DPDK usage discussions
 help / color / mirror / Atom feed
From: 刘峰麟 <liuflin@bupt.edu.cn>
To: users <users@dpdk.org>
Subject: [bug] iperf3 generating udp flows make ovs-dpdk's port can't receive the packets.
Date: Wed, 20 Apr 2022 23:22:15 +0800	[thread overview]
Message-ID: <tencent_0C7BA82A135582F21DD168AD@qq.com> (raw)

[-- Attachment #1: Type: text/plain, Size: 3040 bytes --]

Hi,I encount a bug when I use iperf3 to generate udp flows, which would makes ovs-dpdk port can't receive packets,while tcp flows seem wouldn't. There is details below.
The test topology is 2 VMs installed iperf3 connecting to 1 VM installed ovs-dpdk,which NIC is e1000. 
My iperf3 comand is:
iperf3 -s -i 1
iperf3 10.1.3.4 -u -b 1G -w6m -t 36000&nbsp;


Ovs's flow-table is
&nbsp;cookie=0x0, duration=2319.958s, table=0, n_packets=74802603, n_bytes=104773394087, in_port=dpdk2 actions=output:dpdk1
&nbsp;cookie=0x0, duration=2319.958s, table=0, n_packets=89, n_bytes=5572, in_port=dpdk1 actions=output:dpdk2



But after for about 10s minutes or 1h longer, the iperf3 server can't receive any packets unless restart ovs.Then I stop iperf3, let hosts ping each other, gdb ovs-dpdk and found that rte_eth_rx_burst()&nbsp; always return 0 at one of the ports, meanning it can't receive any packets. 


iperf3 message is like below:
[&nbsp; 5] 2727.00-2728.00 sec&nbsp; 24.2 MBytes&nbsp; &nbsp;203 Mbits/sec&nbsp; 0.042 ms&nbsp; 0/3098 (0%)
[&nbsp; 5] 2728.00-2729.00 sec&nbsp; 26.5 MBytes&nbsp; &nbsp;222 Mbits/sec&nbsp; 0.100 ms&nbsp; 0/3389 (0%)
[&nbsp; 5] 2729.00-2730.00 sec&nbsp; 27.5 MBytes&nbsp; &nbsp;231 Mbits/sec&nbsp; 0.079 ms&nbsp; 0/3522 (0%)
[&nbsp; 5] 2730.00-2731.00 sec&nbsp; 25.1 MBytes&nbsp; &nbsp;210 Mbits/sec&nbsp; 0.213 ms&nbsp; 9/3219 (0.28%)
[&nbsp; 5] 2731.00-2732.00 sec&nbsp; 11.2 MBytes&nbsp; 94.1 Mbits/sec&nbsp; 0.042 ms&nbsp; 0/1436 (0%)
[&nbsp; 5] 2732.00-2733.00 sec&nbsp; 0.00 Bytes&nbsp; 0.00 bits/sec&nbsp; 0.042 ms&nbsp; 0/0 (0%)
[&nbsp; 5] 2733.00-2734.00 sec&nbsp; 0.00 Bytes&nbsp; 0.00 bits/sec&nbsp; 0.042 ms&nbsp; 0/0 (0%)
[&nbsp; 5] 2734.00-2735.00 sec&nbsp; 0.00 Bytes&nbsp; 0.00 bits/sec&nbsp; 0.042 ms&nbsp; 0/0 (0%)
[&nbsp; 5] 2735.00-2736.00 sec&nbsp; 0.00 Bytes&nbsp; 0.00 bits/sec&nbsp; 0.042 ms&nbsp; 0/0 (0%)
[&nbsp; 5] 2736.00-2737.00 sec&nbsp; 0.00 Bytes&nbsp; 0.00 bits/sec&nbsp; 0.042 ms&nbsp; 0/0 (0%)





ovs_log message is:


2022-04-20T14:40:10.483Z|00109|bridge|INFO|ovs-vswitchd (Open vSwitch) 2.13.0
2022-04-20T14:40:10.685Z|00110|connmgr|INFO|br0<-&gt;unix#6: 3 flow_mods in the last 0 s (2 adds, 1 deletes)
2022-04-20T14:40:10.728Z|00111|netdev_dpdk|WARN|Failed to enable flow control on device 0
2022-04-20T14:40:10.728Z|00112|netdev_dpdk|WARN|Failed to enable flow control on device 1
2022-04-20T14:40:15.150Z|00001|netdev_dpdk(dpdk_watchdog1)|DBG|Port 0 Link Up - speed 1000 Mbps - full-duplex
2022-04-20T14:40:15.151Z|00002|netdev_dpdk(dpdk_watchdog1)|DBG|Port 1 Link Up - speed 1000 Mbps - full-duplex
2022-04-20T14:40:18.662Z|00113|memory|INFO|228596 kB peak resident set size after 10.5 seconds
2022-04-20T14:40:18.662Z|00114|memory|INFO|handlers:2 ports:3 revalidators:2 rules:6 udpif keys:6





Could you please tell me how to solve this issue? I would appreciate it!!!











刘峰麟

北京邮电大学/研究生/计算机学院(国家示范性软件学院)



北京




&nbsp;

[-- Attachment #2: Type: text/html, Size: 4470 bytes --]

                 reply	other threads:[~2022-04-26  8:34 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=tencent_0C7BA82A135582F21DD168AD@qq.com \
    --to=liuflin@bupt.edu.cn \
    --cc=users@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).