Cran down indefinitely

[Roger R Dingledine <arma@mit.edu>]

So, people have been teardropping mitnet in a big way. I've been hit with 9
teardrops in the past evening, spaced pretty evenly. (For those who don't know,
teardrop is an implementation of the recent exploit of a bug in the linux
network IP fragmentation code that allows people to crash/reboot linux/windoze/etc
boxen remotely. Check out your favorite linux/bugtraq/etc website for more info.)

Anyway, I'd patched my kernel and recompiled, but hadn't rebooted by the first
hit. But I was safe after that. (And I've been running a tcpdump to log well
over a hundred attacks on various computers on my subnet this past evening. Eesh.)

But anyway, they hit all of Luka's machines, and since Luka isn't there, those
machines are still down. I don't have time to keep breaking into his room to
reboot, and I don't have the inclination to break into and recompile cran's kernel
anytime soon. It's possible I could boot it long enough for people to retrieve
files if they need that.

But anyway, this means cvs and irc are down for a while. Omega's already been
looking at a 'unified machine' approach to these various applications, so all is
not lost.

Let me know if you want an irc server back (and give me a server already patched
with whatever we needed to patch, ideally), and I'll make that happen.

Hopefully, cran will be back in a few days, and all will return to normal.

--Roger

===
SEUL-Leaders list, seul-leaders-request@seul.mit.edu
===