Skip to main content

Split multi homing-ish?

Hurrah for interface aliasing!

The problem: client has three (3) ADSL interfaces, and a Linux mail server. The mail server now wants to do MailMan things, which implies having a web interface. Typing port numbers is too complicated for some of the people using the web interface <cringe>, but we can’t port-forward from tcp/80 on the primary/defaultroute ADSL interface because that already goes to a busy Win2k3 server which would be a PITA to securely set up to proxy the traffic, and we can’t just bolt the secondary or tertiary ADSL router into a second card on the mail server and forward stuff based on interface because they’re also busy with RDP traffic and need to be administered by someone without much of a Linux clue (and who is also seriously busy, so can’t take the time away from his other duties to pick up said clue, even though he’s otherwise quite a bright lad) and because the router might get replaced we can’t rely on MAC addresses either. Traffic forwarded to this box from the non-defaultroute interfaces would be returned through the default route which belongs to a different ISP, which would of course break some things.

The solution: get the tertiary router to port-forward port 80 to an IP alias on the mail server’s ethernet interface, and use iproute2 to answer all traffic to the alias back out through the tertiary router.

Defaultroute belongs to Swiftel, tertiary ADSL belongs to ArachNet, so step one is to add these two definitions to /etc/iproute2/rt_tables:

42     arachnet
43     swiftel

Defining 9.9.9.1 to be the defaultroute, 9.9.9.3 to be the tertiary router's LAN address, 9.9.9.100 to be the primary LAN address (eth0) of the mail server and 9.9.9.200 to be an alias (ie, the address given to eth0:0), set up two very simple iproute2 tables and two very simple rules to reference them:

ip route add default via 9.9.9.3 table arachnet
ip route add default via 9.9.9.1 table swiftel
ip rule add from 9.9.9.200 table arachnet
ip rule add from 9.9.9.100 table swiftel

The rules for swiftel are not strictly necessary since the ordinary “boring” routing rules would take care of it, but they allow the mail server to be seamlessly default-routed through the secondary ADSL router instread if so desired.

I don’t know if this is the very best way to do it, but it’s working and everyone here’s happy. (-:

Comments

Popular posts from this blog

An Open and Shut response to Darl McBride

Hi, Darl. I see you’re being dishonest again . It’d be really nice if you could shoot straight for a change, but I think Kerry’s Dad will be selling snowplows in Hell first. Three years ago, when I first joined The SCO Group, we focused the company on the area that was most profitable and provided the most benefit to customers, investors, resellers, developers and employees: UNIX No, you focused the company on suing people, which was most profitable to lawyers and provided some golden parachutes for your buddies. People thought we were crazy. They were right. But since SCO owns the UNIX operating system The SCO Group does not own UNIX® in any sense of the word. The Open Group owns the UNIX trademark, definition and other rights , The SCO Group does not. The SCO Group doesn’t even own the UnixWare® or OpenServer® code, the rights to those are held by Novell and TSG use them only by permission and under certain conditions — which they have violate...

5x7 text dot-matrix in five minutes...

This is a fairly primitive toy I threw together today for a specific purpose, published in case it’s any use to others... feed it a list of words as command arguments, which it will then display as a 5x7 ASCII dotmatrix with ‘#’ as a dot & ‘_’ as a blank. /* * display text using a 5x7 bitmap font in ASCII letters */ static unsigned char font [] [5] = { { 0x00,0x00,0x00,0x00,0x00 }, // 0x20 32 { 0x00,0x00,0x6f,0x00,0x00 }, // ! 0x21 33 { 0x00,0x07,0x00,0x07,0x00 }, // " 0x22 34 { 0x14,0x7f,0x14,0x7f,0x14 }, // # 0x23 35 { 0x00,0x07,0x04,0x1e,0x00 }, // $ 0x24 36 { 0x23,0x13,0x08,0x64,0x62 }, // % 0x25 37 { 0x36,0x49,0x56,0x20,0x50 }, // & 0x26 38 { 0x00,0x00,0x07,0x00,0x00 }, // ' 0x27 39 { 0x00,0x1c,0x22,0x41,0x00 }, // ( 0x28 40 { 0x00,0x41,0x22,0x1c,0x00 }, // ) 0x29 41 { 0x14,0x08,0x3e,0x08,0x14 }, // * 0x2a 42 { 0x08,0x08,0x3e,0x08,0x08 }, ...

Citizens Augmenting Government Waste

So here we have Microsoft-funded “ Citizens Against Government Waste ” (CAGW) heavily criticising Massachusetts’s switch to an internationally accepted and open document standard (not open source, open standard — but I seem to remember a large monopolist who frequently confuses the two when it suits them). I remember CAGW, they were the organisation who had dead people writing in to support Microsoft in court a few years ago — thanks to sbergman27 for the link — is this “the dead hand of CAGW” at work again? Let’s follow the money and find out. Who stands to lose the most money and control if Massachusetts switches to an unencumbered document format? Big surprise, it’s CAGW sponsor Microsoft, through their dominant MS-Office suite. Why did CAGW list Microsoft’s suite last, after two other much-less-dominant examples which are waning anyway? If you’re inclined to wallow in additional irony, consider that at leas...