the port pairs that make up a port
> group, leading to a cardinali
On Wed, May 25, 2016 at 08:05:38AM +, Chris Packham wrote:
> We've run into a problem trying to compile openvswitch with CFLAGS=-Werror.
Use "configure --enable-Werror", as documented in Open vSwitch's
INSTALL.md file.
___
discuss mailing list
discus
; thanks again!
>
>
>
>
> 祝好!
> from: wangwei
>
> From: Ben Pfaff
> Date: 2016-05-22 03:18
> To: wangweib...@163.com
> CC: discuss
> Subject: Re: [ovs-discuss] can ovs support qos according to ip and
Yes, just set the output queue depending on the IP and MAC.
___
discuss mailing list
discuss@openvswitch.org
http://openvswitch.org/mailman/listinfo/discuss
On Sat, May 21, 2016 at 10:13:06PM +0500, Tooba Ahsen wrote:
> I need to change a small amount of code on the ofproto folder that
> exists in openvswitch. I try to run the ./configure and make commands
> afterwards but it doesn't really seem to be working because the new
> version of the functions
16-05-21 23:36 , Ben Pfaff Wrote: On Sat, May 21, 2016 at
> 02:48:12PM +0800, hanxueluo wrote: > why use recirculation action for
> TCP-mode bond? what's the benefits, > compared to the old implement?
> Performance is much better. > Is there any other situation using
>
On Sat, May 21, 2016 at 02:48:12PM +0800, hanxueluo wrote:
> why use recirculation action for TCP-mode bond? what's the benefits,
> compared to the old implement?
Performance is much better.
> Is there any other situation using recirculation action?
Recirculation is used for MPLS, for connectio
On Fri, May 20, 2016 at 10:50:31AM -0500, Ryan Moats wrote:
> However, both of the above are currently optional things that *could*
> be added by hand. Because of our CI/CD requirements, we'll need to feed
> it with fully decorated files. Maintaining a local branch manually
> in this scenario is no
ost Unreachable
> ^C
> --- 192.168.56.103 ping statistics ---
> 5 packets transmitted, 0 received, +3 errors, 100% packet loss, time 4024ms
> pipe 3
>
>
>
> On Fri, May 20, 2016 at 5:22 PM, Ben Pfaff wrote:
>
> > It seems OK.
> >
> > When I use a sim
t;OVS isn't even opening a UDP socket. What's in your database--can you
> >send along /etc/openvswitch/conf.db (or whatever file you're using)?
>
> I configured 192.168.56.103:2055 for NetFlow.
>
> Kindly check the attached file.
>
> On Fri, May 20, 2016 at
On Fri, May 20, 2016 at 12:50:13PM +0200, Syed Faraz Ali wrote:
> > It's probably easier to debug if you only enable NetFlow for now.
>
> Only NetFlow has been enabled.
>
> > Next, I'd use "strace" to see whether ovs-vswitchd is really trying to
> > send data to the collector. It would be using
On Thu, May 19, 2016 at 05:51:19PM -0500, Ryan Moats wrote:
>
> As I mentioned at today's IRC meeting - we have a need to include measuring
> the performance of the OVN control channel under scale as part of our CI/CD
> test suite, which is driving us to examine the feasibility of some sort of
> l
Please don't drop the mailing list.
On Thu, May 19, 2016 at 09:02:41PM +0200, Syed Faraz Ali wrote:
> The packets are being collected on the sFlow port 6343 at the same time,
> but giving unreachable error for NetFlow. This looks pretty awkward!
> Do you have some suggestions to further look into
On Thu, May 19, 2016 at 03:34:27PM -0700, Ashwin Paranjpe wrote:
> ovs-vswitchd and ovsdb-server both support the --monitor option. According
> to the man page, --monitor creates an additional monitoring-process to
> monitor the daemon.
>
> How do we monitor the health of the monitoring-process it
On Thu, May 19, 2016 at 06:17:45PM +0200, Syed Faraz Ali wrote:
> Hello Ben,
>
> Thanks for the prompt response.
>
> Please have the output from the logs. Below are the logs related to
> collector, I'm not sure if its related to NetFlow collector or used in some
> other context. Kindly suggest fu
On Thu, May 19, 2016 at 04:24:44PM +, Rajasekaran, Monica wrote:
> Today when working on my OVS, all commands were failing. Looks like
> "/usr/local/etc/openvswitch/conf.db" got deleted somehow.
>
> How do I fix this? Or do I have to do a fresh install again ?
As packaged for most distributi
On Thu, May 19, 2016 at 05:08:40PM +0200, Syed Faraz Ali wrote:
> I've configured NetFlow on the openvswitch, but I'm not getting any output
> on the collector machine. I've checked using tcpdump and Wireshark on the
> collector machine but couldn't observe any NetFlow related traffic.
>
> *Enable
On Mon, May 16, 2016 at 11:34:20AM +0200, Alessio Giorgetti wrote:
> Dear Ben,
>
> thank you a lot for your suggestions... I just set the forward-bpdu="true"
> in the bridge using the command:
>
> > ovs-vsctl set bridge br0 other-config:forward-bpdu=true
>
> and this is now correctly configured.
Regards -- Pippin
>
> On Tue, May 17, 2016 at 8:05 PM, Ben Pfaff wrote:
>
> > On Mon, May 16, 2016 at 04:08:53PM -0600, Pippin Wallace wrote:
> > > First off I apologize for submitting a "bug" on such an old version but
> > > what I am trying to
On Mon, May 16, 2016 at 04:08:53PM -0600, Pippin Wallace wrote:
> First off I apologize for submitting a "bug" on such an old version but
> what I am trying to find is the first "stable" OVS version where this bug
> is fixed and move it it for now.
>
>
>
> The situation is that the majority of o
On Tue, May 17, 2016 at 01:24:15PM -0400, Flaviof wrote:
> Question: do you guys oppose enabling color output [1] in the AT unit tests?
> I would like to add that to the top level .at file in the OVS repo
> if you think this is a good idea...
Fine with me.
For best compatibility with older Autoco
On Tue, May 10, 2016 at 12:40:13PM +, O'Reilly, Darragh wrote:
>
> > I think I see the problem. Please try the following patch in addition to
> > the previous one:
>
> Thanks Ben - that fixes it.
Thanks you for testing. I applied this patch to master, branch-2.5, and
branch-2.4.
The othe
On Fri, May 13, 2016 at 02:06:33PM -0400, Krishna Pratap wrote:
> I am trying to use the "ovs-ofctl mod-port switch port " command to
> bring down the patch port I have on my switch . I am using the patch port
> to create a link between two switches. I am using ovs 2.1 version.
>
> When I use the
asses.
> Can you please help ?
> Thanks
>
> On Sat, May 14, 2016 at 9:37 PM, Ben Pfaff wrote:
>
> > You started out by saying that you created the QoS classes. Just do
> > that from your controller.
> >
> > On Sat, May 14, 2016 at 04:01:35PM -0700, Aravind Ga
QoS classes when there is a congestion.
> Can you please help
> Thanks
>
> On Sat, May 14, 2016 at 3:26 PM, Ben Pfaff wrote:
>
> > You could write a controller that monitors link utilization and enables
> > QoS.
> >
> > On Sat, May 14, 2016 at 02:02:39PM -0700
On Sat, May 14, 2016 at 01:12:43PM -0500, Yanko Marín Muro wrote:
> We performed a latency test with cbench to the REFERENCE MiniNet controller
> (OVS-Controlller) and obtained the best result when compared with
> OpenDayLight, Onos, RYU, FLOODLIGHT, POX controllers.
>
>
>
> in which programmi
do it in the controller
>
> Thanks
>
>
> On Sat, May 14, 2016 at 1:58 PM, Ben Pfaff wrote:
>
> > You could use a controller.
> >
> > You are probably wasting your effort, though, because the QoS in use
> > probably does not do anything if the link is not
t;
> Thanks
>
> On Sat, May 14, 2016 at 11:51 AM, Ben Pfaff wrote:
>
> > On Fri, May 13, 2016 at 11:19:19PM -0700, Aravind Ganesan wrote:
> > > I have created QoS classes, and am able to separate the traffic based on
> > > source address.
> > > Wh
> On controller:
> PACKET IN ARRIVED
> OFPPacketIn received: buffer_id= total_len=220 reason=ACTION
> table_id=0 cookie=0 match=OFPMatch(oxm_fields={'in_port': 1})
>
> >Пятница, 13 мая 2016, 18:14 +03:00 от Ben Pfaff :
> >
> >On Fri, May 1
On Fri, May 13, 2016 at 11:19:19PM -0700, Aravind Ganesan wrote:
> I have created QoS classes, and am able to separate the traffic based on
> source address.
> What I want to do now is enable QoS classes only when link is congested.
> For example, if there are 3 hosts, h1, h2, h3, connected by sing
On Fri, May 13, 2016 at 06:49:33PM +0530, Abhijit Bhadra wrote:
> Is Multiple VLAN headers supported in ovs 2.3.0?
No.
> I want to achieve q-in-q by doing push_vlan operations. If it is not
> supported, which version of OVS do I need to upgrade to?
No version of OVS supports multiple VLANs.
___
On Fri, May 13, 2016 at 12:19:34PM +0300, Alexander Beetle wrote:
> Running OF1.3, I set match condition to OVS with certain flow specification
> and an instruction to send that flow to a controller.
>
> $ sudo ovs-ofctl dump-flows s2
> NXST_FLOW reply (xid=0x4):
> cookie=0x0, duration=22.075s,
trying to figure out how to
> match the packet-in with flow-mod or packet-out? If I use xid to match
> packet-in and flow-mod or packet-out, does it violate specs?
>
> - Volkan
> ________
> From: Ben Pfaff [b...@ovn.org]
> Sent: Thursday, May 1
On Fri, May 13, 2016 at 02:34:34PM +0200, Alessio Giorgetti wrote:
> I'm trying to use the Ryu controller on a very simple network composed of
> five Open vSwitch 2.4 (no loop is present currently in the topology).
>
> I would enable STP at the controller, so followed the indications found on
> th
On Thu, May 12, 2016 at 04:53:39PM +0200, Gerhard Stenzel wrote:
> On 05/04/2016 04:54 PM, Ben Pfaff wrote:
> ...
>
> >
> >Thanks, I received it. I don't think it's necessary to copy the list;
> >probably I'm the only one who will look at it.
> >
On Thu, May 12, 2016 at 01:46:32PM +, Ali Volkan Atli wrote:
> When a controller adds a table-miss flow entry with
> instruction->action->max_lenght field set to NO_BUFFER, do I have a
> chance to match a packet-in message and associated flow-mod message
> via buffer_id?
I don't understand th
t noticed your answer.
>
> Peter Gubka
>
> -Original Message-
> From: Peter Gubka -X (pgubka - PANTHEON TECHNOLOGIES at Cisco)
> Sent: Wednesday, May 04, 2016 9:00 AM
> To: 'Ben Pfaff'
> Cc: b...@openvswitch.org
> Subject: RE: [ovs-discuss] controller'
On Tue, May 10, 2016 at 05:48:51AM +, Kai Ma (kaima) wrote:
> In OVS, there is a command to add meter to switch, but seems no command to
> attach the meter to flow.
> So, how to attach a meter to a flow? Else the meter is specific to a
> switch only.
The "meter" instruction is part of a flow.
On Mon, May 09, 2016 at 06:44:18PM -0700, Dio Vinra wrote:
> I used eclipse IDE with CDT. I choose "makefile project with existing code"
> option and point it to OVS folder where the makefile.am is stored, but the
> console said "make: *** No rule to make target 'all'. Stop." when I tried
> to buil
On Thu, May 05, 2016 at 02:39:34PM +, Dominick Krachtus (TT) wrote:
> Below are the steps i have taken to setup our current testing environment.
> As of now this environment is up and functioning with multicast working.
> There are however some oddities with multicast. Testing with omping
> pro
t;
> On Monday, 9 May 2016, Ben Pfaff wrote:
>
> > On Mon, May 09, 2016 at 03:25:09AM -0700, Dio Vinra wrote:
> > > Can anybody point me to some openvswitch development tools or unit
> > testing
> > > tools? Are there any script/tools to facilitate deployment, tes
On Mon, May 09, 2016 at 04:33:54PM +, O'Reilly, Darragh wrote:
>
> > I sent out a patch for formal review:
> >http://openvswitch.org/pipermail/dev/2016-May/070666.html
>
> Thanks Ben. I tested on master and it fixes the mdb problem.
>
> The next problem is that any existing flows don
On Mon, May 09, 2016 at 10:33:24PM +, Maurice Qureshi wrote:
> Could this be a sandbox issue? Similar setup (Linux containers and
> OVS) works with ODL controller by programming the flows manually.
You're in the sandbox? The sandbox doesn't work with real system ports
(that's the point).
On Mon, May 09, 2016 at 03:25:09AM -0700, Dio Vinra wrote:
> Can anybody point me to some openvswitch development tools or unit testing
> tools? Are there any script/tools to facilitate deployment, testing or so?
Some developers use the ovs-dev.py script, from the utilities directory.
> Also, are
On Mon, May 09, 2016 at 03:02:02PM +0900, Borasy Kong wrote:
> So I've been trying to log in OVS using both OVS_NLERR() and printk() but I
> cannot see my message in 'dmesg' at all.
> Can someone tell me if there is a proper way to log in OVS?
There's nothing special about logging in the OVS kerne
On Mon, May 09, 2016 at 10:50:04AM +0530, jagadish krishnappa wrote:
> To increase the number of flow-mods per second, we are planning to spawn a
> new thread to handle the entire connection manager functionality.
> A multi thread functionality to handle all OF connections to a OVS switch.
I don't
On Thu, May 05, 2016 at 04:04:42PM +, O'Reilly, Darragh wrote:
>
> > Does the following patch make a difference?
>
> Yes it does - thanks. But now we have another problem when the video
> app restarts. Still trying to pin it down.
Thanks for testing.
I sent out a patch for formal review:
On Fri, May 06, 2016 at 11:02:54PM +, Maurice Qureshi wrote:
> Thank Ben.
>
> Now in order to see if traffic is really getting forwarded by the flows
> created by OVN and OVS, I have done following.
>
> 1) Created two Linux Containers with MAC addresses as 00::01 and 00::02. They
> are in
You'd want to enhance netdev_get_queue() to support these properties,
and then push that down to every netdev implementation via netdev_class.
On Sat, May 07, 2016 at 12:05:49PM +, fatt 3issam wrote:
> It's obvious that it's hard coded. The correct values should be read
> from the actual queue
On Thu, May 05, 2016 at 06:09:13PM +, Chen, Weiwen wrote:
> Try to find out how to configure that for vswitch so I can adjust ARP
> cache TTL per my needs. Please let me know if you know the answer.
OVS has an ARP cache used for userspace tunneling. If that's the one
you mean, then I don't th
On Fri, May 06, 2016 at 09:14:53AM +, fatt 3issam wrote:
> Could you please tell me what this means in ofproto/ofproto.c in the method
> put_get_config_reply(...)
> (https://github.com/openvswitch/ovs/blob/master/ofproto/ofproto.c#L6271) :
> /* None of the existing queues have compatible pro
On Thu, May 05, 2016 at 09:19:56PM +, Maurice Qureshi wrote:
> I am tracing the packet using the ovn/env1/packet1.sh. Trace seems to
> show that it is hitting the entry in flow table and eventually action
> is output port 2. However, The n_packets counters for any of the entry
> in flow table d
ing an old thread. To clarify, does the packet itself
> > contain the tunnel_key of the datapath binding? If so, what exactly is the
> > structure of a packet in this case or where in the packet would I find this
> > field?
> >
> > Thanks,
> >
> > Brendan
On Wed, May 04, 2016 at 02:44:38PM +0200, Gerhard Stenzel wrote:
>
>
> On 05/03/2016 05:25 PM, Ben Pfaff wrote:
> >On Tue, May 03, 2016 at 01:53:42PM +0200, Gerhard Stenzel wrote:
> ...
> >>[openvswitch 2.5.90] testsuite: 744 745 818 949 980 1026 1087 failed
> >
I don't think that HTB works on a virtual interface anyhow; I believe
that it needs a fixed link speed.
I think that mininet can do what you want, so you might want to ask on a
mininet mailing list.
On Tue, May 03, 2016 at 07:29:58PM +, fatt 3issam wrote:
> Thank you.If I understood, No versi
On Tue, May 03, 2016 at 07:02:45AM +, Peter Gubka -X (pgubka - PANTHEON
TECHNOLOGIES at Cisco) wrote:
> Are you sure about "reconnecting" switches? As i wrote before , to reproduce
> the problem, i had to use 2 switches/bridges.
You're right, now that I look again. I missed the differences
On Tue, May 03, 2016 at 12:22:26AM -0700, Aravind Ganesan wrote:
> Hi
> I am trying to develop QoS in mininet OVS switches. I am using Mininet VM
> 2.2.1.
> I referred to earlier messages in this list and from that, I tried the
> "dpctl add-queue s1 1 q-id 1 " but it says "command not found".
OVS
On Tue, May 03, 2016 at 11:38:49AM +, fatt 3issam wrote:
> When executing the command: # ovs-ofctl queue-get-config swicth switch-port,
> i only get the queue id!!
> OFPT_QUEUE_GET_CONFIG_REPLY (OF1.3) (xid=0x4): port=2queue 0:
> I want to get the queue properties also!
> my queue configurati
On Tue, May 03, 2016 at 01:53:42PM +0200, Gerhard Stenzel wrote:
>
> On 05/02/2016 10:36 PM, Ben Pfaff wrote:
> >What if we just make ofp_port_t 32 bits? That has to be the long-term
> >trend anyway because OpenFlow 1.1+ uses a 32-bit port number. The
> >following pass
On Sat, Apr 23, 2016 at 10:19:13AM +0200, Tomasz Łukojko wrote:
> Problem start when i want tio add port to br:
> "vs-vsctl add-port br0 gre1 -- set interface gre1 type=gre
> ovs-vsctl: Error detected while setting up 'gre1'. See ovs-vswitchd log
> for details."
> and now i cant add any port. I t
On Tue, Apr 19, 2016 at 05:30:21PM +, Farhad Sunavala wrote:
> >"normal" consists of a number of more or less independent steps:
> > * Drop certain malformed or invalid frames.
> > * Check VLAN.
> > * Check for other inadmissible frames.
> > * Learn source MAC.
> >
On Fri, Apr 22, 2016 at 09:32:26AM +, Peter Gubka -X (pgubka - PANTHEON
TECHNOLOGIES at Cisco) wrote:
> Hello,
>
> I had to use 2 switches/bridges to reproduce the problem. Logs in attachments.
>
> Just for the time orientation:
> Enabling 2 masters for 2 switches (controller firstly sent sl
which handles the arp request but when it comes
> to send the reply from the kernel via br0 it doesn't work
>
>
> thank you
>
> 2016-05-02 21:29 GMT+01:00 Ben Pfaff :
>
> > Please don't drop the mailing list.
> >
> > What does it mean to "out
What if we just make ofp_port_t 32 bits? That has to be the long-term
trend anyway because OpenFlow 1.1+ uses a 32-bit port number. The
following passes all the unit tests on i386 (though I haven't carefully
looked at all the uses of ofp_port_t), what about big-endian systems?
diff --git a/inclu
Please don't drop the mailing list.
What does it mean to "output the response of the kernel"?
On Mon, May 02, 2016 at 09:19:47PM +0100, oualid hjira wrote:
> thanks Ben,
> but how can we output the response of the kernel ??
>
> 2016-05-02 20:28 GMT+01:00 Ben Pfaff
On Tue, Apr 26, 2016 at 04:43:38PM +, O'Reilly, Darragh wrote:
> We have an application that we need to be able to restart quickly. It
> listens on a multicast address, so a restart causes IGMPv3 leave
> packets to be sent and join packets soon after. OVS250 receives the
> leave packet and remo
controller(if mismatch). what i'm
> interested about is to send packets directly to the kernel where i can
> activate a routing daemon.
>
> Best regards
>
> 2016-05-02 19:01 GMT+01:00 Ben Pfaff :
>
> > On Tue, Apr 26, 2016 at 05:42:28PM +0100, oualid hjira wrote:
&g
On Tue, Apr 26, 2016 at 05:42:28PM +0100, oualid hjira wrote:
> im using OvS and i want to send packets directly to the kernel, i've made
> some researches i found that OFPP_NORMAL action will handle this problem
> but the packet are either sent to controller neither to the kernel
What does "the p
On Thu, Apr 28, 2016 at 11:21:24PM +, Kai Ma (kaima) wrote:
> > 在 2016年4月28日,下午10:01,Ben Pfaff 写道:
> >
> > On Thu, Apr 28, 2016 at 01:50:31PM +, Kai Ma (kaima) wrote:
> >>> 在 2016年4月28日,下午9:36,Ben Pfaff 写道:
> >>>> On Thu, Apr 28, 2016 at 08:3
On Mon, Apr 18, 2016 at 07:38:43AM +, Ali Volkan Atli wrote:
> @Ben, the proposal is from
> (https://mailman.stanford.edu/pipermail/openflow-discuss/2010-June/thread.html#1024).
> Also
> I don't think it violets the Openflow specs. Could you please give us
> more information?
It breaks LLDP:
On Sat, Apr 30, 2016 at 08:55:41PM +0200, Guillermo Chica wrote:
> Hi,
>
> I have created a topology with mininet. When I've added the links, I've
> used the delay parameter, as in self.addLink( node1, node2, delay='5ms').
> After running the topology with sudo mn --custom topo5g.py --topo topo5g
On Mon, May 02, 2016 at 08:35:39AM +, Majid Taleqani wrote:
> i have a host that sends 100 tcp syn packets per second which every packet
> has spoofed source IP and as a result they all will be sent to the POX
> controller. as i monitor by Wireshark i see that the first 250s packetIns
>
On Thu, Apr 28, 2016 at 07:08:08PM +0800, Haosen Wen wrote:
> I defined a new action with type=0x1000(4096) in loxigen through floodlight
> controller. But the decode result in ovs is type=0x10(16).
What do you mean, "the decode result in ovs is type=0x10(16)." OVS
doesn't decode actions to a raw
On Thu, Apr 28, 2016 at 08:32:32AM +, Kai Ma (kaima) wrote:
> I want to implement a HW ASIC switch by porting OVS.
> For ³add flow², I want to use HW ASCI TCAM to set ACL rule for flow, not
> linux CPU.
>
> How can I porting this part?
Did you read PORTING.md?
> In OVS-ofctl.c line616, tran
od to
> > return "ODP_FIT_TOO_LITTLE" due to "expected but not present" attribute,
> > which makes perfect sense. However, this was not taken any further and the
> > flow entry is still cached at the kernel.
> >
> > Is there any additional logic
itch.
>
> Evgeny
>
> On 26 ? 2016, at 18:17, Ben Pfaff mailto:b...@ovn.org>>
> wrote:
>
> On Tue, Apr 26, 2016 at 02:55:51PM +, Evgeny Sandler wrote:
> I'm using OVS-2.1.3 and I'm trying to push additional .1q VLAN tag to
> Ethernet frame wh
riginal Message-
> From: discuss [mailto:discuss-boun...@openvswitch.org] On Behalf Of Ben Pfaff
> Sent: יום ה 21 אפריל 2016 19:54
> To: Yi Ba
> Cc: b...@openvswitch.org
> Subject: Re: [ovs-discuss] ovs get stuck when running traffic from VM to VM
> on same compute
>
&g
On Tue, Apr 26, 2016 at 02:55:51PM +, Evgeny Sandler wrote:
> I'm using OVS-2.1.3 and I'm trying to push additional .1q VLAN tag to
> Ethernet frame which already has one VLAN tag by using ovs-ofctl
> command as follows:
OVS doesn't support QinQ.
___
On Tue, Apr 26, 2016 at 01:46:26PM +0900, Borasy Kong wrote:
> Can someone help show me how the flow stats is generated. For example,
> where does the ovs-ofctl dump-flows get the data like n_packet, n_bytes
> from each time this command is issued. Is there any particular spot in the
> OVS source c
On Tue, Apr 26, 2016 at 07:09:39PM +0800, 307419...@qq.com wrote:
> Apr 26 03:51:17 ppmaster ovsdb-server[711]:
> ovs|2|stream_unix|ERR|/var/run/openvswitch/db.sock: binding failed: No
> such file or directory
/var/run/openvswitch doesn't exist. You'll have to create it.
(The scripts that
On Tue, Apr 26, 2016 at 02:10:20PM +0200, Petr Horacek wrote:
> is there any difference between setting an IP on OVS bridge and setting in
> on its internal iface?
>
> ovs-vsctl add-br br0
> ip addr add 192.168.1.10 dev br0
>
> vs.
>
> ovs-vsctl add-br br0 -- add-port br0 mgmt0 -- set Interface
On Tue, Apr 26, 2016 at 01:56:58PM +0200, Joel Frank wrote:
> I got some questions regarding ovs and flow entry monitoring. I found
> several patches back in Jun 2014 implementing Flow Monitors
> (http://openvswitch.org/pipermail/dev/2014-June/041673.html). However I
> can't get them to work with c
On Sat, Apr 23, 2016 at 11:45:12AM -0700, Ben Warren wrote:
> Thanks for the explanations!
> > On Apr 23, 2016, at 10:15 AM, Ben Pfaff wrote:
> >
> > On Fri, Apr 22, 2016 at 11:25:26PM -0700, Ben Warren wrote:
> >> As you can see, the disconnects happe
On Fri, Apr 22, 2016 at 11:25:26PM -0700, Ben Warren wrote:
> As you can see, the disconnects happen when the datapath ID changes,
> presumably due to adding interfaces? Only the addition of the first
> two interfaces seem to bring it down.
OVS disconnects OpenFlow connections when the datapath I
On Fri, Apr 22, 2016 at 03:55:07PM +0800, Haosen Wen wrote:
> Hi all,
>
> I'm trying to add a new action, "FLAT" to ovs-2.3.2 action definitions.
> When compiling, I met the following errors:
>
> *lib/ofp-actions.c: In function 'ofpact_from_openflow11':*
> *lib/ofp-actions.c:1155:5: error: duplic
On Fri, Apr 22, 2016 at 04:52:29PM +0800, lol wrote:
> Hi, any one can tell me the arctitecture of flow table in ovs?
https://www.usenix.org/conference/nsdi15/technical-sessions/presentation/pfaff
___
discuss mailing list
discuss@openvswitch.org
http://o
On Thu, Apr 21, 2016 at 02:55:46PM +, Peter Gubka -X (pgubka - PANTHEON
TECHNOLOGIES at Cisco) wrote:
> Is something wrong with my approach or is it a bug in openvswitch?
> Is role parameter reliable?
Please enable OpenFlow debug logging with "-vvconn" or "ovs-appctl
vlog/set vconn", then rep
On Thu, Apr 21, 2016 at 01:24:35PM +, Yi Ba wrote:
>
>- What you did that make the problem appear.
> launching 2 VMs on the same host. The host has ovs-dpdk and the VM ports are
> vhostuser. Generating bi-directional traffic (each VM sends to the other)
> using dpdk from within t
On Wed, Apr 20, 2016 at 03:04:52PM +0800, zhao.j...@zte.com.cn wrote:
> I think VLAN splinters are needed for the buggy network drivers, and i
> find ixgbe
> is one among them. suppose network driver is buggy, and VLAN splinters are
> used
> on the nics, i think ovs should not deals with those
On Wed, Apr 20, 2016 at 05:25:23PM +0800, Troy wrote:
> Hi, Dear all:
> I have installed my OVS in Ubuntu-14.04. The OVS version is 2.5.0.
> I installed two VMs in the system, and the ifconfig list contains 'vnet0'
> and 'vnet1'.
> When I add a port named 'vnet1' to br0, it shows:
> ovs-vsctl:
On Wed, Apr 20, 2016 at 09:19:16AM +0800, zhao.j...@zte.com.cn wrote:
> I'm using Linux 3.10.0-229 with intel 82599 10 gigabit nics, and the
> network driver is ixgbe, version 3.21.2
Linux 3.3 and later should not need VLAN splinters.
___
discuss ma
On Tue, Apr 19, 2016 at 08:42:28PM +0800, zhao.j...@zte.com.cn wrote:
> Hi,dear all:
> I'm using ovs2.5.0 in centos7.2, for the reason of drivers, i has no
> choice but to use function
> of vlan splinters.
You are using Linux 3.2 or earlier?
___
di
On Tue, Apr 19, 2016 at 04:27:19PM +, Farhad Sunavala wrote:
> >> Proposal:
> >> We introduce the interface keyword "learn= no".
> >> E.g. ovs-vsctl set interface foo learn=no
> >>
> >> This will instruct br-int NOT to do MAC learning on packets received on
> >> interface foo.
>
> >This seem
Please don't drop the mailing list.
On Mon, Apr 18, 2016 at 11:24:22PM -0400, Rüdiger Birkner wrote:
> I see. This requires me to use OFv1.5, right?
Yes.
We'd take patches, if you want to make it available in earlier versions.
> Is there a way to configure ovs to use the five-tuple and then
>
On Tue, Apr 19, 2016 at 01:06:25PM +, Jan Wickbom wrote:
>
>
> > -Original Message-
> > From: Ben Pfaff [mailto:b...@ovn.org]
> > Sent: den 18 april 2016 17:37
> > To: Jan Wickbom
> > Cc: discuss@openvswitch.org
> > Subject: Re: [ovs-discuss
On Tue, Apr 19, 2016 at 08:08:46AM +, Farhad Sunavala wrote:
> Proposal:
> We introduce the interface keyword "learn= no".
> E.g. ovs-vsctl set interface foo learn=no
>
> This will instruct br-int NOT to do MAC learning on packets received on
> interface foo.
This seems less flexible than add
On Mon, Apr 18, 2016 at 10:42:21PM +, Tarren, Jacob A. (LARC-B703)[LITES
II] wrote:
> Hello!
>
> I've just started experimenting with OpenVSwitch on Xen, and so far I like
> it, but I'm running into an issue setting up port mirroring and promiscuous
> mode.
>
> I'm trying to get all traffi
On Mon, Apr 18, 2016 at 04:57:16PM -0400, Rüdiger Birkner wrote:
> I am planning to use the OF 1.3 select group feature. It is important for
> my application that the buckets are assigned based on the five tuple (or at
> least the mac and ip addresses). Is the selection of the bucket still based
>
301 - 400 of 3929 matches
Mail list logo