X is still dead...

I've emerged the following from the x11 overlay:

media-libs/mesa-9999
x11-libs/libdrm-9999
x11-drivers/xf86-video-intel-9999
x11-base/xorg-server-1.6.2

I enabled tiling (it was manually disabled before, as specified in this
gentoo-wiki article)

X is definitely more stable than before, I am able to log in to gnome and use
as normal, although sooner or later the GPU locks up again and I am just left
with a pointer and no keyboard.

Fortunately, this box's main purpose is as a server (primarily development :-),
so running X is not critical. For now, I am going to sit tight and just make
sure to continually run updates. I am going to wait a bit longer before I
upgrade to 2.6.31, simply because I need the kernel to be stable (beyond r2)
and do not want to sacrifice that just to get X running a bit smoother.

To see a more documented recap of what all I did with X, check out this gentoo bugzilla page

I started a couple of small projects this weekend. The first was getting SAMBA/cifs working again in the new network environment. I couldn't figure out why I couldn't see/access my SAMBA share on my gentoo box: nothing had really changed within the local network, other than a slightly different routing prefix... Why would that matter?

Then I remembered setting up my stateful firewall and documenting it on this very blog. Of course SAMBA isn't working, as I explicitly allow traffic over one particular Class C subnet, and since that change, all traffic is blocked. So after a quick few edits to my firewall configuration script and to smb.conf, I was again all set up.

The other small project was deploying some of the past sites I have completed on my web server, to act as a mini portfolio. I haven't decided whether I want to get the sites up and running as they do/did in production, or if I should just take screenshots. Right now I'm leaning towards screenshots, as I don't want to have to reconstruct each site's respective database and data.

This week I hope to get the chance to play around with Windows 7 a bit more and try and figure out why Gnome is running so slowly on my FreeBSD VM. I also plan on finally updating my Dell C610 laptop from Ubuntu 7.04 Feisty Fawn to 9.04 Jaunty Jackalope. This will mark the 5th release of Ubuntu I have used, as I started with the first release, 4.10 Warty Warthog, back in '04. I should take this opportunity to note that without Ubuntu, I would be nowhere near where I am today in terms of experience with Linux. I've always been a fan of Ubuntu's cheesy credo of "Linux for human beings," as it aims to simplify many of the complexities that are typically associated with Linux. Although I had used a few other distros prior to Ubuntu, namely RedHat, SuSE, and Mandrake, it was not until Ubuntu that I felt comfortable with the Linux kernel and using bash. So thank you Canonical for Ubuntu, and thank you to the Debian team for apt-get!

Lastly, this week, I think I'm going to give OSX on a VM another shot.

iptables complete

So after a weekend primarily away from my computer (probably for the better :-) I had the chance to sit down and work out iptables. Although the Gentoo Wiki article was helpful in getting the kernel configured correctly, I did not want to write an iptables start script- after all, iptables is a service, so why not utilize it as such? Instead, I used Linux Home Networking's tutorial to create my own firewall script, complete with logging. I was also fortunate to reference a pre-built firewall script, authored by a friend and coworker, that he uses as a standard template for all of his RHEL5 servers at UGA.

I created three chains, LOG_DROP, LOG_ACCEPT, and LOG_REJECT, to log packets when needed. I did not use LOG_DROP, nor LOG_REJECT (maybe in the future?), and instead am just using REJECT with --reject-with tcp-reset for TCP packets and icmp-port-unreachable for UDP packets to help mask the fact that I'm using a firewall. If the packets are simply dropped, it'd be pretty easy for an intruder to realize that I'm running a firewall because his packets would not result in the standard --tcp-reset or icmp-port-unreachable to indicate a nonexistent service. In the future, I may consider logging some of the rejected packets.

Right now the only services I am logging when accepted are SSH & SFTP (both use port 22), as logging every http request is impractical.

Unlogged accepts include http (port 80) and SAMBA (netbios-ssn, microsoft-ds, UDP 137 & 138). One key problem I had was allowing SAMBA to continue to work with the firewall enabled. With iptables running, I could only access my SAMBA shares by using the machine's IP, and not its hostname. While this problem was frustrating, it forced me to better manage my system by ensuring that samba was configured correctly (/etc/samba/smb.conf), my hostnames were set properly (it was not- only localhost was set in /etc/hosts), as well as resolv.conf (/etc/resolv.conf). This forum post eventually led me to realize that I needed to accept traffic on ports 137 & 138 on top of netbios-ssn (UDP 139), and microsoft-ds (TCP 445).

Further securing SAMBA, I specified the allowed source IPs on my network, both within smb.conf and iptables. If there's anything I've learned, its that redundancy in terms of security is never a bad idea.

Below is a (modified) iptables firewall script:

#!/bin/bash

# Stateful firewall for hostname

########################
# Ethernet Information #
########################

# Device: lo
# IP: 127.0.0.1
# Hostname: localhost

# Device: eth0
# IP: 192.168.2.3
# Hostname: hostname

# Flush all tables
iptables -F

# Remove all non-default cahins
iptables -X

#+-------------------------------+
#| Setup Firewall Process chains |
#+-------------------------------+

# Create a LOG_DROP chain for dropped incoming requests to be logged
iptables -N LOG_DROP
iptables -A LOG_DROP -j LOG --log-level info --log-prefix "Firewall-LOG_DROP: "
iptables -A LOG_DROP -j DROP

# Create a LOG_REJECT chain for rejected incoming requests to be logged
iptables -N LOG_REJECT
iptables -A LOG_REJECT -j LOG --log-level info --log-prefix "Firewall-LOG_REJECT: "
iptables -A LOG_REJECT -j REJECT

# Create a LOG_ACCEPT chain for accepted incoming requests to be logged
iptables -N LOG_ACCEPT
iptables -A LOG_ACCEPT -j LOG --log-level info --log-prefix "Firewall-LOG_ACCEPT: "
iptables -A LOG_ACCEPT -j ACCEPT


# Accept all connections from localhost
iptables -A INPUT -i lo -j ACCEPT

# Accept reply packets
iptables -A INPUT -m state --state ESTABLISHED,RELATED -j ACCEPT

# Accept PING requests
iptables -A INPUT -p icmp --icmp-type echo-request -j ACCEPT

#+-----------------------------------+
#| Accepted services that ARE Logged |
#+-----------------------------------+

#SSH & SFTP (TCP)
iptables -A INPUT -p tcp --dport ssh -j LOG_ACCEPT

#+---------------------------------------+
#| Accepted services that are NOT Logged |
#+---------------------------------------+
# HTTP (TCP)
iptables -A INPUT -p tcp --dport http -j ACCEPT

# HTTPS (TCP)
#iptables -A INPUT -p tcp --dport https -j ACCEPT

# TOMCAT (TCP)
#iptables -A INPUT -p tcp --dport 8080 -j ACCEPT
# If we're only using TOMCAT (no apache), we can forward traffic to 8080
#iptables -t nat -A OUTPUT -d localhost -p tcp --dport 80 -j REDIRECT --to-ports 8080
#iptables -t nat -A OUTPUT -d hostname -p tcp --dport 80 -j REDIRECT --to-ports 8080
#iptables -t nat -A PREROUTING -d hostname -p tcp --dport 80 -j REDIRECT --to-ports 8080

# RAILS (TCP)
#iptables -A INPUT -p tcp --dport 3000

# netbios-ssn (UDP)
#iptables -A INPUT -p udp --dport netbios-ssn -j ACCEPT
iptables -A INPUT -p udp --source 192.168.2.0/24 --dport netbios-ssn -j ACCEPT

# netbios-ssn (TCP)
#iptables -A INPUT -p tcp --dport netbios-ssn -j ACCEPT
iptables -A INPUT -p tcp --source 192.168.2.0/24 --dport netbios-ssn -j ACCEPT

# microsoft-ds (TCP)
#iptables -A INPUT -p tcp --dport microsoft-ds -j ACCEPT
iptables -A INPUT -p tcp --source 192.168.2.0/24 --dport microsoft-ds -j ACCEPT

# microsoft-ds (UDP)
#iptables -A INPUT -p udp --dport microsoft-ds -j ACCEPT
iptables -A INPUT -p udp --source 192.168.2.0/24 --dport microsoft-ds -j ACCEPT

# nmbd (UDP) required for SAMBA to send requests via broadcasting
iptables -A INPUT -p udp --source 192.168.2.0/24 --dport 137:138 -j ACCEPT
#iptables -A OUTPUT -p udp --dport 137:138 -j ACCEPT

# LOG & Drop malicious IPs
#iptables -A INPUT --source xxx.xxx.xxx.xxx -j LOG_DROP

# Reject remaining packets, do so with tcp-reset and icmp-port-unreachable
# so hackers don't know we're running a firewall
iptables -A INPUT -p tcp -i eth0 -j REJECT --reject-with tcp-reset
iptables -A INPUT -p udp -i eth0 -j REJECT --reject-with icmp-port-unreachable

I really enjoyed this project and learned a lot. After getting my basic firewall set up, I feel confident that I could do it again, now that I understand most of the iptables flags and commands
NAT translation and masquerading may be a project for a different day, but at least now I'm familiar with what they are and how they work.

Furthermore, I think this project will serve as a good segway into my next one, which is better learning SAMBA and how to configure it.

IPTables

It is high time I really sat down and figured out IPTables. Similar to my experience with SAMBA, I've used IPTables in the past, though I was more just shooting in the dark than masterfully configuring my ever-important firewall.

I think I'll get started on this first, as it is 1.) interesting 2.) important, and 3.) something I should have done a long time ago.

So, my initial project line-up as of right now is as follows:

1.) Learn how to properly and diligently manage ip tables
2.) Explore SAMBA further and get it working pristinely at my home set up
3.) Set up SSHFS and figure out how to use keys instead of tunneled clear-text passwords

Some useful links to aid me in my quest:
Gentoo-Wiki HOWTO_IPtables_and_stateful_firewalls
IP Tables Tutorial