As mentioned previously, that is not supported for connects currently.
Regards,
Florin
> On Mar 31, 2022, at 9:44 PM, weizhen9...@163.com wrote:
>
> Hi,
> I describe our scene in detail. We use nginx in vpp host stack as a proxy.
> And we add some features in nginx. For example, nginx close
Hi,
I describe our scene in detail. We use nginx in vpp host stack as a proxy. And
we add some features in nginx. For example, nginx close upstream tcp links
actively and this causes a lot of TIME_WAIT states in nginx proxy when we test
the performance of nginx using vpp host stack. So we
Hi,
Given that 20k connections are being actively opened, those on main thread, and
40k are established, those on the workers, suggests that tcp runs out of ports
for connects. If possible either increase the number of destination IPs for
nginx or try “tcp src-address ip1-ip2” and pass in a
Hi,
>From this test results, can you see some errors? Why is the performance of
>nginx using vpp low?
Thanks.
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21174): https://lists.fd.io/g/vpp-dev/message/21174
Mute This Topic:
TCP accepts connections in time-wait (see here [1]) but won’t reuse ports of
connection in time-wait for connects. If you expect lots of active opens from
nginx to only one destination ip and you have multiple source ips, you could
try to use "tcp src-address” cli.
Regards,
Florin
[1]
Hi VPP fellows:
I got problems when I tried to create and delete AF_XDP interface. Pls give me
some suggestions to figure out the problem. The issue description is as below:
Step 1: create af_xdp interface based on host if wlan0: create interface af_xdp
host-if wlan0 num-rx-queues 1
Step 2:
virtio device is using Linux Kernel driver “virtio-pci”. You need to rebind it
to ‘vfio-pci’ be able to use it in user-space.
There is a rebinding script in vpp:
https://git.fd.io/vpp/tree/extras/scripts/pci-nic-bind
-br
mohsin
From: on behalf of "long...@gmail.com"
Date: Thursday, March
[Edited Message Follows]
Thank Mohsin,
I create interface without DPDK like below:
>
> create int virtio :00:08.0 gso-enabl
>
but I got *error* :
*create interface virtio: device not bound to 'vfio-pci' or 'uio_pci_generic'
kernel module
* I tried modprobe virtio_pci but still got it.
Thank Mohsin,
I create interface without DPDK like below:
>
> create int virtio :00:08.0 gso-enabl
>
but I got *error* :
*create interface virtio: device not bound to 'vfio-pci' or 'uio_pci_generic'
kernel module
* I tried modprobe virtio_pci but still got it.
Do you know how to fix that?
I didn’t mean you should switch to envoy, just that throughput is pretty low
probably because of some configuration. What that configuration is is not
obvious unfortunately.
Regarding the kernel parameters, we have time wait reuse enabled (equivalent to
tcp_tw_reuse) but that should not
Forgot to mention, your config have enabled gso on tap interface while is not
enabled on dpdk interface.
Either enable it on dpdk interface if the underneath vhost device supports it.
or, disable it from tap interface too.
In case of vpp native virtio:
create int virtio :00:08.0
Please add following to dpdk stanza in startup.conf
“DPDK { dev :00:08.0 { tso on } enable-tcp-udp-checksum }”
Or else try VPP native virtio without using dpdk:
create int virtio :00:08.0
From: on behalf of "long...@gmail.com"
Date: Thursday, March 31, 2022 at 4:40 PM
To:
Hi Mohsin,
Please see below:
>
>
>
> Can you please share the numbers you are getting?
>
>
*Without VPP, I get 5Gbps. With VPP I get ~2Gbps.*
>
>
>
> And also please share the output of:
>
>
>
> sh hardware
>
>
>
> sh tap
>
>
>
> sh threads
>
>
>
> sh cpu
>
>
Please see
Hi Jerome,
>
> you enabled gso but didn't turn gro on. IOW, in this mode, VPP accepts GSO
> packets sent by linux but won't coalesce packets to reduce the load on
> linux side.
> You may be interested in this article
>
Hi,
The version of vpp is 22.06. The configs are in the attachment.
The nginx using vpp host stack is low.
The most important test indicators is RPS.
How can we increase the performance of nginx using vpp host stack?
nginx.conf
Description: Binary data
startup.conf
Description: Binary data
Hi,
Could you provide a bit more info about the numbers you are seeing and the type
of test you are doing? Also some details about your configs and vpp version?
As for tcp_tw_recycle, I believe that was removed in Linux 4.12 because it was
causing issues for nat-ed connections. Did you mean
Now, we develop some functions in nginx and test nginx performance in kernel
host stack and vpp host stack. We found the performance of nginx in vpp is
lower than nginx in kernel.
So I want to ask how to debug to increase the performance of nginx in vpp.
Besides, we modify some parameter in
Hello,
Can you please share the numbers you are getting?
And also please share the output of:
sh hardware
sh tap
sh threads
sh cpu
-br
Mohsin
From: on behalf of "long...@gmail.com"
Date: Thursday, March 31, 2022 at 10:20 AM
To: "vpp-dev@lists.fd.io"
Subject: Re: [vpp-dev] Fastest way to
Hello,
In your config you enabled gso but didn't turn gro on. IOW, in this mode, VPP
accepts GSO packets sent by linux but won't coalesce packets to reduce the load
on linux side.
You may be interested in this article
Thank you Ole,
>
> NIC - VPP - BSD socket app (using LD_PRELOAD
> https://s3-docs.fd.io/vpp/22.02/developer/extras/vcl_ldpreload.html
> )
==> Can any my application (Java, Python,...) run on vlc ld preload like NIC -
VPP - vcl ld_preload socket - application?
>
> But the answer here is very
Hoi Suresh,
As a datapoint to this - route synchronization from VPP -> Linux is not
implemented. The other way around (setting routes in Linux, using Netlink,
will be propagated into the VPP dataplane).
On Thu, Mar 31, 2022 at 5:29 AM suresh vuppala wrote:
> I have two loopback interfaces in
Dear Mr Longtrb,
> I am a newbie in VPP. Many thanks to you all for making great Vpp.
> I have many applications in user space (written by java, C, python,...) and I
> want to connect those apps to VPP so what is the fastest way that VPP can
> support without changing the application?
>
> I
I put an assert just to be sure that virtio-input got a frame with the
ETH_INPUT_FRAME_F_SINGLE_SW_IF_IDX
flag.
Attaching to process 324596
[New LWP 324612]
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
0x7fc5c21cf1dc
23 matches
Mail list logo