Pi Hole
-
@scottalanmiller said in Pi Hole:
@scottalanmiller said in Pi Hole:
@marcinozga said in Pi Hole:
@aaronstuder said in Pi Hole:
What about installing it on Vultr?
You can install it on any supported Linux system, mine runs in Centos 7 LXD container. It's how you set it up as DNS proxy/server. Although local DNS is always preferred, once your queries are cached, latency is much lower than querying cloud instance.
Plus, the Raspberry Pi will pay for itself within a year of Vultr use.
Not realistically. Figure $60 for a working RP setup minimum, not including space, heat or power draw (all small, but non-zero.) Vultr would be $30. It would take two years to break even, three years to RP cost savings minimum.
If it lasts 3 years. Typically, power failure will cause the SD card to corrupt. Then, you'd have to reinvest into another SD card and start all over.
And with the time/value of money, it makes the Vultr approach better and better, unless you already have a place on your network where you can easily run the Pi Hole. Running dedicated RPs rarely makes sense as they are too low powered and costly compared to a tiny VM on Vultr or similar. An RP doesn't really have any more power than a tiny Vultr VM. And Vultr VMs are growing in capacity faster than RPs are version to version.
They're also more capable because more applications are designed for x86_64-bit procs instead of the RPi's ARM procs. If you're doing something with an RPi, then you probably have intentions of using the GPIO pins instead of anything more software-based that would be better suited for a VM, either on a local laptop, a home lab, or Vultr.
-
@scottalanmiller said in Pi Hole:
@marcinozga said in Pi Hole:
@aaronstuder said in Pi Hole:
What about installing it on Vultr?
You can install it on any supported Linux system, mine runs in Centos 7 LXD container. It's how you set it up as DNS proxy/server. Although local DNS is always preferred, once your queries are cached, latency is much lower than querying cloud instance.
Plus, the Raspberry Pi will pay for itself within a year of Vultr use.
Not realistically. Figure $60 for a working RP setup minimum, not including space, heat or power draw (all small, but non-zero.) Vultr would be $30. It would take two years to break even, three years to RP cost savings minimum.
If it lasts 3 years. Typically, power failure will cause the SD card to corrupt. Then, you'd have to reinvest into another SD card and start all over.
I mean, its not hard to install Raspbian,
sudo apt-get update && sudo apt-get upgrade, then curl Pi Hole. Could have another setup in an hour after failure.With containers and ansible, I had mine ready in probably less than 3 minutes. You can probably do it by hand just as fast.
-
@scottalanmiller said in Pi Hole:
@scottalanmiller said in Pi Hole:
@marcinozga said in Pi Hole:
@aaronstuder said in Pi Hole:
What about installing it on Vultr?
You can install it on any supported Linux system, mine runs in Centos 7 LXD container. It's how you set it up as DNS proxy/server. Although local DNS is always preferred, once your queries are cached, latency is much lower than querying cloud instance.
Plus, the Raspberry Pi will pay for itself within a year of Vultr use.
Not realistically. Figure $60 for a working RP setup minimum, not including space, heat or power draw (all small, but non-zero.) Vultr would be $30. It would take two years to break even, three years to RP cost savings minimum.
If it lasts 3 years. Typically, power failure will cause the SD card to corrupt. Then, you'd have to reinvest into another SD card and start all over.
And with the time/value of money, it makes the Vultr approach better and better, unless you already have a place on your network where you can easily run the Pi Hole. Running dedicated RPs rarely makes sense as they are too low powered and costly compared to a tiny VM on Vultr or similar. An RP doesn't really have any more power than a tiny Vultr VM. And Vultr VMs are growing in capacity faster than RPs are version to version.
They're also more capable because more applications are designed for x86_64-bit procs instead of the RPi's ARM procs. If you're doing something with an RPi, then you probably have intentions of using the GPIO pins instead of anything more software-based that would be better suited for a VM, either on a local laptop, a home lab, or Vultr.
Anything with a script based language is architecture agnostic. One of the many benefits of Python.
-
Pi hole OpenVPN server
https://github.com/pi-hole/pi-hole/wiki/Pi-hole---OpenVPN-server
-
Anyone try on Fedora? Installer did not work here.
-
@scottalanmiller said in Pi Hole:
Anyone try on Fedora? Installer did not work here.
Mine was always on Raspbian.
-
On Fedora there were errors on exit, but it said that it finished. And SELinux is not supported, but you can turn that off or whatever. But it never seemed to configure anything. Testing Ubuntu now.
-
It fails on setting static IPv4 address in Fedora.
-
@marcinozga said in Pi Hole:
It fails on setting static IPv4 address in Fedora.
It doesn't even go through as much of the script.
-
It "just worked" on Ubuntu.
-
@scottalanmiller said in Pi Hole:
It "just worked" on Ubuntu.
That's what I have it installed on and it was a breeze.
-
@scottalanmiller said in Pi Hole:
It "just worked" on Ubuntu.
No luck with CentOS 7 on Vultr.
Debian 9 on Vultr worked perfectly.
-
what a name....hahaha...
-
@aaronstuder said in Pi Hole:
@scottalanmiller said in Pi Hole:
It "just worked" on Ubuntu.
No luck with CentOS 7 on Vultr.
Debian 9 on Vultr worked perfectly.
I think that they are a bit ambitious on their list of OSes that work with that script
-
I'm on Vultr as well.
-
@stuartjordan said in Pi Hole:
what a name....hahaha...
I personally never took a look at the project because I thought it only worked with a Pi...
Bad name really...
-
-
Pi-hole will run on most Debian-based distro's and is the preferred platform for it.
We officially support the following:
Raspbian: Jessie (lite / with pixel)
Ubuntu: 14.04 / 16.04 / 16.10
Fedora: 24 / 25
Debian: 8.6
CentOS: 7.2.1511 / 7.3.1611 -
It's working on 17.04.
-
I just installed in on Fedora 25. Installer is expecting ifcfg-eth0 (or whatever your interface is) file in /etc/sysconfig/network-scripts and if that doesn't exists, it fails. I touched that file and installation went smooth.
