So last week I screwed up in a royal fashion. I caused 1 core on 12,000 servers to be running at 100% with a zombie process. In conversation with a friend he asked “So how many BogoMIPS were you wasting?”, of course I had to find out.
So I grepped out the first core on each box’es BogoMIPS value and summed it up.
Per second: 61099300 BogoMIPS
If you take that figure and multiply it up for 12 hours.
Total: 2639489760000 Million Instructions
For more information on BogoMIPS please see this nice FAQ.
“If your ISP has registered with the appropriate databases, our servers are updated on a regular basis so you will have to wait until the next refresh to gain access.”
Would you interpret that to mean they are using whois lookups on IP addresses/blocks to check its location for limiting content to certain contries?
So is it a test? No. It is something I have been pondering for a while. It never really gives me anything, it just provides me with something to-do, but it apparently could cause some illness. What I hear you say! Like all the other things that might cause illness? Well I don’t know if they do or don’t. Just that I have avoided one substance the past 80 or so hours. At the same time though I do miss it, it was my always with friend. (Now that just sounds a bit psyco, but some people might understand what I mean.) I am gonna see how long I can mange to go without.
If like me you like to pipe things around in the shell then when I looked at aircrack-ng I wanted to script using it. Therefore I wrote a simple patch for the current version that disables buffering to allow the stdout/err to act like they are writing to a log file so good for timely greping around in a bash script. Once I have my WEP scripts polished I will post them up. 
You can find the patch here.
I often work on many machines and end up with multiple sessions to each. Therefore I finally got around to checking out the ControlMaster and ControlPath configuration of ssh. I added the following to my ~/.ssh/config which is for a virtual machine I use on my laptop, so no network issues in play.
Host vm
Hostname 192.168.100.101
User dan
ControlMaster auto
ControlPath /Users/dan/.ssh/master-vm
Then in a terminal I ran the following, twice. Once on its own then with an ssh session running in another terminal.
for X in `seq 1 5`; do time ssh vm "pwd; uptime; hostname;" >/dev/null; done;
Without avg.: 0.268s
With avg.: 0.035s
I think that is quite impressive.
For the raw terminal output, please see here.
Update
Of course you can do this dynamically.
Host *
ControlMaster auto
ControlPath ~/.ssh/master-%r@%h:%p
Then each socket will be used only when the user, host and port are the same.
Spotted today on Soho Street.
Looks like they are heading out to get more people on the Wagon
Took this picture last night, I really liked the lines in the sky, almost looks like pencil.
Sunset
Not sure how I was tinkering with the exposure, but I think this looks cool.
Green sky at night?
So I am setting up something to automatically download iPlayer episodes and needed to install iplayer-dl onto a Centos 5 machine. I don’t like installing stuff on a host without using package management therefore I have created a RPM for it.
You can find the SPEC, RPM and SRPM here.
Well would be apart from hosing the box within 3 hours.
Once Xen is up an running, mini boxes will be doing it all.
Update: An image I took a while back.
Chinatown Lanterns Jan 2009
