I think The ARP abnormal is caused by the entire FIB of VPP can not process any
packets;
During these days, I fount when I use tcpdump on both side interface when VPP
NAT multithreads mode,
50% out2in reply packets are lost, the lost packets are handoff to
vpp_wk_1(thread_2);
the “show run “ results show that these have on active “nat44-out2in “;
Is this the root cause?
DBGvpp# show run
Thread 0 vpp_main (lcore 0)
Time 403.2, average vectors/node 0.00, last 128 main loops 0.00 per node 0.00
vector rates in 0.0000e0, out 0.0000e0, drop 0.0000e0, punt 0.0000e0
Name State Calls Vectors
Suspends Clocks Vectors/Call
acl-plugin-fa-cleaner-process event wait 0 0
1 3.61e4 0.00
admin-up-down-process event wait 0 0
1 5.37e4 0.00
api-rx-from-ring any wait 0 0
24 7.56e5 0.00
bfd-process event wait 0 0
1 2.92e4 0.00
cdp-process any wait 0 0
55 5.25e5 0.00
dhcp-client-process any wait 0 0
5 3.69e4 0.00
dns-resolver-process any wait 0 0
1 1.34e4 0.00
dpdk-ipsec-process done 1 0
0 2.13e5 0.00
dpdk-process any wait 0 0
133 3.18e6 0.00
fib-walk any wait 0 0
199 1.02e4 0.00
flow-report-process any wait 0 0
1 1.34e4 0.00
flowprobe-timer-process any wait 0 0
1 2.40e4 0.00
ikev2-manager-process any wait 0 0
395 7.59e3 0.00
ioam-export-process any wait 0 0
1 1.99e4 0.00
ip-route-resolver-process any wait 0 0
5 2.75e4 0.00
ip6-icmp-neighbor-discovery-ev any wait 0 0
395 7.52e3 0.00
l2fib-mac-age-scanner-process event wait 0 0
1 9.48e3 0.00
lisp-retry-service any wait 0 0
199 1.47e4 0.00
lldp-process event wait 0 0
1 2.12e7 0.00
memif-process event wait 0 0
1 3.80e4 0.00
nat-det-expire-walk done 1 0
0 1.23e4 0.00
nat64-expire-walk any wait 0 0
41 3.54e4 0.00
nat64-expire-worker-walk interrupt wa 40 0
0 1.47e4 0.00
send-garp-na-process event wait 0 0
1 4.37e3 0.00
startup-config-process done 1 0
1 1.41e9 0.00
udp-ping-process any wait 0 0
1 2.80e4 0.00
unix-cli-local:0 active 0 0
48 2.12e7 0.00
unix-epoll-input polling 170368 0
0 5.27e6 0.00
vhost-user-process any wait 0 0
1 1.42e4 0.00
vhost-user-send-interrupt-proc any wait 0 0
1 1.18e4 0.00
vpe-link-state-process event wait 0 0
4 1.08e4 0.00
vpe-oam-process any wait 0 0
194 1.03e4 0.00
vxlan-gpe-ioam-export-process any wait 0 0
1 2.21e4 0.00
wildcard-ip4-arp-publisher-pro event wait 0 0
1 1.45e4 0.00
---------------
Thread 1 vpp_wk_0 (lcore 1)
Time 403.2, average vectors/node 1.71, last 128 main loops 0.00 per node 0.00
vector rates in 1.4584e0, out 3.2244e-1, drop 1.1484e0, punt 0.0000e0
Name State Calls Vectors
Suspends Clocks Vectors/Call
TenGigabitEthernet83/0/0-outpu active 12 64
0 9.52e2 5.33
TenGigabitEthernet83/0/0-tx active 12 64
0 1.25e3 5.33
TenGigabitEthernet83/0/1-outpu active 8 66
0 1.63e3 8.25
TenGigabitEthernet83/0/1-tx active 8 66
0 2.38e3 8.25
arp-input active 2 2
0 7.03e4 1.00
dpdk-input polling 509908889 588
0 8.09e8 0.00
error-drop active 427 463
0 5.21e3 1.08
ethernet-input active 425 431
0 8.32e3 1.01
ip4-glean active 2 32
0 5.23e4 16.00
ip4-input-no-checksum active 15 157
0 1.08e4 10.47
ip4-load-balance active 12 64
0 1.02e3 5.33
ip4-lookup active 19 160
0 2.44e3 8.42
ip4-rewrite active 17 128
0 1.64e3 7.53
llc-input active 394 396
0 2.89e3 1.01
lldp-input active 30 33
0 1.17e4 1.10
nat44-in2out active 9 96
0 5.42e3 10.67
nat44-in2out-slowpath active 8 96
0 3.86e4 12.00
nat44-in2out-worker-handoff active 5 94
0 6.04e4 18.80
nat44-out2in active 12 64
0 3.63e3 5.33
nat44-out2in-worker-handoff active 11 63
0 5.01e4 5.73
nat64-expire-worker-walk interrupt wa 40 0
0 9.92e3 0.00
---------------
Thread 2 vpp_wk_1 (lcore 7)
Time 403.2, average vectors/node 12.79, last 128 main loops 0.00 per node 0.00
vector rates in 4.0429e-1, out 1.5874e-1, drop 7.9370e-2, punt 0.0000e0
Name State Calls Vectors
Suspends Clocks Vectors/Call
TenGigabitEthernet83/0/1-outpu active 6 64
0 7.76e2 10.67
TenGigabitEthernet83/0/1-tx active 6 64
0 9.85e2 10.67
dpdk-input polling 504121313 163
0 2.93e9 0.00
error-drop active 2 32
0 3.53e3 16.00
ip4-glean active 2 32
0 7.68e2 16.00
ip4-input-no-checksum active 11 163
0 1.21e4 14.82
ip4-lookup active 8 96
0 1.54e3 12.00
ip4-rewrite active 6 64
0 1.43e3 10.67
nat44-in2out active 9 96
0 1.44e4 10.67
nat44-in2out-slowpath active 8 96
0 5.40e4 12.00
nat44-in2out-worker-handoff active 4 98
0 5.61e4 24.50
nat44-out2in-worker-handoff active 7 65
0 4.74e4 9.29
nat64-expire-worker-walk interrupt wa 40 0
0 1.08e4 0.00
发件人: "Matus Fabian -X (matfabia - PANTHEON TECHNOLOGIES at Cisco)"
<matfa...@cisco.com>
日期: 2018年1月5日 星期五 下午3:23
收件人: 李洪亮 <lihongli...@360.cn>
抄送: "vpp-dev@lists.fd.io" <vpp-dev@lists.fd.io>
主题: RE: The performance problem of NAT plugin
I tested ARP with stable/1801 and it works fine
23:20:10.885539 ARP, Request who-has 3.3.3.5 tell 3.3.3.1, length 28
23:20:10.885769 ARP, Reply 3.3.3.5 is-at 08:00:27:c9:ea:36 (oui Unknown),
length 46
Matus
From: 李洪亮 [mailto:lihongli...@360.cn]
Sent: Friday, December 22, 2017 4:36 PM
To: Matus Fabian -X (matfabia - PANTHEON TECHNOLOGIES at Cisco)
<matfa...@cisco.com>
Cc: vpp-dev@lists.fd.io
Subject: Re: The performance problem of NAT plugin
I find that when I use NAT , the ARP is abnormal;
AS, I use the 3.3.3.5 as the NAT pool address: nat44 add address 3.3.3.5
On target NIC(3.3.3.1), the tcpdump can capture the ARP request without reply:
23:21:57.783994 ARP, Request who-has 3.3.3.5 tell 3.3.3.1, length 28
_______________________________________________
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
