Stefano Palazzo's questions

I'm trying to share a GnuPG key pair by importing it into each machine. This is how I'm doing it:

gpg --allow-secret-key-import --import secret.gpg.key
gpg --import public.gpg.key

The keys have been exported with -a.

After doing this, the public key is shown correctly when I do a gpg --list-keys, but the private key isn't (gpg --list-secret-keys).

What am I doing wrong?

By the way: I'm doing this with Puppet, so any solution that doesn't require me to type stuff in (--edit-key and the like) would be appreciated.

To indent a whole block of code in nano, after selecting it with alt+a, I have to press alt+{, which doesn't work on my german layout macintosh keyboard (and would be pretty annoying even if it did).

How do I change and customize nano's keyboard shortcuts?

On a multi user system, I want to measure each user's CPU usage in seconds of cpu time. For the purpose of this measurement, I assume that if a PID belongs to a user, this user is causing the CPU time - that is I'm ignoring daemons and the kernel.

Currently I'm doing this, every five seconds:

1. Get each user and the PIDs they are running via ps aux
2. For each PID, get x, the sum of utime, cutime, stime and cstime from /proc/[pid]/stat
3. calculate t = x / interval (interval isn't always exactly 5 seconds when there's high load)

If I run this, I get sensible looking values. For instance: A user on this system was spinning in python (while True: pass), and the system was showing round about 750 milliseconds of CPU time per second. When the system hung for a bit, it reported 1600ms for one 1-second inverval. Which seems about right, but I undestand that these values can be deceiptful, especially given I don't really understand them.

So my question is this:

What is a fair and correct way to measure CPU load on a per-user basis?

The method has to be rather accurate. There might be many hundreds of users on this system, so extracting percentages from ps aux will not be accurate enough, especially for short-lived threads which many pieces of software like to spawn.

While this might be complicated, I absolutely know it's possible. This was my starting point:

The kernel keeps track of a processes creation time as well as the CPU time that it consumes during its lifetime. Each clock tick, the kernel updates the amount of time in jiffies that the current process has spent in system and in user mode. — (from the Linux Documentation Project)

The value I'm after is the amount of seconds (or jiffies) that a user has spend on the CPU, not a percentage of system load or cpu usage.

It's important that we measure CPU time while the processes are still running. Some processes will only last for half a second, some will last for many months - and we need to catch both sorts, so that we can account for users' CPU time with fine granularity.

I tried loggin into an SSH server via Nautilus' "Connect to Server" feature by just leaving the password blank, but that didn't work.

How do I connect to my server using the key I have on my system?

I've set up a virtual machine configured with bridge networking. After making eight clones of it, the new machines can't connect to the network.

I've instructed VirtualBox to re-generate the machines' mac-addresses.

My router can do port-forwarding based on MAC addresses. That is, a specific MAC will get a specific IP, for which I can configure a set of ports to be forwarded. In order to easily change that set of ports, I'd like to have different connections in the Network manager.

How do I change the MAC address for a network connection?

I want to encrypt a file using AES-256. How can I do that quickly and easily, and how can I - or someone else -decrypt it again?

Assuming I have a Virtual Private Server from someone like Linode, running Ubuntu, how do I easily set up a PPTP VPN so that Ubuntu clients can connect to it?

By default, Ubuntu offers to connect to these PPTP networks. I want to set it up in such a way that it's very easy to connect to (I.e. without installing any packages on the client).

My VPS is running Ubuntu, and I'd like to be able to receive email to my domain.

How do I easily set up a mail server to do this?

From time to time, SSH sessions to my servers just freeze up and then neither Ctrl+C nor Ctrl+\ do anything.

How do I quit the session without just closing the Terminal it's in?

Clarification:

The session is locked as in, no keys are transmitted to the server. I'd like to shut it down gracefully, such that the user isn't logged in anymore.

I know I can kill the SSH client, but that will, in time, make the "2 users" info message I get from uptime meaningless. I'd like to know how many users are really logged in.

When looking for hardware, I often try to find out if the manufacturer employs open source developers and things like that. I know Canonical has relations with a couple of big computer companies, but who, apart from Canonical, is supporting the community effort directly? Is there a list?

I've set up SSL on my webserver, now I need two files:

• a certificate
• a certificate Key

How do I create a self-signed certificate for testing purposes?

My block devices are named sda, sdb and so on.

What happens when I get to sdz? Is the next one going to be called sdaa?

I usually edit Python files with nano, so I need four spaces instead of a tab. How can I get nano to automatically insert spaces for tabs, and control how many of them it inserts?

When I plug an external monitor into my laptop, the laptop screen is still considered my primary display. Since my laptop is on the right of my external monitor, I now have the Launcher sort of "in the middle".

How do I get the Launcher and the BFB onto my external monitor?

At the moment, I'm using mimetex to automatically convert a bunch of equations into gif-images.

The program is fast, but the images look pretty bad:

How can I (quickly) convert a LaTeX equation into an SVG image on the command line?

I often use Python 3000's interactive interpreter, but it's missing the convenience of tab-completion I'm used to from ipython, which isn't available for Python 3.x.

How do I enable tab completion in all of them, 2.6 and 3.x?

Since I rarely use caps-lock, I'd like to map the key to a middle mouse click instead. I would also like to map Alt+Caps Lock to the original function of the caps lock key, should I ever need it.

I can map any keyboard shortcut to xdotool click 2, but the Gnome Keyboard Shortcuts dialog won't let me assign a command to the caps-lock key, even with modifiers.

I know this is a bit of a strange undertaking; How would I go about doing it?

Putting any gtk.Widget inside a gtk.Menu works, somewhat, by putting an empty gtk.MenuItem in first. It doesn't seem to be reliable - but it basically works. With an appindicator.Indicator, it doesn't work. The entry just stays empty. Even if you set the child[border_size] to 32, for example.

The only widget that works is a gtk.Label (plus the MenuItem, ImageMenuItem and so on of course).

Now the Me Menu has got a gtk.Entry in it, so it must surely be possible to do this Properly. How do I do it?

Here's some example code to explain what I'm on about:

    self.item = gtk.MenuItem()
    self.item.add(gtk.Label("hello world!"))
    self.menu.append(self.item)

The above code works, this however doesn't:

    self.item = gtk.MenuItem()
    self.item.add(gtk.Entry())
    self.menu.append(self.item)

Where self.menu is an appindicator.Indicator, if it were a gtk.Menu, the Entry would be at least displayed.

Here's some working code:

import gtk
import appindicator

class AppIndicator (object):

    def __init__(self):
        self.ind = appindicator.Indicator("hello world client",
            "distributor-logo", appindicator.CATEGORY_APPLICATION_STATUS)
        self.ind.set_status (appindicator.STATUS_ACTIVE)
        self.menu = gtk.Menu()
        item = gtk.MenuItem()

        item.add(gtk.Label("hello world"))
        # item.add(gtk.Entry())

        self.menu.append(item)
        self.menu.show_all()
        self.ind.set_menu(self.menu)


indicator = AppIndicator()
gtk.main()

I'd like to connect to a wireless access point from the command-line. Ideally, I'd only need the name of the AP. But the hardware-address would work as well.

I know I can use nmcli to connect to a managed network connection, but in my case, the access point may not be configured for Network-Manager yet (See the difference between the output of nm-tool and nmcli con).

Example output of nmcli:

Auto pwln       3a3d62b1-bbdf-4f76-b4d2-c211fd5cfb03   802-11-wireless [...]  
Wired Network   aa586921-accf-4932-98c4-c873c310f08e   802-3-ethernet [...]  
Cisco-UDP Uni   7f94847b-04dc-40b7-9955-5246fb77cc65   vpn [...]  
T-mobile (D1)   867f345a-cbbf-4bd4-b883-a5e5ae0932f0   gsm  [...]

Example output of nm-tool:

State: connected
- Device: eth1  [Auto pwln] ----------------------------------------------------
[...]
  Wireless Access Points (* = current AP)
    *pwln: Infra, [...], Freq 2472 MHz, Rate 54 Mb/s, Strength 80 WPA WPA2
    WLAN:  Infra, [...], Freq 2422 MHz, Rate 54 Mb/s, Strength 20 WPA WPA2

[...]

How do I connect to an access point that may or may not be known to NM?

Extra: Finding out if the connection needs a pass-phrase, and submitting it on the command-line as well would be great too (that is to say It'd be nice if network-manager wouldn't pop open any keyring dialogues or errors on the gui)