透明防火牆架設的完全攻略（bridge＋iptables＋squid）
架設透明代理和防火牆是linux平台上很熱的話題,在水木上也有相關文章,但完全的攻略在公網上也很少,最近架了一台,前後花去一個多星期（我這人手腳慢,別笑,中間笑話也頗多）覺得還是把過程寫下來的好,可以讓人依葫蘆畫瓢.
先把網路環境說一下,一個200人左右的區域網,一個C class,一台路由器做NAT（一個公網ip）.網路環境還是很簡單的,目的就是在路由器和區域網間加一台透明防火牆進去,同時完成cache server的功能.我不想讓防火牆做nat,一是因為已經有了一台路由器,要物盡其用.二是nat其實也是很耗資源的事（尤其當下面的clients特別多時,對router cpu的要求還是很高的,我就碰到過一台cisco 75XX 路由器拖1000個用戶5分鐘死一次機的事,後來不得已架了一台PIX做NAT）還是分分開的好.三是萬一cache server 趴下了,只要把鏈路重新旁接一下,網路照樣用（頂多性能不好）,不會影響用戶.
防火牆的平台為一P4 2.4G server, 512M內存,2×80GHD,兩塊NIC.安裝的是RH8.0 kernel 2.4.20（原來想裝gentoo的,但要命的1.4到現在才剛到RC3,等不及了,偷了一把懶~_~）
先下載bridge做橋接,bridge的作用就是讓兩塊網卡變成一個橋設備,讓兩端的網路埠完全透明地轉發packets,而讓iptables起到blocking的作用.下載地址為http://bridge.sourceforge.net（什麼？sourceforge？不知道上不去啊！沒辦法,因為眾所周知的原因,大夥找https proxy繞上去吧！）需下載bridge-utils-0.9.6.tar.gz和bridge-nf-0.0.7-against-2.4.19.diff.（這個很重要,否則iptables無法攔截轉發過來的packets）
再到http://netfilter.samba.org下載iptables,併到www.linuxhq.xom下載最新的kernel及patch.
在/usr/src下釋放kernel,進入/usr/src/linux,並把bridge的patch打上,patch –p1 < bridge-nf-0.0.7-against-2.4.19.diff.
還有個問題,此時用ifconfig看到的流量,在過了4G之後就會被置0,然後從頭開始,很不爽.這是因為 'RX bytes' and 'TX bytes'等變數的數據類型為 unsigned long,即2^32,到了4G就重置,其實把它們改成unsigned long long（2^64）就可以了.更改
/usr/src/linux/include/linux/netdevice.h 和../linux/net/core/dev.c,以下是打patch時的記錄,找到文件改掉相應代碼即可.
diff -urN linux-orig/include/linux/netdevice.h linux/include/linux/netdevice.h
--- linux-orig/include/linux/netdevice.h
linux/include/linux/netdevice.h
@@ -96,10 96,10 @@
struct net_device_stats
{
- unsigned long rx_packets; /* total packets received */
- unsigned long tx_packets; /* total packets transmitted */
- unsigned long rx_bytes; /* total bytes received */
- unsigned long tx_bytes; /* total bytes transmitted */
unsigned long long rx_packets; /* total packets received */
unsigned long long tx_packets; /* total packets transmitted */
unsigned long long rx_bytes; /* total bytes received */
unsigned long long tx_bytes; /* total bytes transmitted */
unsigned long rx_errors; /* bad packets received */
unsigned long tx_errors; /* packet transmit problems */
unsigned long rx_dropped; /* no space in linux buffers */
diff -urN linux-orig/net/core/dev.c linux/net/core/dev.c
--- linux-orig/net/core/dev.c
linux/net/core/dev.c
@@ -1689,7 1689,7 @@
int size;
if (stats)
- size = sprintf(buffer, "%6s:%8lu %7lu %4lu %4lu %4lu %5lu lu %9lu %8lu %7lu %4lu %4lu %4lu %5lu %7lu lun",
size = sprintf(buffer, "%6s:%8llu %7llu %4lu %4lu %4lu %5lu lu %9lu %8llu %7llu %4lu %4lu %4lu %5lu %7lu lun",
dev->name,
stats->rx_bytes,
stats->rx_packets, stats->rx_errors,
這樣編譯后,ifconfig的4G重置問題就被消除了.
下面就是重新編譯kernel了,make menuconfig.....
注意,除了要把該編譯進去的驅動編進去（象NIC什麼的,我第一次編時就漏了一塊網卡,汗ing）還要把最重要的iptables給enable起來,進入Network Options,選中Network Packet filtering (replaces ipchains),這樣就會打開IP: Netfilter Configuration 子菜單,進入,選中裡面所有的選項,然後回到Network Options菜單,選中「802.1d Ethernet Bridging」 和 「netfilter (firewalling) support」兩個選項（一看就知道很重要.）
退出,開編.
make dep
make modules
make modules_install
make bzImage
完成後cp /usr/src/linux/arch/i386/bzImage /boot/bzImage.bridge
然後把System.map拷至/boot下再在lilo或grub中更改相應配置即可（這就不細說了,重編kernel的文章到處都有.）
重起,如果正常,開始下一步,tar –xzvf bridge-utils-0.9.6.tar.gz.
cd ./bridge-utils-0.9.3
make
make install（編譯bridge並安裝,具體的安裝目錄可在make文件生成時指定.）
完畢后可以先試試bridge的基本命令.如brctl addbr mybridge （創建bridge mybridge,然後用ifconfig查看會發現多了一個mybridge埠）
brctl addif mybridge eth0
brctl addif mybridge eth1 （把eth0和eth1包含到mybridge中）
然後需要對各埠做以下動作
ifconfig eth0 0.0.0.0 promisc
ifconfig eth1 0.0.0.0 promisc（打開混雜模式）
這時橋接即已成功,從router和區域網轉發過來的消息開始通過機器,這時你會發覺滿屏的tcp/ip轉發消息讓你無法下手,硬碟也狂轉,這是syslogd和klogd做的怪,log的priviledge設得太低,改掉它.telnet 進入機器（這時你只能telnet,console上已經泛濫成災.）kill 掉syslogd和klogd,在/etc/syslog.conf中屏蔽掉*.info至/var/log/messages的rules（一開始我不知道這個,結果第一天下午連到網路上去,第二天上午一看,/var/log下的messages居然有2.7G,My God）,並修改/etc/init.d/syslog,在daemon klogd 一項后加上-2 -c 4（提高klogd的消息轉發級別）,重起syslogd,哈哈,整個世界清凈了！！！
其實我們也可以寫個啟動腳本放在/etc/init.d中,如下：
#!/bin/bash
. /etc/init.d/functions
return=$rc_done
start() {
echo "Starting service bridge tranfw"
brctl addbr tranfw || return=$rc_failed
brctl addif tranfw eth0 || return=$rc_failed
brctl addif tranfw eth1 || return=$rc_failed
ifconfig eth0 0.0.0.0 promisc || return=$rc_failed
ifconfig eth1 0.0.0.0 promisc || return=$rc_failed
brctl sethello tranfw 1 || return=$rc_failed
brctl setmaxage tranfw 4 || return=$rc_failed
brctl setfd tranfw 4 || return=$rc_failed
ifconfig tranfw 192.168.9.11 broadcast 192.168.9.255 promisc up || return=$rc_failed
/sbin/route add default gw 192.168.9.1 netmask 0.0.0.0 metric 1
echo -e "$return"
}
stop() {
echo "Shutting down service bridge tranfw"
brctl delif tranfw eth0 || return=$rc_failed
brctl delif tranfw eth1 || return=$rc_failed
brctl delbr tranfw || return=$rc_failed
echo -e "$return"
}
status(){
ifconfig tranfw
brctl show tranfw
}
restart(){
$0 stop && $0 start || return=$rc_failed
}
case "$1" in
start)
start
;;
stop)
stop
;;
restart)
restart
;;
status)
status tranfw
RETVAL=$?
;;
*)
echo $"Usage: $0 {start|stop|status|restart}"
RETVAL=1
esac
exit $RETVAL
這樣啟動時就自己起來了.
下一步,我們可以開始配iptables了
tar –xjvf iptables-1.2.7.tar.bz2
cd ./iptables-1.2.7
make
make install
安裝很簡單.關鍵是如何用,我懶,抄了個腳本就開用了,如下：
#!/bin/bash
#Flush all rules from the chains
iptables -F
#Delete all user created chains (mainly KEEP_STATE chain)
iptables -X
#############################################################
##Create special chain KEEP_STATE
iptables -N KEEP_STATE
iptables -F KEEP_STATE
##Drop bad states
iptables -A KEEP_STATE -m state --state INVALID -j DROP
iptables -A KEEP_STATE -m state --state RELATED,ESTABLISHED -j ACCEPT
#############################################################
#Drop RST/ACKs to limit OS detection through pinging
iptables -A FORWARD -p tcp --tcp-flags RST RST,ACK -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "RST/ACK: "
iptables -A FORWARD -p tcp --tcp-flags RST RST,ACK -j DROP
#Deny pings from outside
iptables -A FORWARD -p icmp --icmp-type 0/0 -d 192.168.9.0/24 -j ACCEPT
iptables -A FORWARD -p icmp --icmp-type 0/0 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Drop Echo Reply: "
iptables -A FORWARD -p icmp --icmp-type 0/0 -j DROP
#Drop potential SQL Worm
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 1433 -j ACCEPT
iptables -A FORWARD -p tcp --dport 1433 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Possible SQL Worm: "
iptables -A FORWARD -p tcp --dport 1433 -j DROP
#Pass all boxes to the keep_state chain
iptables -A FORWARD -j KEEP_STATE
######################################################################
##Set up UDP
#Outgoing Traceroute
iptables -A FORWARD -p udp -s 192.168.9.0/24 --sport 32769:65535 --dport 33434:33523 -j ACCEPT
#Incoming Traceroute
iptables -A FORWARD -p udp -s 192.168.9.0/24 --dport 32769:65535 --sport 33434:33523 -j ACCEPT
#Time exceeded
iptables -A FORWARD -p udp -s 192.168.9.0/24 --dport 11 -j ACCEPT
#Port not found
iptables -A FORWARD -p udp -s 192.168.9.0/24 --dport 3 -j ACCEPT
#DNS
iptables -A FORWARD -p udp -s 192.168.9.0/24 --dport 53 -j ACCEPT
#SNMP
iptables -A FORWARD -p udp -s 192.168.9.0/24 --dport 161 -j ACCEPT
#DHCP
iptables -A FORWARD -p udp -s 192.168.9.0/24 --sport 68 --dport 67 -j ACCEPT
#Time Server
iptables -A FORWARD -p udp -s 192.168.9.0/24 --sport 1024:65535 --dport 123 -j ACCEPT
######################################################################
##Allow outward browsing
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 80 -j ACCEPT
##Allow outward ssh
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 22 -j ACCEPT
##Allow outward ftp
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 21 -j ACCEPT
##Allow outward telnet
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 23 -j ACCEPT
##Allow outward smtp
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 25 -j ACCEPT
##Allow outward pop
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 110 -j ACCEPT
###Block HTTP Request from outside that are not authorized
iptables -A FORWARD -p tcp --dport 80 -j DROP
iptables -A FORWARD -p tcp --dport 80 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Dropped HTTP: "
#Allow Telnet out
iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 23 -j ACCEPT
#Deny ports
##telnet
iptables -A FORWARD -p tcp --dport 23 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Denied Telnet: "
#Deny ports
##telnet
iptables -A FORWARD -p tcp --dport 23 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Denied Telnet: "
iptables -A FORWARD -p tcp --dport 23 -j DROP
##Deny BO
iptables -A FORWARD -p udp --dport 31337 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Denied BO: "
iptables -A FORWARD -p udp --dport 31337 -j DROP
#Deny and log ftp on all others
#iptables -A FORWARD -p tcp -s 192.168.9.0/24 --dport 21 -j DROP
#iptables -A FORWARD -p tcp --dport 21 -m limit --limit 5/minute -j LOG --log-level notice --log-prefix "Denied FTP: "
#Deny suspicious traffic
iptables -A FORWARD -p tcp --destination-port 1080 -j DROP
iptables -A FORWARD -p tcp --destination-port 8080 -j DROP
iptables -A FORWARD -p tcp --destination-port 8000 -j DROP
iptables -A FORWARD -p tcp --destination-port 8081 -j DROP
##Allow all outgoing traffic
iptables -t nat -A PREROUTING -p tcp --dport 80 -j REDIRECT --to-ports 3128 （轉發至squid）
iptables -A FORWARD -s 192.168.9.0/24 -j ACCEPT
#Set to drop all packets not accepted by rules above
iptables -A FORWARD -j DROP
也把它放到/etc/init.d中,這樣就可以隨著xinet.d一起啟動了.iptables的語法這裡就不詳細說了,網上很多.
就是squid的配置,這需要仔細說一下,隨RH發布的squid的配置有問題,性能受限.建議先卸除,到www.squid-cache.org下載最新的2.5stable1.然後重新編譯,編譯前需做以下準備工作：
1.編輯/usr/include/bits/types.h中__FD_SETSIZE值至32768
2.ulimit -HSn 32768
然後到squid src包目錄編譯squid ,./configure, make all, make install.....
squid會安裝在預設的/usr/local/squid下,squid的可執行文件在安裝目錄的bin子目錄下,配置文件在etc子目錄下
Squid配置文件為：/usr/local/squid/etc/squid.conf,以下為我的配置：
http_port 8080
cache_mem 32 MB
cache_swap_low 90
cache_swap_high 95
maximum_object_size 4096 KB
cache_dir ufs /usr/local/squid/cache 10000 16 256
cache_access_log /usr/local/squid/logs/access.log
cache_log /usr/local/squid/logs/cache.log
dns_nameservers 202.96.209.5
unlinkd_program /usr/local/squid/bin/unlinkd
acl acllist src 192.168.9.0/255.255.255.0
acl regular_days time MTWHF 8:00-19:00
acl movie urlpath_regex "/etc/squid/banned.list"（把過濾關鍵字寫到文件中去）
acl banned url_regex iij4u.or.jp(做些控制)
acl cache_prevent1 url_regex cgi-bin /?
acl cache_prevent2 url_regex Servlet
acl all src 0.0.0.0/0.0.0.0
acl manager proto cache_object
acl localhost src 127.0.0.1/255.255.255.255
acl SSL_ports port 443 563
acl Safe_ports port 80 # http
acl Safe_ports port 21 # ftp
acl Safe_ports port 443 563 # https, snews
acl Safe_ports port 70 # gopher
acl Safe_ports port 210 # wais
acl Safe_ports port 1025-65535 # unregistered ports
acl Safe_ports port 280 # http-mgmt
acl Safe_ports port 488 # gss-http
acl Safe_ports port 591 # filemaker
acl Safe_ports port 777 # multiling http
http_access deny movie regular_days
no_cache deny cache_prevent1
no_cache deny cache_prevent2
http_access deny banned
http_access allow all
http_access deny !Safe_ports
http_access deny CONNECT !SSL_ports
icp_access allow all
http_access allow manager localhost
http_access deny manager
http_access allow localhost
http_access deny all
client_lifetime 2 hours
half_closed_clients off
cache_effective_user squid
cache_effective_group squid(請注意squid對cache及log目錄有讀寫權)
httpd_accel_host virtual
httpd_accel_port 80
httpd_accel_with_proxy on
httpd_accel_uses_host_header on
同時在/etc/init.d中的啟動腳本里加上這句ulimit -HSn 32768
這樣squid啟動時在cache.log中就可以看到如此語句：
2003/03/05 09:30:53| Starting Squid Cache version 2.5.STABLE1-20030303 for i686-pc-linux-gnu...
2003/03/05 09:30:53| Process ID 12939
2003/03/05 09:30:53| With 32768 file descriptors available
此時的squid擁有32768個file descriptors（預設為1024）,如果用戶很多,預設值會成為瓶頸,極度影響squid的速度（它會報錯說running out of file descriptors）,我曾經在板上發問,結果沒人理我:-((,還是自力更生,豐衣足食.
啟動squid：/usr/local/squid/bin/squid -D
大功告成！整個出internet的http流量就全朝squid走了,用戶不知道,同時又可完成封鎖的功能.
好辛苦啊！
