I found a tool called
pv
via a Hacker News thread. pv or pipe viewer
allows you to view data as it is pulled out of a Unix pipe. This is really
helpful when dealing with long running commands. I used it today to check on
the progress of encrypting a large tar archive.
$ pv archive.tar.xz | gpg --sign --symmetric - > archive.tar.xz.gpg
6.51GiB 0:08:52 [6.23MiB/s] [======================> ] 70% ETA 0:03:47
While pv is running you get a progress, time elapsed, speed, a progress bar,
70% complete and an estimation of time until complete.
One of the facilities at
campGND
is going to be a wireless network. The
hope is to have the network running for the majority of the time. I have built
a wireless network at a campsite before, that was made easier by having
guaranteed bandwidth from a satellite terminal.
The plan is to have a wireless network for the campsite served by a
MikroTik
. Using a wireless bridge to reach to the farmhouse. The
farmhouse is out of site of the fields we are planning to use. Instead of
having wifi doing the full jump I am going to run ethernet as far as possible.
At
campGND
we are depending on a few things that could be fickle.
-
BT Home Broadband
-
A long run of ethernet
-
Solar Cells and a battery for network power.
Our final back haul is the BT network the site is pretty off the grid for phone
reception so we are stuck with BT. We have to be able to make a long hop form
the farmhouse before we can do a wireless link down to the site. The solar
cells will provide enough to run wireless access points during the day. I think
at night we might be a little drunk to care.
I still need to do some testing of the wireless hardware but the plan is to use
the following.
-
100M run of Ethernet.
-
Injected POE, then split POE.
-
2x WRT54G's.
-
A
MikroTik
access point.
-
Solar Cells for with battery backing for night time.
campGND
is coming up and it is time to start talking about my projects for
the weekend. With our remote location I thought it would be fun to play with
something flaming and dangerous.
Rockets were the first thing that came to mind, I haven't done much with
rockets beyond launching fireworks a couple of times. Doing my first launches
at campGND would probably slow everything down somewhat. I got myself a starter
kit from
Model Rocket Shop
and some extra motors, for a bigger bang.
Iain and myself went out to
Balmedie Beach
to have a test run with my new
toy. We got a couple of videos of the rockets going up, excuse the portrait
slow-mo.
On the first launch the recovery canopy got slightly melted by the rocket
motor. This meant we didn't really have any recovery mechanism for the rest of
the launches. The beach was pretty deserted in the dunes so this wasn't a big
deal. At campGND loosing recovery could make things a little tricky.
For campGND I am planning on adding some telemetric data to the rockets, using
an Arduino and some sensors. I also want to try adding a camera to the nose
cone on a rocket.
For my new business cards I wanted to impose data from an experiment onto the
background of an image. For the best results I wanted to render the plot of data
onto a transparent png without any axis, values or the standard box.
After a while I got to
> png(filename="transplot.png",width=900,height=400,bg="transparent")
> plot(timestamp[0:1000],snd_cwnd[0:1000],type="h",yaxt="n",xaxt="n",ann="F",frame.plot="F")
> dev.off()
Which generates the following image.
I have been building kernels on my imac in virtual machines. This can take
a while and I wanted notifications when the build had finished running.
On my imac
$ while true
nc -l 4000 | say
done
This makes netcat(1) wait in a loop for any connections then pipes the output
into say(1).
On the vm
$ make buildkernel; echo "Build Complete" | nc -N imac 4000
Replace imac with the hostname or ip of your machine.
I have seen the build side hang and not close the connection until killed, I am
not sure why.