May 8, 2021
The use-case might be a little specific but I’ll describe it nonetheless, hoping it may help someone. A while ago I read Practical Forensic Imaging by Bruce Nikkel and tried out the sfsimage script that he presented therein.
The idea is as simple as it is brilliant: combine open-source disk recovery tools like dc3dd with the great compression and usability of readonly squashfs filesystems. The imaged disk is piped directly into a virtual file inside the archive and some metadata of the acquisition is added alongside it to create a “forensic image”. This image can be mounted and inspected without elevated privileges using squashfuse – all the while usually only taking a fraction of the original disk size.
...November 6, 2019
To keep things short: I’ve had an Anet A8
3D printer for a while now and wanted to flash an updated version of its
Marlin firmware.
For some reason however, the bootloader was broken. So I couldn’t flash it with
avrdude by simply using a USB cable. For this case you can find lots of
tutorials on how to flash Arduino boards with a second Arduino. I didn’t have a
second Arduino at the time though … so why not just bit-bang it from a Raspberry
Pi instead?
...November 6, 2018
A while ago I stumbled upon podman, which touts itself as an alternative to
Docker. Not only does podman not use any big fat daemons™ but it makes it rather easy to run
containers in a user-namespace, i.e. with greatly restricted privileges on your system. The fun
thing is: you are still root within the container!
...October 5, 2018
Trying to implement some sort of automatic versioning based on your git commits or tags is not as
easy as it should seem. The idea is to use a feature built into the revision control system to
modify your project files and increment version counters automatically or embed commit information
into software builds - in my case: the --version output of Go applications built with
cobra.
...June 29, 2017
This is an update over my earlier post about PXE booting. I’ve learned a few more things and refined some rough edges.
Update: By now I’ve learned even more. See homelab/bootstrap for an updated guide.
My Motivation is similar to the last post: I started building my homelab with virtual machines. Most of them are based on a minimal CentOS 7 installation, and as such I have a lot of very similar systems. Yes, I could probably use containers to great effect. But I prefer the separation/isolation that I get from virtual machines on ESXi.
...March 9, 2017
Today I set out to setup a local CentOS mirror for quicker PXE installations of my virtual machines. In the long run this will probably be superseded by a Spacewalk machine (update: discontinued on May 31, 2020) and until now netboot.xyz has served me well. For the time being I just wanted a faster alternative.
...June 6, 2016
Sooner or later when setting up a server you’ll want to create some MySQL databses and users. If you’re not proficient in writing SQL queries or just wanted to use a nice GUI tool for the task, you’d need to connect remotely to your databse host. But of course you do not want to expose your MySQL port to the internet … Or suppose you want to debug some remote service, which is only accessible locally on the remote machine …
...May 16, 2016
Introduction
#
Some context
#
I have a NAS at home which runs on Debian Jessie. A NAS wouldn’t be a NAS without some storage, so I put in two disks with 4 TB each when I built it. Those two disks have actually been used in a setup with OpenMediaVault before that and already had a software-RAID on them. When migrating the disks (to Ubuntu at first) I learned about mdadm and that OpenMediaVault uses it. Great, that was a rather painless transition!
...