Re: [ovs-discuss] DPDK bandwidth issue

Geza, Thanks for your insight. - What is the packet size you see these bandwidth values? A: I've tried various packet sizes with iperf, no significant differences. - What endpoints do you use for traffic generation? A: The bandwidth in question was measured from host to host, no VMs involved.

Re: [ovs-discuss] Theory of OVS DPDK forward packets in VXLAN

Regards _Sugesh From: discuss [mailto:discuss-boun...@openvswitch.org] On Behalf Of zhi Sent: Wednesday, October 19, 2016 4:57 AM To: discuss@openvswitch.org Subject: [ovs-discuss] Theory of OVS DPDK forward packets in VXLAN Hi all I configured native tunneling by this guide[1]. And follow

Re: [ovs-discuss] OVS - Userspace packet processing

Regards _Sugesh From: Advith Nagappa [mailto:advith.naga...@gmail.com] Sent: Wednesday, October 19, 2016 1:02 AM To: Chandran, Sugesh Cc: discuss@openvswitch.org Subject: Re: [ovs-discuss] OVS - Userspace packet processing Hello Sugesh, Thank you for the response.

Re: [ovs-discuss] DPDK bandwidth issue

Regards _Sugesh From: discuss [mailto:discuss-boun...@openvswitch.org] On Behalf Of Zhang Qiang Sent: Wednesday, October 19, 2016 7:55 AM To: geza.ge...@gmail.com Cc: discuss Subject: Re: [ovs-discuss] DPDK bandwidth issue Geza, Thanks for your insight. - What is the

Re: [ovs-discuss] OVS with DPDK for Vxlan

Regards _Sugesh From: discuss [mailto:discuss-boun...@openvswitch.org] On Behalf Of Rohan Gandhi Sent: Tuesday, October 18, 2016 7:44 PM To: discuss@openvswitch.org Subject: [ovs-discuss] OVS with DPDK for Vxlan Hello All, I am currently using OVS for Vxlan tunneling, and it works fine. I am

Re: [ovs-discuss] OVS - Userspace packet processing

Regards _Sugesh From: Advith Nagappa [mailto:advith.naga...@gmail.com] Sent: Wednesday, October 19, 2016 9:15 AM To: Chandran, Sugesh Cc: discuss@openvswitch.org Subject: Re: [ovs-discuss] OVS - Userspace packet processing Thank you again, I am using a mininet set

Re: [ovs-discuss] OVS - Userspace packet processing

Regards _Sugesh From: Advith Nagappa [mailto:advith.naga...@gmail.com] Sent: Wednesday, October 19, 2016 11:11 AM To: Chandran, Sugesh Cc: discuss@openvswitch.org Subject: Re: [ovs-discuss] OVS - Userspace packet processing Thank you again. [Sugesh] I guess this

Re: [ovs-discuss] Ｑｕｅｓｔｉｏｎ　ａｂｏｕｔ　to set the miss_send_len

I mean that the OpenFlow 1.0 message is the way to change miss_send_len. If you need to change it, use that message. If your controller does not support that message, add support. It is a very simple message so any controller should be able to easily support it. On Thu, Oct 20, 2016 at

Re: [ovs-discuss] DPDK bandwidth issue

Thanks Sugesh, But would you please help me further with why dpdkbond affects bandwidth? With the bond the bandwidth is only around 30Mbps, configurations are shown in previous post. > On 19 Oct 2016, at 3:46 PM, Chandran, Sugesh > wrote: > > > > Regards >

Re: [ovs-discuss] Ｑｕｅｓｔｉｏｎ　ａｂｏｕｔ　to set the miss_send_len

Thank you for your reply. Do you mean that I can't change the value of miss_send_len? Can I change it by the other way? For example, by rewriting config file or changing code and rebuilding. -作成者: Ben Pfaff - 宛先: Naoyuki NS Shimizu/R/RICOH@RICOH 送信元: Ben Pfaff

Re: [ovs-discuss] [ovs-dev] packets getting dropped in vhostuser port

There seem to be some version dependency between the two. dpdk-16.04 with ovs2.6 is not possible as ovs is referencing some newer dpdk code. i tried the reverse, dpdk16.07 and ovs2.5.9, but getting the following error (rpm based build) + cd dpdk + make config T=x86_64-atm-linuxapp-gcc make[1]:

Re: [ovs-discuss] Ｑｕｅｓｔｉｏｎ　ａｂｏｕｔ　to set the miss_send_len

On Wed, Oct 19, 2016 at 12:30:54AM +, Naoyuki NS Shimizu wrote: > I'd like to change the value of miss_send_len. > > I use trema as Openflow v1.0 controller and trema ｉｓ not implemented > SetConfig messege. > > So I can't send SetConfig messege from Openflow v1.0 controller. > > Please

[ovs-discuss] [ovs-dev] build failure with DPDK on Ubuntu 16.04 with GCC 5.4

Hi, I met an issue when building OVS with DPDK on Ubuntu 16.04, the commit ID is as below, also I'm using DPDK R16.11 RC1 commit ID: commit fed622dfd90ffc754808ecb464d7e62e77f0da58. We can pass build on Fedora 23 but can't pass build on Ubuntu 16.04, kernel: 4.4, gcc: 5.4. Could anyone help

Re: [ovs-discuss] [ovs-dev] packets getting dropped in vhostuser port

One suggestion, could you keep dpdk as dpdk-16.04, only ovs upgrade, then see if the case still happen. Or you can keep only dpdk changed but no OVS change. We need to narrow down the issue. From my knowledge, qemu-2.4.1 is fine for the case. -Original Message- From: dev

Re: [ovs-discuss] OVS - Userspace packet processing

Thank you again. *[Sugesh] I guess this must be something to do with how mininet configure OVS. Please have a look at the ovs-vswitchd.log file for more details(this shows how OVS is being configured by mininet).* *Also check if ovs-vswitchd and ovsdb-servers are running on the machine*

Re: [ovs-discuss] OVS - Userspace packet processing

Thank you again, I am using a mininet set up (mn --switch user), which I think uses the userspace ovs switch. (Is this assumption right? All available documentation uses wording which gives this impression.) To confirm this, I also have removed the openvswitch.ko in my VM. But curiously

[ovs-discuss] adding gtp flow

Dear Members, I have applied the gtp-u patch for OVS version 2.5 and I added the following flows and my topology is sudo mn --mac. H1 connected to port 1 and H2 connected to port 2. Port 3 is a gtp tunnel port. - sudo ovs-vsctl add-port s1 gtp0 -- set interface gtp0 ofport_request=3

[ovs-discuss] packets getting dropped in vhostuser port

Hi, In a PHY-VM-PHY setup, i was using dpdk16.04 + OVS 2.5.90 + qemu 2.4.1, everything was working fine. When i upgraded to dpdk 16.07 + ovs 2.6 (branch-2.6) without any change to the setup/configuration; now all the packets(1200B) are getting dropped at vhostuser ports. Though, i am able to send

Re: [ovs-discuss] Time a packet spends in the switch

I've tried in datapath.c, bridge.c, learning-switch.c and flow-table.c but nothing works. For kernel module i'm using printk() to print the messages and for userspace simply writing to a file but for some reason nothing happens in either modules. Any other ideas that i can try to get the

Re: [ovs-discuss] [vxlan]openvswitch-2.5.0 coredump issue

On Tue, Oct 18, 2016 at 5:53 AM, Huanglili (lee) wrote: > Hi, all: > > I have met a core dump issue when I ifconfig down vxlan-port > durning sending vxlan packets. > > We know that when ifconfig down vxlan port, the udp socket will be > freed, but

Re: [ovs-discuss] Time a packet spends in the switch

i'm using vmware for the testing. So i have a state saved before compiling the source code for virtual machine. Each after doing all the modification, i compile, install and load the module. send some traffic between couple of hosts then see all the log like kernel, dmesg, syslog etc, if nothing

Re: [ovs-discuss] Time a packet spends in the switch

Hi Danish, Please post four changes. I guess all packets receives via netdev_frame_hook() function in openvswith.ko module. Sateesh On Wed, Oct 19, 2016 at 10:17 AM, Joe Stringer wrote: > Are you sure that you loaded your newly compiled kernel module? > (insmod