观察网络性能时如何选择工具
转载自：http://linuxperf.com/?cat=4
Linux系统上的网络工具甚多，如何根据实际需要选择称手的工具呢？在此作一个简单介绍：
观察网络流量：
- “sar -n DEV 1 5” 可以统计每个网卡上的网络流速：
|
1
2
3
4
5
6
7
|
# sar -n DEV 2 5
...
02:47:12 PM IFACE rxpck/s txpck/s rxkB/s txkB/s rxcmp/s txcmp/s rxmcst/s
02:47:14 PM lo 0.00 0.00 0.00 0.00 0.00 0.00 0.00
02:47:14 PM eth0 2.51 0.00 0.17 0.00 0.00 0.00 0.00
...
|
- iptraf 是观察网络流速的强力工具，它可以让你的观察逐步深入，从硬件层（网卡），到网络层（IPv4，IPv6），到传输层（TCP，UDP etc.），一直到每一对socket pair。
观察网络连接的状态：
- “netstat -a“
这是传统的工具，但是它无力处理海量的网络连接。所以在大规模网络连接的主机上，建议使用ss。
|
1
2
3
4
5
6
7
|
# netstat -a
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 0 0 *:mysql *:* LISTEN
tcp 0 0 *:5901 *:* LISTEN
...
tcp 0 104 bj71s060.chn.hp.com:ssh 16.169.16.67:52681 ESTABLISHED
|
- “ss -a” 列出所有的网络连接。ss特别适合海量连接的主机。
如果加上”-p”选项，还可以显示对应的进程号。
|
1
2
3
4
5
6
7
|
# ss -a
State Recv-Q Send-Q Local Address:Port Peer Address:Port
LISTEN 0 50 *:mysql *:*
LISTEN 0 5 *:5901 *:*
...
LISTEN 0 128 :::38246 :::*
ESTAB 0 184 16.187.252.58:ssh 16.169.16.67:52681
|
观察静态统计值：
- “netstat -i“
|
1
2
3
4
5
|
# netstat -i
Kernel Interface table
Iface MTU Met RX-OK RX-ERR RX-DRP RX-OVR TX-OK TX-ERR TX-DRP TX-OVR Flg
eth0 1500 0 572421962 0 0 0 16831237 0 0 0 BMRU
lo 16436 0 556534 0 0 0 556534 0 0 0 LRU
|
- ifconfig 可以看到网卡层面的少量统计值，packet数量，collision，errors等：
|
1
2
3
4
5
6
7
8
9
10
|
# ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:19:BB:5B:07:34
inet addr:16.187.252.58 Bcast:16.187.255.255 Mask:255.255.252.0
inet6 addr: fe80::219:bbff:fe5b:734/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:572422657 errors:0 dropped:0 overruns:0 frame:0
TX packets:16831252 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:50512945748 (47.0 GiB) TX bytes:11194760435 (10.4 GiB)
Interrupt:19 Memory:f0500000-f0520000
|
- “ip -s link” 看到的信息基本类似：
|
1
2
3
4
5
6
7
8
9
10
11
12
13
|
# ip -s link
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
RX: bytes packets errors dropped overrun mcast
2255610398 556534 0 0 0 0
TX: bytes packets errors dropped carrier collsns
2255610398 556534 0 0 0 0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:19:bb:5b:07:34 brd ff:ff:ff:ff:ff:ff
RX: bytes packets errors dropped overrun mcast
3268332087 572422979 0 0 0 16733497
TX: bytes packets errors dropped carrier collsns
2604833405 16831285 0 0 0 0
|
- “netstat -s” 提供了各个协议下的统计信息，有些统计值比如retransmit是很有用的，只有 “netstat -s” 能看到：
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
|
# netstat -s
...
Tcp:
71479 active connections openings
7181 passive connection openings
13723 failed connection attempts
407 connection resets received
1 connections established
18969163 segments received
11210435 segments send out
85883 segments retransmited
0 bad segments received.
80162 resets sent
...
475 fast retransmits
30 forward retransmits
110 retransmits in slow start
22772 other TCP timeouts
5 sack retransmits failed
...
|