Hardware / OS

  • Dell Precision WorkStation 390

  • Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz

  • 4GB RAM DDR 667 MHz

  • HDD (going to be 4TB 2x)

  • with Ubuntu 18.04


  • root

  • jelko

  • luca

  • leo


SSH is running and available to the internet (!). Only Pubkey authentication is allowed.


ufw is administering the firewall rules. Default: Deny all. Internal networks: (HfK Inhouse) + (HfK VPN)

Rules (by 2019-02-04):

To Action From
-- ------ ----
[ 1] 22/tcp ALLOW IN Anywhere
[ 2] 8080/tcp ALLOW IN Anywhere
[ 3] 8000/tcp ALLOW IN Anywhere
[ 4] 80/tcp ALLOW IN Anywhere
[ 5] 445/tcp ALLOW IN
[ 6] 445/tcp ALLOW IN
[ 7] 8001/tcp ALLOW IN Anywhere
[ 8] 8002/tcp ALLOW IN Anywhere
[ 9] 8088/tcp ALLOW IN
[10] 5060 ALLOW IN
[11] 5060/udp ALLOW IN
[12] Anywhere ALLOW IN
[13] Anywhere ALLOW IN
[14] 5060/tcp ALLOW IN
[15] 5038/tcp ALLOW IN
[16] 5038/tcp ALLOW IN
[17] 8088/tcp ALLOW IN
[18] 5039/tcp ALLOW IN
[19] 9090 ALLOW IN
[20] 2812 ALLOW IN
[21] 8003 ALLOW IN Anywhere
[22] 2812 ALLOW IN
[23] 5083 ALLOW IN
[24] 5038 ALLOW IN
[25] 8010/tcp ALLOW IN Anywhere
[26] 8011/tcp ALLOW IN Anywhere
[27] 8012/tcp ALLOW IN Anywhere
[28] 8013/tcp ALLOW IN Anywhere
[29] 8014/tcp ALLOW IN Anywhere
[30] 8015/tcp ALLOW IN Anywhere
[31] 8016/tcp ALLOW IN Anywhere
[32] 8099 ALLOW IN Anywhere
[33] 2812/tcp ALLOW IN Anywhere
[34] 22/tcp (v6) ALLOW IN Anywhere (v6)
[35] 8080/tcp (v6) ALLOW IN Anywhere (v6)
[36] 8000/tcp (v6) ALLOW IN Anywhere (v6)
[37] 80/tcp (v6) ALLOW IN Anywhere (v6)
[38] 8001/tcp (v6) ALLOW IN Anywhere (v6)
[39] 8002/tcp (v6) ALLOW IN Anywhere (v6)
[40] 8003 (v6) ALLOW IN Anywhere (v6)
[41] 8010/tcp (v6) ALLOW IN Anywhere (v6)
[42] 8011/tcp (v6) ALLOW IN Anywhere (v6)
[43] 8012/tcp (v6) ALLOW IN Anywhere (v6)
[44] 8013/tcp (v6) ALLOW IN Anywhere (v6)
[45] 8014/tcp (v6) ALLOW IN Anywhere (v6)
[46] 8015/tcp (v6) ALLOW IN Anywhere (v6)
[47] 8016/tcp (v6) ALLOW IN Anywhere (v6)
[48] 8099 (v6) ALLOW IN Anywhere (v6)
[49] 2812/tcp (v6) ALLOW IN Anywhere (v6)

Output active numbered rules: ufw status numbered. You may find an introduction to UFW at Digital Ocean.

You can monitor incoming bandwith with: sudo tcptrack -i enp4s0 port 8001.


Backups of the are composited in three stages:

  1. Daily staging of changes via rsync

  2. Weekly staging of changes via rsync

  3. Monthly backup to external HDD via tar

For the monthly backups to happen the Backup HDD (labeled as such) needs to be plugged into the server.

Now on:


The cronjobs are scheduled as such. Update (2020-05-06): They are not scheduled in /etc/cron* but via monit (!) in /etc/monit/monitrc (Group Backup). → Disabled and moved to /etc/cron.d/angrezi-backup

# m h dom mon dow command
0 4 * * * /mnt/backup/
0 6 * * 0 /mnt/backup/
0 0 1 * * /mnt/backup/

Locations and Script

Locations backed up: /etc /var/angrezi /root /boot /opt /usr/local /srv /var/lib /var/mail /var/www /var/backups /var/local /var/opt /var/log Excluded are: --exclude="*.wav" --exclude=/home/*/.gvfs --exclude=/home/*/.cache --exclude=/home/*/.local/share/Trash --exclude=/media

The script for montly backup mounts the HDD, stores the last staged weekly changes to the HDD and unmounts the HDD. Do not fiddle with the HDD! If you want to initiate a manual backup, run ./mnt/backup/ as root.

# Where to backup to
dest="/mnt/angrezi_backup_hdd/$(date +%Y)"
# Create archive filename
date=`date +%Y-%m-%d`
hostname=$(hostname -s)
echo "" >> $Log
echo "---------------------------------------------------" >> $Log
echo "$date:: Full backup to $dest/$archive_file" >> $Log
now=`date +"%T"`
echo "$now:: starting..." >> $Log
echo "$now:: mounting Backup-HDD with UUID c59e6838-79ae-4994-bf1c-729ff9fae423 to /mnt/angrezi_backup_hdd" >> $Log
mount UUID="c59e6838-79ae-4994-bf1c-729ff9fae423" /mnt/angrezi_backup_hdd/
now=`date +"%T"`
echo "$now:: Mounting successful" >> $Log
now=`date +"%T"`
echo "$now:: starting backup..." >> $Log
tar -cvpJf "$dest/$archive_file" /mnt/backup/weekly #storing permissions, using xz as compression, verbose
now=`date +"%T"`
if [ $rc != 0 ]; then
echo "$now:: Backup failed!" >> $Log
if grep -qs '/mnt/angrezi_backup_hdd ' /proc/mounts; then
echo "HDD is mounted, Backup still failed" >> $Log
echo "HDD is not mounted! Is it plugged in?" >> $Log
exit 1
echo "$now:: Finished." >> $Log
now=`date +"%T"`
echo "$now:: Unmounting..." >> $Log
umount /mnt/angrezi_backup_hdd
if [ $rc != 0 ]; then
echo "Unmounting not successful. Yikes. Exiting." >> $Log
echo "Unmounting disc successful. Exiting. Thanks for choosing our services." >> $Log
exit 1


We use monit to monitor all "important" services at


Domains Mapped:






  • AirTime

  • Angrezi Controller

  • Fileserver (Directory Listing with styling)

  • Icecast (Port 8000)

  • Cockpit (Port 9090)

Angrezi Controller

Self-build controller interface. Source on GitHub.


  • darkice (installed via apt service manually placed in /etc/systemd/system/darkice.service and removed in init.d)

  • angrezi-master-recorder (self made)

  • angrezi-stream-monitor (self made)

  • angrezi-file-watch (self made)

After changes to the services in /etc/systemd/system/ you can reload them with systemctl daemon-reload + systemctl enable <service name>


angrezi-file-watch uses inotifywatch -m -r -e create to monitor the directories of recordings. If a file has been created (aka a new show has been recorded) it automatically appends chattr +a to prevent deletion.


  • onboard sound card (Intel) disabled in /etc/modprobe.d/blacklist-angrezi.conf

  • Pulse Audio disabled in /etc/pulse/client.conf

Using ALSA with Behringer UMC202HD with following config in asound.conf

pcm.stream_in_32 { # the Behringer hardware we use does only support 32 bit by default.
type dsnoop
ipc_key 5978293
ipc_key_add_uid false
hint.description "Angrezi: USB Card hw1 for Stream IN, 32 bit, dsnooped for mulitple listeners"
slave {
pcm "hw:1,0"
channels 2
format S16_LE # this will only affect the plug pcm.stream_in_16.
rate 44100
pcm.stream_in_16 {
type plug
hint.description "Angrezi: USB Card hw1 for Stream IN, 16 bit, dsnooped for mulitpl e listeners"
slave.pcm "stream_in_32"

two capture devices are available:

  • stream_in_32 with S32_LE (hardware default)

  • stream_in_16 with S16_LE (software conversion via plug)

Test with (software monitor): arecord -f S16_LE -r44100 -c2 -D stream_in_16 | aplay -D hw:1 -


The local studio audio in form stream_in_16 is broadcasted to Icecast via Darkice 1.3. Darkice only supports 16 bit sampling (which makes the downconversion necessary). Recording in DarkIce was disabled after a bug corrupted the recording on server reconnection occurred.

This local Darkice stream functions as master source in AirTime / liquidsoap.


FM Relay to Radio Weser TV

Radio Weser TV gets a special relay stream out (with a defined URL), which is made to never fail (go silent):

The stream is produced by relaying our live stream in a seperate liquidsoap instance (service: angrezi-relay-out-radioweser) which falls back to a playback of mp3 files if the stream goes silent for more than 10 seconds.

Fallback music (MP3, 44100, min. 192 kbit/s, Stereo) must be placed in /media/storage/share/Automation/Live-Out-RWTV/Music-MP3/ and fallback Jingles in /media/storage/share/Automation/Live-Out-RWTV/Jingles-MP3/. Jingles and Music are randomly picked in the radio 1:5. (If all fallback is empty as well, the script will fallback to a single audio file placed at /var/angrezi/relay-out-radioweser-fallback.mp3.)

Notice: The service needs to be restarted if the music folder is updated.

TODO: We could add a special config on Icecast so even if liquidsoap fails, there would be a fallback. TODO: Hide the mountpoint from Icecast.