#!

A little tech blog

DragonFlyBSD - Hammer2 snapshot first steps

I am running DragonFlyBSD for quite some time now, but I never took the time to explore how Hammer2 really works concerning its snapshot abilities. This post is my first try at it and I intend it to be a little memo. First let’s gather some info about our mounted filesystems. $ mount | grep hammer2 serno/QM00005.s1d on / (hammer2, local) /dev/serno/QM00005.s1e@DATA on /build (hammer2, local) There are 2 different mountpoints interesting for us : / and /build.

DragonFlyBSD - Install Firefox syncserver with PostgreSQL

This article has been inspired by this article and by the official documentation of Mozilla Firefox syncserver. After reading the article by Mirabelette, I thought it would be nice to have my own syncserver, the only difference is that: - I run DragonFlyBSD - I want a PostgreSQL database as I know how to make backups of it. That led me to two different steps : - building syncserver under DragonFlyBSD - making syncserver use PostgreSQL as its database instead of SQLite.

Access CouchDB 2.x stats

After installing CouchDB I wanted to be able to access the /_stats page which is supposed to contain a lot of statistics about the running instance. The thing is that page is now displaying {"error":"not_found","reason":"Database does not exist."} After looking for a long time, the solution came from IRC channel #couchdb, and I need to access /_node/couchdb@127.0.0.1/_stats, where couchdb@127.0.0.1 is the node name you find in vm.args on ArchLinux. # Licensed under the Apache License, Version 2.

Build a live arch iso with sshd

I’m the owner of a little pc that I use as a server. It’s a headless machine, that means I don’t have a screen linked to it and that’s a pain when something goes wrong. Yesterday I decided to migrate the hard drive to a new 500 GB one. I wanted to be able to do it without plugging the screen that is so far away and any keyboard. To be able to do it I needed an iso that’s using dhcp and spawning a sshd daemon.

How to backup a btrfs filesystem using Amazon S3 and s3backer

A while ago, a friend of mine introduced me to s3backer, a tool that makes you able to use a s3 bucket as a support for any filesystem you want. I was looking for a way to backup my btrfs filesystem using btrfs send/receive commands, this looked like a good opportunity to do so. In addition I’ll use cryptsetup to encrypt the data on this filesystem. In order to achieve this you need a few things ready :

Install Boinc on CentOS 7

Boinc is a piece of software that allows to get involved into distributed computing in projects such as WorldCommunityGrid, Rosetta@Home or even LHC@Home. It’s both interesting and put your lazy computers to work. There’s of course a package in the EPEL repository allowing us to install the software easily. Let’s install it ! yum install epel-release yum install boinc-client Now that this is done, there’s the tricky part. Right after you’ve created your account on the website of the project that you’d like to join, you have to get the account key.

Extract and encode a cd to mp3 via the Linux cli

Ripping a cd and encoding it to mp3 is so easy it takes only 3 lines to be achieved. mkdir cd01 && cd cd01 # create a directory and put yourself in it cdparanoia -B # extract all the cd tracks in current directory as .wav files for t in *.wav; do lame -b 320 $t; done # encode all the files to mp3 320kbps It makes it rather easy to script and encode all your cd collection easily if needed.

Hugo - how I run it with systemd

I was looking for a solution to run a little blog, I did try wordpress but it’s not what I was looking for, then I discovered Hugo. It’s a Go program that will parse Markdown files (your articles in fact) and generate a static website for you. I don’t want to describe the full configuration on this article, but I want to share the systemd unit file I wrote. [Unit] Description=Hugo webserver After=network.

OpenBSD - Munin graphs generated without text

I did install munin and at first my graphs looked like this on this image. So as you can see this looks like something is wrong here. The solution was to modify my cron job to look like the following. su -s /bin/ksh _munin -c 'LC_CTYPE=en_US.UTF-8 /usr/local/bin/munin-cron' It seems that munin-cron does require a LC_CTYPE value, otherwise it generates without any text.