Hello,
I am experiencing an issue with squid 3.5.5 and FreeBSD 10.1 where tcp_outgoing_address correctly rewrites the source address of outgoing packets, but fails to bind the socket to the correct interface. I've been using this kind of setup/configuration for quite some time (since the squid 2.7 days), so I believe something between squid 3.3 and squid 3.5 has broken this behavior, or it is a problem with FreeBSD 10.x that I do not understand. FWIW squid 3.3.3 on FreeBSD 9.x behaves correctly with the same config.
I've been able to replicate this on a non-production VM with a simple config. For example, I have two WAN interfaces: WAN1 (em0/192.168.92.246) and WAN2 (em1/10.8.8.10), and a LAN interface (em2/192.168.5.1/24). My default route points to the gateway for WAN1/em0. I have configured a tcp_outgoing_address for a single host on the LAN (192.168.5.2) to use outgoing address 10.8.8.10 (WAN2/em1). HTTP request packets from squid end up going out em0 (the default route interface) with an ip of 10.8.8.10. The source IP is correct but the packets egress the wrong interface, so they are obviously dropped by the upstream router on that incorrect interface.
The following is my test squid.conf and some basic network config. While I normally use squid as a transparent proxy in conjunction with the PF firewall and routing enabled, for the purposes of replicating the issue in my VM test environment I have disabled PF, routing, and connected directly to squid via my browser's proxy config. Attempts to load webpages time out unless I remove use of the tcp_outgoing_address directive.
# begin test squid.conf
acl SSL_ports port 443
acl Safe_ports port 80 # http
acl Safe_ports port 21 # ftp
acl Safe_ports port 443 # https
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
acl CONNECT method CONNECT
http_access allow localhost manager
http_access deny manager
http_access deny !Safe_ports
http_access deny CONNECT !SSL_ports
http_access allow all
acl wan2 src 192.168.5.2
tcp_outgoing_address 10.8.8.10 wan2
follow_x_forwarded_for allow localhost
acl_uses_indirect_client on
log_uses_indirect_client on
http_port 3129
cache_replacement_policy heap LFUDA
maximum_object_size 768 MB
cache_dir aufs /squid/cache 2048 32 512
access_log daemon:/squid/logs/access.log squid
cache_log /squid/logs/cache.log
cache_store_log none
logfile_rotate 0
pid_filename /var/run/squid.pid
# end test squid.conf
em0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=9b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM>
ether 00:0c:29:a3:33:93
inet 192.168.92.246 netmask 0xffffff00 broadcast 192.168.92.255
nd6 options=9<PERFORMNUD,IFDISABLED>
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
em1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
options=9b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM>
ether 00:0c:29:a3:33:7f
inet 10.8.8.10 netmask 0xffffff00 broadcast 10.8.8.255
inet 10.8.8.11 netmask 0xffffff00 broadcast 10.8.8.255
inet 10.8.8.12 netmask 0xffffff00 broadcast 10.8.8.255
inet 10.8.8.13 netmask 0xffffff00 broadcast 10.8.8.255
inet 10.8.8.14 netmask 0xffffff00 broadcast 10.8.8.255
nd6 options=9<PERFORMNUD,IFDISABLED>
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
root# netstat -rn | grep default
default 192.168.92.2 UGS em0
I am hoping someone can shed some light on whether or not this may be a squid bug or something going on with FreeBSD. Thank you.
-Nick
_______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
