Anonymous's questions

I have a Linux system (Slackware Linux 13.0) with two network interfaces. Let's call them NIC0 and NIC1

My goal is to make NIC0 to appear as eth0 in the system.

I know this can be achieved via udev rules that map network aliases to MAC addresses of network interfaces. In Slackware Linux the file /etc/udev/rules.d/70-persistent-net.rules contains such rules.

The trickiest part of my problem is that I need to fake the MAC address of NIC0. I know I can dynamically change the MAC addres of a network interface with the command:

ifconfig eth0 hw ether <new MAC address>

Do you see the problem? This supposes that the network interfaces are already set up.

So my question is: If I would have an udev rule for NIC1(the one that shall go up as eth1, with its original MAC address), would it be enough for the system to bring the other network interface (NIC0) as eth0 by default? This way I could change its MAC address later, after the udev machinery completes and the network aliases are brought up.

Last week I had to deal with the following situation: some of my users complained that they got emails telling them that their computers are infected and they should go to a certain site for a free scan. Further investigation revealed that the site in question is a typical malware spreading site: it displays a fake scan window and tricks the inexperienced users into downloading a trojan.

By pinging the domain, I've got the IP address of that site. A whois on that IP revealed to me the hosting company that hosted the malware site. I've emailed their abuse team pointing them at that site. They choosed to ignore my email and the malware site is still up and running ATM.

Of course I'm upset by such attitude. I'd like to hear from you how you deal with such situations. I'd also like you to comment my strategy.

My users access the Internet over a proxy server. That malware site is blocked now. But I want to go one step further. I want to find out all sites that hosting company is hosting and block them. Can you suggest a workflow and tools for performing this research?

I have a LAN of Linux computers that get access to the Internet over NAT. The gateway runs a restrictive firewall. I have ICMP echo requests and replies allowed to pass through the firewall on the gateway, so I can ping any external host from a LAN computer, but mtr and traceroute to an external host hang at the gateway hop.

What additional firewall rules should I have so these tools would work properly?

Many package managers in Linux distributions and BSD flavors track package dependencies. But before a package goes into a repository, it is investigated by a maintainer, who tracks the dependencies manually.

What software can be used to track package dependencies? Plain text files are not very good, because they are flat. Outliners that can keep data structured in trees are better, but not ideal, since dependencies do not always form a tree, sometimes they form a graph.

Please bear in mind that I'm not looking for something that can resolve dependencies automatically. I'm looking for an app that could store the dependencies between packages and allows queries to display the dependences for packages (like package managers).

I have a computer running Linux. It has qemu installed. It has a bridge interface br0 configured with the following settings:

IP: 10.1.1.1
Netmask: 255.255.255.0

There is also a tap0 interface that is added to br0. tap0 is supposed to be used for networking guests running inside qemu.

I'm running Slax Linux from its LiveCD inside qemu like this:

qemu -kernel-kqemu \
 -net nic,vlan=0,macaddr=aa:aa:aa:aa:aa:aa,model=pcnet \
 -net tap,vlan=0,ifname=tap0,script=/etc/qemu-ifup \
 -m 512 \
 -cdrom slax-6.0.7.iso \
 1>stdout.log 2>stderr.log 

I'm booting Slax in text mode (don't need X, because my goal is to test networking). I'm configuring the network interface in the guest like this:

IP: 10.1.1.4
Netmask: 255.255.255.0
Gateway: 10.1.1.1 

Everything works well, I can ping in both directions:

10.1.1.1 -> 10.1.1.4
10.1.1.4 -> 10.1.1.1

Now I'm replacing the Slax LiveCD with a qemu image, that contains a freshly installed OpenBSD 4.5. I run it like this:

qemu -kernel-kqemu \
 -net nic,vlan=0,macaddr=aa:aa:aa:aa:aa:aa,model=pcnet \
 -net tap,vlan=0,ifname=tap0,script=/etc/qemu-ifup \
 -m 512 \
 -hda obsd.img \
 1>stdout.log 2>stderr.log 

OpenBSD boots, I'm configuring its network interface exactly the same way:

IP: 10.1.1.4
Netmask: 255.255.255.0
Gateway: 10.1.1.1

Now I can't ping in any direction.

The qemu logs don't contain any hints.

Also this is not a firewalling issue: OpenBSD 4.5 doesn't have pf enabled by default. Snippet from its /etc/rc.conf:

pf=NO # Packet filter / NAT

If one doesn't want to dig into sophisticated backup solutions, it is always possible to write a shell script that gathers data and compresses it. Because of privacy issues one might want to encrypt the backup archive by using a suitable tool (zip/rar with password, gpg etc).

Now when it comes to automation, the problem of revealing the password steps in which creates the following issues:

• When running the backup script from console, the password is just entered interactively.
• When running the script via cron, this is impossible.
• Hardcoding the password in the shell script is bad and ugly.

I believe this is a common situation for this problem. What is the generic or typical solution for the problem of encrypting data with a password, but in automated way?

Such known tools like iftop/iptraf display network I/O per interface and per connection. Is there a way to see network I/O statistics per process?